The Official SaaStr Podcast: SaaS | Founders | Investors

I think after 100+ days and with 9 apps vibe coded into production with @replit used over 500,000 times we’re just getting going. And … I think key to that is that I’m now “Replit Fluent”.

What does that mean? It’s a state where I know how to vibe well enough (without a developer), and I know the app and its capability and limits well enough, that I can basically see any app I want to build in my head, and now know how to completely prompt it and shepherd it to production … before I start.

I can now will almost any ‘normal’ app into existence that I want to build. Without a developer.

I remember in the early days of Cursor my son actually paid out of his own pocket for the first time since ChatGPT (he’s awfully smart). He said back then Cursor could now do 90% of his coding for him, but “for most people, it might be 10% or less.” I didn’t get what he meant at the time, but today I do. It’s more than being “good at prompting”. It’s understanding the system well enough to already understand its outputs, its limitations, and exactly how it works in practice — before you start. It’s being truly fluent in the agent.

This isn’t to say I don’t still have bumps and that some things end up harder or longer than anticipated. But pretty much now I can will most things into production on Replit very predictably with the agent — because I can already plan them out fully in my mind before I first “prompt” the agent.

What that means in practice is 3 things:

First, I 100% know if a project will work now before I start in Replit. If I can see it to completion in my mind, now I can finish it to a reasonably high standard. That is a huge boon. When I started ‘prosumer’ vibe coding all of 100+ days ago :), I couldn’t finish my first project. In part, it was because I picked a very complex product to start. But there are many stories of others in similar boats. They can’t finish their vibe coded app. But now — I have a 100% chance of finishing a project, and in roughly the amount of time I budget for it in mind before starting.

Second, now I’m merely just time constrained in what I can build. I already have a couple of jobs. I have $100m+ to invest in fresh capital at SaaStr Fund, and running SaaStr itself is an eight figure business. Both take a lot of time. But I set aside about 1.5-2 hours a day to vibe code. That’s my budget. That’s what I can build now.

Third, maintenance and new features and upgrades to existing apps I’ve vibe’d consume more and more of my time budget. This of course is true of any software. It just catches up to you with prosumer vibe coding. So now that I can basically truly build anything I want, the question becomes — do I have enough time to make it great? Or should that time go into making my existing 9 apps even better? I now can make a pretty good v1 of anything I want to. But getting to great takes time. It always has.

I think it’s a big deal to be able to get to this state without a developer and without being a developer. Yes, I co-founded a B2B startup that went from $0 to $200m+ ARR so I had that going for me (which is a lot). But the fact I am now “Replit Fluent” and that now my only limit on building is my own time … is super interesting.

The world is so, so much different than 12 months ago.

There is a learning curve in AI. In almost every great tool and every great agent, at least for now. Don’t let anyone tell you there isn’t. Ignore any marketing that tells you there isn’t. If you want to get most AI Agents into production, e.g. a great AI SDR or great AI Support, and have it work well — you have to learn and often train the AI. IMHO you also have to learn tools like Replit to truly get out of them what they promise. You will be much, much better 100 hours in.

But if you get to a State of Fluency in a prosumer vibe coding app … it’s like a superpower you never had before.

Thanks for reading SaaStr AI: How To Sell, Scale, and Win! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

Marchelle Mooney, VP of Sales at Mangomint joined us at 2025 SaaStr Annual + AI Summit for one of our best deep dives on sales and GTM in vertical SaaS yet.

Marchelle brings a unique perspective to vertical SaaS sales—she’s a former hairdresser and salon owner who transitioned into SaaS sales leadership. At Mangomint, she leads sales for a deep vertical SaaS platform serving salons and spas. Her journey from thinking “SaaS just meant you had attitude” to becoming a VP of Sales at a fast-growing vertical SaaS company gives her insights that bridge the gap between traditional enterprise SaaS playbooks and the reality of selling to SMB vertical markets. She credits much of her learning to Jason Lemkin and the SaaStr community.

MangoMint today processes over 1,000,000 appointments a month for over 5,000 salon and spa customers.

Top 5 Learnings: The Vertical SaaS Sales Playbook

1. Deep Domain Credibility is Non-Negotiable—Build Trust From the First Click

The insight: Your founding team needs deep insider knowledge of the vertical before shipping a single line of code. This isn’t about market research—it’s about speaking the language fluently.

Marchelle emphasizes that credibility builds trust instantaneously. From the moment a prospect lands on your website, they should see language that feels like their own world. At Mangomint, this means understanding the difference between how a hair salon refers to their “client” versus how a med spa calls them a “patient.”

She’s witnessed demos end immediately because a rep used the wrong terminology. In one case, calling a “patient” a “client” was enough for the prospect to disconnect—that’s how critical getting the language right is in vertical SaaS.

The key principle: You’re solving problems for customers who don’t know what they don’t know. Many verticals have been using antiquated playbooks for years. Your product needs to answer questions they haven’t even thought to ask yet, and you can only do that with genuine insider knowledge.

Why it matters: When Marchelle started at Mangomint in 2018, every tenth call involved unseating pen and paper booking systems. Today, it’s maybe once a year. The market has evolved, but the need for domain expertise hasn’t—it’s just shifted to different problems, like introducing AI to customers who’ve only used ChatGPT to decide between sushi or Thai for dinner.

2. Win by Eliminating Choice—Your SMB Customers Have Analysis Paralysis

The insight: Vertical SaaS customers, especially SMBs, struggle with too many options. Unlike enterprise buyers who might want flexibility and customization, your vertical customers want you to tell them the right way to do things.

Marchelle uses a simple analogy: “What are we doing for dinner? We could do sushi. We could do Thai.” As soon as there are two options, she’s in analysis paralysis. Her customers feel the same way.

The approach: At Mangomint, they’ve eliminated choice not by limiting features, but by presenting a clear, opinionated path forward. They’re going after a vertical where customers aren’t using any AI products except ChatGPT for personal decisions. This creates an opportunity to be prescriptive about the right solution.

Why this works in vertical SaaS: Your customers are busy operators. A salon owner is literally doing a demo while bleach is processing on a client five feet away, and they might need to pause the call to shampoo someone. They don’t have time to evaluate ten different workflow configurations—they need you to show them the one that works.

3. PLG + SLG = The Perfect Vertical SaaS Motion (When Done Right)

The insight: The debate about “PLG is dead” versus “SLG is back” creates a false dichotomy. In vertical SaaS, you need both—but PLG serves a different purpose than in traditional SaaS.

Mangomint offers a 21-day free trial (not a freemium product), but they don’t use it to replace sales—they use it to supercharge sales conversations.

How it works:

* Prospects can start a trial and explore the product independently

* But the trial becomes a data goldmine for sales reps

* When Sally from the waxing studio goes straight to the memberships tab, rage clicks, and leaves, the AE gets an alert

* The rep sends a personalized text: “Hey Sally, I’m Marshelle with Mangomint. Thanks for starting a trial. Looks like you’re building memberships. What memberships do you currently offer? Would you like to see how Jessica’s spa down the street does $30k a month in memberships using Mangomint?”

The framework: “The product proves, the humans frame the value.”

The trial isn’t about product-led conversion—it’s about conversation-led conversion informed by product data. This is personalization at scale, using PLG as a tool rather than a replacement for sales.

4. Master Sub-Vertical Dominance Through Micro-Niche Fluency

The insight: Within your vertical, there are sub-verticals, and each one thinks they’re unique and special. Treat them like the unicorns they believe they are.

At Mangomint, a hair salon owner doesn’t use the same terminology as a med spa owner. These aren’t just semantic differences—they represent different mental models, different workflows, and different Facebook groups (literally—hair salons don’t join the same Facebook groups as med spas).

The strategic approach: Marchelle calls this “breaking through the silence.” Unlike horizontal SaaS where you’re cutting through noise, in vertical SaaS you’re often dealing with silence—isolated communities that don’t cross-pollinate. Each sub-vertical lives in its own bubble.

The execution: Go on “sub-vertical dominant sprints.” Master one micro-niche completely before expanding. Know the words they use, train your reps with 100% fluency, and dominate that sub-vertical before moving to the next.

Why this matters: This isn’t about having a tighter ICP or a cuter logo—it’s about fundamentally different go-to-market execution. The playbook is flipped upside down from traditional SaaS.

5. Make Onboarding Your Sales Weapon—Trust Before Revenue

The insight: In 2022, Mangomint moved their customer success team into the new revenue organization and transformed them into an onboarding team. This wasn’t just an org chart change—it was a strategic repositioning of how they build trust.

The approach: Mangomint offers completely complimentary white-glove onboarding. They migrate client lists, client history (sometimes medical history), sales history, service menus, staff schedules, and payroll setup—all for free, often before the customer has paid a single dollar.

Why SMB vertical SaaS is different: These customers don’t want to pay to migrate their mission-critical data. Unlike enterprise customers who might have budgets for implementation services, SMB customers in vertical markets expect you to make it easy.

The payoff: By building trust first through free onboarding, you can later layer in charging for premium outcomes and growth features. Trust is Mangomint’s number one company value “by design, not by default.”

This approach also extends to support. While many SaaS companies are moving to AI-first support, Marchelle heard the HubSpot CEO and others say “support is done, it’s AI-owned” multiple times at SaaStr. But when she asked her VP of Customer Operations about their AI support tool Finn, the response was revealing: “Finn doesn’t know who canceled the haircut on Tuesday at 2pm with Jessica.”

The vertical SaaS reality: You need a different support playbook. Problems that come up might not be software problems—they’re vertical problems. You need to show customers you can help them with these issues and teach them to solve them over time. (Good news: They eventually have a report for that, so the person won’t reach out again about canceled appointments.)

Top 3-4 Mistakes Marshelle Made (and What She Learned)

1. Initially Underestimating the Importance of Vertical Athletes in Hiring

The mistake: Early on, there was likely too much emphasis on finding reps with SaaS experience and big logos on their resumes, without enough weight on their ability to operate at the altitude of the vertical.

The learning: A Salesforce enterprise rep with an impressive resume might be completely wrong for selling to a salon owner doing a demo while processing bleach. The mismatch isn’t about capability—it’s about altitude and agility.

The correction: Marchelle now focuses on hiring “vertical athletes”—reps who are agile enough to understand the customer and operate at their level. While she still values SaaS sales acumen, the persona is most important: “I’m looking for the Grade-A top tier USDA approved beef, grass-fed, all of it.”

She notes that all their top performers now have prior vertical experience, but she’s confident they could perform in other companies too. The key is being holistic about team building, not fixating on either domain expertise OR SaaS experience—you need both.

Current approach: She’s now focused on immersing reps in the trade during onboarding and exploring AI simulators that can plug in actual conversations from the vertical, allowing reps to practice until they’re fluent in the language their customers speak.

2. Not Moving Fast Enough on Trial Response Times

The mistake: Not initially understanding that the window for connecting with business owner/operators is measured in minutes, not hours or days.

The learning: When a salon owner is taking 30 seconds out of their busy day to search for software and start a trial, they’re looking RIGHT NOW. If you don’t call them within minutes of trial signup, you’ve probably missed the initial window.

The correction: Marchelle implemented a strict SLA: contact within minutes, with no exceptions. The sequence is automated message + phone call + text follow-up—always from a human, always immediately.

Why this matters differently in vertical SaaS: Unlike traditional SaaS buyers who might be doing structured vendor evaluation, vertical SMB customers are often impulse shopping during a brief break in their operational chaos. Speed isn’t just important—it’s everything.

3. Trying to Be Everything to Everyone (Learning to Say No to Non-Ideal Customers)

The mistake: Getting caught up in the hype cycle and not proactively disqualifying prospects who aren’t a great fit, particularly franchises in Mangomint’s case.

The learning: Growth loves focus, and this is even more critical when you’re trying to be the default leader for a specific vertical. Being “proudly not for everyone” isn’t just positioning—it’s a crucial strategic decision.

The correction: Now, if Marchelle sees any indication that a prospect might be a franchise (something that “gives her an itch on her spine”), she proactively has that conversation and lets them know Mangomint isn’t great for them—before they self-identify this during the trial.

The framework: Don’t push into markets that aren’t pulling you there. Stay niched down enough that you can be genuinely best-in-class for your ICP rather than mediocre for a broader market.

4. Misunderstanding What “Customer Success” Means in Deep Vertical SaaS

The mistake: Initially thinking about CS as a separate team function rather than as a company-wide responsibility.

The learning: In vertical SaaS, you can’t afford to have CS as a siloed function. The problems customers face are too intertwined with product, sales, and operations.

The correction: Mangomint “killed their CS team” but made customer success more important as an organizational principle. They implemented shared ownership across the entire company:

* Product owns success metrics and time-to-value

* Sales owns reality—helping customers understand both what the product can and can’t do for their specific situation

* Customer Operations (live in-app support) owns the inevitable problems that arise

The vertical SaaS insight: This isn’t just a “nice to have” or aspirational—it’s foundational. When you’re serving very specific verticals with deep operational needs, everyone needs to be aligned on customer success, because the problems don’t fit neatly into traditional support categories.

The Big Takeaway

Marchelle’s final point encapsulates the entire vertical SaaS philosophy: Don’t be just another software—be the only software they’ll need.

Mangomint’s mission is to make every salon and spa more profitable. Right now, they’re focused on being best-in-class, the default software for their ICP. But down the line, they’ll be the software that teaches every salon and spa to bolt on AI.

In the future, customers will likely book appointments through AI agents—but Mangomint will get there by building trust first, ensuring their vertical feels safe migrating in that direction, and making sure customers don’t feel like they’re ripping out the human element of their business.

As Macshelle jokes with a woman working the speaker room about having “a job AI cannot take,” only to learn about a shop down the street where robots are doing eyelashes—the future is coming. But in vertical SaaS, you guide your customers there. You don’t just sell them software. You become their trusted partner for navigating change in their industry.

That’s the vertical SaaS playbook. It’s not just a tighter ICP or a cuter logo. It’s fundamentally different from the ground up.

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

The VC fundraising landscape has completely transformed in the last 18 months, and most founders still don’t realize just how dramatically the rules have changed.

After analyzing 1,000 VC pitch decks and calculating 400,000+ startup valuations on SaaStr.ai, and having countless conversations with both sides of the table, the data is unambiguous: traditional paths to venture funding have essentially closed for the majority of B2B companies.

The capital isn’t gone—there’s actually more money in the market than ever. It’s just almost all flowing to a completely different type of company than it was 24 months ago.

Jason Lemkin joined us for a LIVE SaaStr AI Wednesday to walk us through the data, and do live Q+A

If you’re a B2B founder growing 80% annually at $10-30M ARR with solid unit economics and happy customers, you might think you’re in a strong position to raise. You’re probably not.

If you’re planning to raise your Series B based on performance that would have secured funding in 2022, you need to recalibrate immediately.

Benchmarks have shifted so dramatically that roughly 80% of VCs who would have enthusiastically funded solid B2B companies 18-24 months ago are now passing—not because those companies got worse, but because an entirely new category of hypergrowth AI-native startups has reset every expectation in the market.

This deep dive breaks down exactly what’s happening, why it’s happening, and—most importantly—what you need to do about it right now. We’ll walk through the actual data from top-tier growth funds, show you the real benchmarks you’re being measured against, and give you a clear-eyed assessment of your options whether you’re pre-revenue, scaling past $10M, or approaching $100M ARR.

The worst thing you can do is stay in the dark about where you actually stand. Let’s fix that.

Top 5 Takeaways

* The bar for VC funding has skyrocketed dramatically. AI-native companies are now scaling from $1M to $100M ARR in 8-11 quarters (versus the previous “top quartile” benchmark of 19-20 quarters), fundamentally resetting investor expectations across the board.

* 80% of traditional B2B VCs who would have funded you 18-24 months ago won’t fund you today. Capital is flooding into hypergrowth AI-native companies, leaving traditional SaaS startups—even strong ones—struggling to raise regardless of solid fundamentals.

* The latest top quartile metrics are higher. And harder. At $1m-$5m ARR, VCs expect 500% growth. At $10-25M ARR, VCs expect 100%+ growth. At $50-100M ARR, they want 90% growth with 120% net revenue retention. Below these numbers, you’ll face an uphill battle regardless of market or team quality.

* Three consecutive great months can flip you from unfundable to fundable. If you’re close to the benchmarks but not quite there, demonstrating acceleration over 3-4 months can completely change investor perception—but you need to be honest about where you actually stand first.

* Don’t assume the next round is coming. The single biggest founder mistake is burning cash with the expectation they’ll raise another round. If your odds are below 60-70%, run your company with the cash and revenue you have—not the funding you hope to get.

The New Reality: AI-Native B2B Speed vs. Traditional SaaS Growth

Let’s start with the chart that explains everything happening in VC right now. Iconiq Capital, one of the leading growth funds originally backed by Mark Zuckerberg’s money, recently published benchmark data that crystallizes the entire market shift.

The Old Top Quartile: Getting from $1M to $100M ARR in 19-20 quarters (roughly 5 years) used to represent elite performance—the kind of company VCs actively competed to fund.

The New Standard: AI-native companies are doing it in half that time or less:

* Glean (AI enterprise search): 11 quarters

* 11 Labs (AI speech): 8 quarters

* Perplexity: Even faster

* Cursor, Anthropic, Lovable, Replit: Less than 4 quarters from $1M to $100M

Here’s the question every VC is now asking: If you have $100M, $500M, or $2B to deploy, where on this spectrum are you putting your money? The answer is obvious for most—and it’s not with the companies on the right side of that chart, even though they’re very strong “top quartile” performers.

These hypergrowth AI companies didn’t even exist 24 months ago. You literally couldn’t invest in them. So the top quartile traditional SaaS companies were about as good as it got. But now? That entire world has been reset.

Two Categories of Fundable Companies

Bessemer Venture Partners broke down their investment thesis into two distinct categories, and the contrast is striking:

Supernovas (AI-Native Companies)

* Growth expectations: $1M to $40M in year one, $125M+ by year two

* Gross margins: Often 25% or even negative (yes, negative)

* Revenue per employee: $1M+

* Examples: Gamma hit $50M ARR in one year; Replit and Lovable are burning massive cash but scaling at unprecedented rates

Shooting Stars (The Best of Traditional Software Startups)

* Growth expectations: $1M → $3M → $12M → $40M → $100M (still better than triple-triple-double-double over 5 years)

* Gross margins: 60%+ required

* Revenue per employee: $164K minimum

* Key point: Even for these “traditional” companies, the bar has never been higher

These AI-native companies often have terrible gross margins because of token costs and infrastructure expenses. But VCs don’t care. They’re remarkably headcount-efficient (often under 100 employees at $50-100M ARR), and the growth trajectory is so explosive that margin concerns become secondary.

The Paradox of Public vs. Private Expectations

Here’s where things get confusing for founders: while VC expectations for startups have skyrocketed, expectations for public B2B companies have actually declined.

Public company reality:

* Average growth rates are at all-time lows

* “High growth” for public companies is now just 30% (it was 70%+ in 2021)

* Companies growing 20-30% are getting 20x+ ARR multiples

* HubSpot is growing 19% and trading at premium valuations

Private company reality:

* VCs expect 150-250%+ growth depending on stage

* The bar keeps rising, not falling

* You need to massively outperform public company benchmarks to even get a meeting

Why the disconnect? Public markets value profitability and predictability at scale. Private markets are chasing the next category-defining winner, and AI has created more potential winners than ever before—meaning the competition for capital is fiercer despite there being more capital available.

The Real Benchmarks: What Top Quartile Actually Looks Like Now

From Iconiq’s data analyzing their portfolio of venture-backed B2B companies, here are the top quartile benchmarks you need to beat to get funded:

Less than $10M ARR:

* Growth: 515%

* Net Revenue Retention: N/A (too early)

* Magic Number: N/A

* Revenue per employee: N/A

$10M-25M ARR:

* Growth: 170%

* Net Revenue Retention: 130%

* Payback period: 1.0x

* Revenue per employee: $160K

$25M-50M ARR:

* Growth: 100%+

* Net Revenue Retention: 125%

* Payback period: 1.2x

* Revenue per employee: $180K

$50M-100M ARR:

* Growth: 90%

* Net Revenue Retention: 120%

* Payback period: 1.5x

* Revenue per employee: $220K

Read that carefully. This is not “top 1%” or “exceptional” performance. This is top quartile—meaning 75% of top venture-backed B2B startups fall below these numbers. And yet, this is effectively the minimum bar to get funded by most VCs today.

If you’re at $30M ARR growing 60-70%, you are building a truly great company. You might even have excellent fundamentals, happy customers, and strong unit economics. But you’re walking through the desert when it comes to VC funding. Maybe—maybe—you can raise if you have extraordinary circumstances (dominant market share, insane NRR, clear acceleration path), but you’ll need exceptional reasons beyond being likable and having a solid business.

The Data from 1,000 Pitch Decks and 400,000 Valuations

Through SaaStr AI’s pitch deck review tool (which has now analyzed 1,000+ VC pitch decks in just two weeks), we’re seeing the market reality play out in real-time:

Series A averages (across ~108,000 calculations):

* AI-native companies: ~$5-6M ARR, growing 180%

* AI-enhanced SaaS: ~$5M ARR, growing 100%

* Traditional SaaS: ~$5M ARR, growing 75%

The traditional SaaS number (75% growth) won’t get you funded in most cases. It’s just not enough.

Grade distribution from pitch deck reviews:

* 17% received A- or above (these are “top tier” and highly fundable)

* Average grade: B-

* Median grade: B-

Here’s what that means: Most founders are in the B range—they have something, they’re in the zone, but it’s not enough. A B- is not enough to get funded in today’s market. You need to be brutally honest about where you actually stand.

What Revenue Efficiency Really Looks Like

One trend that’s accelerating: companies are getting dramatically more headcount-efficient as they scale. The Iconiq data shows this progression:

* Early stage: ~$90K revenue per employee

* Growth stage: $220K revenue per employee

* Pre-IPO: $400K revenue per employee

The most successful companies—especially AI-native ones—are proving you can get to $50-100M ARR with under 100 employees total. That level of efficiency is becoming the new expectation.

How to Navigate This Market: Practical Advice

#1. Get Brutally Honest Feedback on Your Actual Funding Odds

Don’t go into fundraising blind or optimistic. Use these approaches:

Ask your existing investors directly: Put them on the spot. “On a scale of 1-10, how good is our startup? More importantly, what are the exact odds—give me a percentage—that we can raise our next round?” Push for a number. Most experienced investors know exactly what your odds are.

Use objective AI tools: Upload your pitch deck to SaaStr AI’s review tool. It will benchmark you against all other founders, tell you where you stand, show you what VCs will think of your traction, and give you honest odds of getting funded. Don’t let ego or optimism cloud your judgment—you need to know where you actually stand.

#2. Figure Out What You’re Number One At

This has always mattered in venture, but VCs are obsessed with it now. With hundreds of AI SDR tools, thousands of AI startups, and constant innovation noise, you must be number one at something specific.

Not “we’re going to be better than the other guy eventually.” Not “we have a great team and we’re executing well.” What are you number one at right now? And show it with data, even if you’re early.

Every VC is looking for the next category winner, the next OpenAI of their vertical. If you can’t articulate—and prove—what you’re number one at, you’ll get lost in the noise.

#3. Understand Your Zone

If you’re above the Iconiq benchmarks: You’ll likely get funded. Maybe not from everyone, but you’ll get multiple term sheets if you have good founders and an interesting space.

If you’re within 20% of the benchmarks and burning very little: You’re in the “maybe” zone. Not everyone will want to invest, but you’ll find someone—especially if you have other compelling factors (new product coming, obvious acceleration path, strong market position). Just be prepared to take a lot of meetings.

If you’re significantly below the benchmarks: Be honest. It’s going to be incredibly difficult if not impossible. Don’t pretend otherwise. Don’t run out of money because you assumed you could raise. The worst outcome in startup life is going from funded to unfunded and not knowing it until you’re out of cash.

#4. The Three-Month Rule

If you’re close but not quite fundable, here’s the good news: VCs have incredibly short memories for previous underperformance if you can show recent acceleration.

If you’re at $20M ARR growing 70% (below the 90% benchmark), but people are taking meetings because they believe in you and your space, three consecutive great months can completely flip the narrative.

Show them one month where instead of growing 6% MoM you grew 10%. Then another month at 10%. Then a third at 10%. By month three or four, investors will stop caring about the previous 30+ months of slower growth.

You don’t need a year of great performance. You don’t need 10 years. But you probably do need three solid months, maybe four to be safe. Two months of acceleration will get you meetings, but investors will want to see one more month before committing.

Define a clear inflection point (launched AI agents, hired key exec, released version 3.0), explain what changed, and show sustained acceleration from that point forward.

#5. Not Everyone is AI-Native or Bust (But 80% Are)

The VC world has split into rough camps:

~80% of VCs: Only interested in AI-native, hypergrowth companies. If you’re not building the next Replit or Cursor, they’re out.

~20% of VCs: Still believe in traditional triple-triple-double-double SaaS. These investors (like Rory on the 20VC podcast) argue that if you can hit those benchmarks, they can see a path to IPO at multi-billion dollar valuations, and that’s good enough to make money.

The takeaway: Don’t jump off a bridge if you’re not AI-native. You just need to realize the fundraising dating game will be harder. You’ll need more meetings, more outreach, and thicker skin. But the investors exist—they’re just the minority now.

#6. Don’t Assume the Next Round is Coming

This is perhaps the most critical point: The biggest mistake founders make is assuming they’ll be able to raise another round.

Look, if you’re at $20M ARR growing 60%, you’re still compounding toward something potentially great. If you don’t need outside capital and can grow at 50-60% for several years, you’ll build a hundreds-of-millions-dollar company. That’s objectively a great outcome.

But no VC will fund you at those metrics in today’s market.

So stop operating like the next round is guaranteed. Stop burning cash with the assumption you’ll raise. If your odds are below 60-70% (and be honest—they probably are if you’re not hitting those benchmarks), run your company on the cash and revenue you have.

This isn’t bootstrapping shame. This isn’t giving up. This is basic financial reality. Either you can raise money to run at a loss in exchange for hypergrowth, or you can’t. If you can’t, don’t run at a loss.

#7. Be Wary of Dated Advice

This is subtle but important: many investors haven’t been actively investing recently, or they’re disconnected from what’s actually getting funded in 2025. They’ll give you advice that sounds reasonable but is actually from 2021 or 2022.

They’ll tell you your company is “hot as Travis Kelce and Taylor Swift” because you have great unit economics and a strong team. But if you don’t have an ounce of AI in your product and you’re growing like a 2021 company, nobody will actually fund you at the valuations you expect.

The same goes for executives you hire. Be extremely careful about VPs of Sales or CMOs who haven’t been in the market since 2021. They’ll want to build teams of 20-40 people, hire 12 people in finance, create 21-person demand gen teams. That model is increasingly dated and unsupportable.

Get advice from people who are actively in the market right now. Not people who were successful three years ago but haven’t adapted to the current reality.

Special Considerations for Different Stages

Pre-Revenue and Pre-Seed Companies

If you’re pre-revenue or pre-seed, the signaling game becomes even more important:

Join a top accelerator if possible: Y Combinator, Neo, South Park Commons, or Entrepreneur First. Not because the programs themselves will make you a better company (though they might), but because the investor attention is at an all-time high. Yes, it’s somewhat unfair that 20-year-old founders with two weeks of revenue and massive hubris can close rounds at $25-60M post-money valuations just because they’re YC-backed. But that’s the reality. These accelerators attract VCs like flies.

If you can’t do an accelerator, you need alternative signals:

* Epic team (proven founders or engineers with strong reputations can raise pre-revenue)

* Exceptional pilots (three pilots that could each become multi-million dollar contracts)

* Early traction metrics (100K users for a free product, 10 real paying customers, something demonstrable)

The “Almost There” Companies

This might be the toughest spot to be in right now: You’re at $10-20M ARR, growing 50-70%, maybe even cash flow positive or close to it. Through 2023, this is exactly when private equity firms would start calling to buy you. Companies at $15-30M ARR growing 30-50% were getting acquired at 6-10x revenue multiples.

That’s largely stopped. PE firms learned that many of these acquisitions (made in 2020-22) haven’t played out as expected. Revenue hasn’t proven as durable, AI is making categories unstable, and growth has decelerated faster than models predicted.

What to do:

* Take PE calls if they come, but don’t expect them like you would have 18 months ago

* Push forward and get bigger—the PE desert between $10-30M ARR is real right now, but it may not be permanent

* AI-enhance your product aggressively so you can potentially access either VC funding or PE acquisition in 12-24 months with better metrics

* Study companies like Filevine (legal case management) that infused true AI agents and raised $400M from Excel at a $3B valuation

The Professional Services Question

Founders often worry about having 50% professional services revenue versus 50% software ARR. Don’t.

Palantir—worth over $300 billion today—was almost entirely low margin software + services until just before going public. Their gross margins were in the teens or 20% range … until suddenly they weren’t. Every VC will tell you services are bad, but Palantir literally proved the opposite at massive scale.

More importantly, “forward deployed engineers” are the hottest hiring trend right now (up 12x in the last year according to Iconiq). Why? Because AI products require training, configuration, and month-long deployments. This isn’t technically professional services, but it’s close.

Don’t worry about the revenue mix. Worry about whether you have real product-market fit and can scale very rapidly.

The SAFE vs. Equity Question

Quick primer: Most VCs don’t want to do SAFEs unless the check is immaterial or they have to.

SAFEs offer investors no protection, no control over dilution, limited (if any) pro-rata rights, and no board seat. There have been cases where founders simply kept all the SAFE money because legally “it’s neither equity nor debt, so investors have no rights.”

When do SAFEs make sense?

* Large funds writing tiny checks ($100K from a multi-billion dollar fund—they don’t care)

* Forced by accelerators (YC makes you use SAFEs; everyone accepts it for small amounts)

For a $2M raise, you might be able to do it on a SAFE, but unless you have a hot hand, don’t push it. Optimize for getting the deal done, not the specific instrument. Don’t over-optimize on terms that don’t really matter if it means losing the deal.

Understanding Valuation Expectations

Here’s an uncomfortable truth: Not only have VC expectations gone up, but founder valuation expectations have also skyrocketed—especially for AI companies.

Recent Y Combinator batches have had founders with minimal revenue raising at $25-60M post-money valuations, hoping to triple that in another round within a year. From the VC side, that’s an extraordinarily difficult bet to make work economically.

So you have VCs expecting faster growth and founders expecting higher prices with lower dilution. This puts pressure on the entire model. You might get the term sheet, but the performance expectations will be equally extreme.

The Bottom Line

The venture capital market in 2025 has fundamentally bifurcated:

The winners: AI-native companies scaling at unprecedented rates (even with negative gross margins), and exceptional traditional SaaS companies hitting 100%+ growth at $25M+ ARR with strong margins and capital efficiency.

Everyone else: Walking through a funding desert, regardless of having solid businesses, happy customers, or reasonable growth rates.

This isn’t fair. It isn’t rational in many ways. A company growing from $20M to $35M in a year (75% growth) is still building something substantial and valuable. But VCs are human, capital is following the most explosive growth stories, and you can’t fight the market.

What you can do is be ruthlessly honest about where you stand, understand your actual odds, make informed decisions about burn rate and runway, and either position yourself to hit the benchmarks or run a sustainable business without assuming the next round will materialize.

The worst outcome isn’t building a $100M revenue company over 10 years instead of 3 years. The worst outcome is running out of money because you didn’t realize the rules had changed.

Top Things Founders Need to Know Right Now

* Use objective tools to assess your fundability. Don’t rely on gut feel, friendly advisors, or outdated advice. Upload your deck to SaaStr AI, ask your investors point-blank for percentage odds, and benchmark yourself against current market data—not 2021 metrics.

* If you’re not AI-native and not hitting top quartile growth, prepare for 5x more meetings. You’ll need to talk to everyone. Get comfortable with rejection. Have a thick skin. The investors who will fund you exist, but they’re now the minority.

* Three great months can change everything. If you’re close but not quite there, focus obsessively on showing acceleration over 90-120 days. Define your inflection point clearly, show consistent execution, and investors will forgive previous underperformance.

* Make the cash you have last. Run your company as if the next round isn’t coming unless you’re highly confident (>70% odds) you can raise. The biggest failure mode is burning through capital while assuming funding is inevitable when it’s not.

* Be number one at something specific and provable. “We’re going to be great” doesn’t cut it. “We have 10x the accuracy of competitors as measured by [specific metric]” or “We’re the fastest-growing solution in [narrow category] with [specific data]” is what VCs want to hear.

* Don’t confuse PE recapitalizations with venture rounds. If someone wants to give you $20M but take 50% of your company, that’s not growth capital—it’s a partial buyout. These deals exist and might make sense, but they’re fundamentally different from venture funding.

* The bar is high and getting higher, but it’s not arbitrary. VCs are comparing you against the fastest-growing companies in history. They’re not trying to be mean—they’re trying to make returns in a market where AI-native companies are redefining what’s possible. Your job is to either meet that bar or build a great business without venture capital.

The market has spoken. The only question is whether you’re going to listen and adapt accordingly.

For more resources, benchmark your company for free at SaaStr AI (saastr.ai), where you can get instant valuations, pitch deck reviews, and see how you stack up against thousands of other B2B startups in real-time.

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

We’re back! Our latest deep dive with Jason Lemkin (SaaStr), Rory O’Driscoll (Scale Venture Partners), and Harry Stebbings (20VC).

And come join the team LIVE at SaaStr AI London, Dec 1-2!! More here.

Bottom Line Up Front

The classic venture playbook is almost … no more.

* Traditional SaaS metrics like burn multiples—once the gold standard for evaluating capital efficiency—are being rendered obsolete by AI-native companies growing at unprecedented speeds with radically different unit economics.

* Meanwhile, founders with objectively good numbers (triple-triple-double-double growth, solid burn multiples) are getting rejected by VCs focused exclusively on AI breakouts.

* The message is stark: if you’re not AI-first, raise capital now at any reasonable price, consolidate where possible, and prepare for a world where even “perfectly good” $15M ARR companies have “zero value to VCs.”

The stakes extend beyond individual companies:

* With 6,700+ unicorns and only 15 IPOs year-to-date, massive portfolio consolidation is inevitable.

* Tech PE firms face existential questions as AI agents reduce seat-based revenue and products that stayed static for a decade now require constant reinvention.

* And towering above it all: OpenAI’s plan to consume more energy than India within 8 years, raising fundamental questions about whether the economics of AI can ever pencil out at scale.

The Burn Multiple Paradox: Why AI Breaks the Rules

The conversation opened with Iconiq’s 73-page State of Software report, which revealed a counterintuitive finding: AI-native companies under $100M ARR have terrible free cash flow margins (-126% versus -56% for non-AI companies), yet their burn multiples are actually better because they’re growing so explosively fast.

Rory O’Driscoll explained the fundamental concept: “It’s basically how many dollars of ARR do you get out of each dollar that you’re spending, right? What is the efficiency you’re creating for each dollar of venture capital you’re lighting on fire?”

The math seems simple: if you’re valued at 10x ARR and you spend $2 to add $1 of ARR, you’ve created $10 of market cap from $2 invested. But as Rory cautioned, this only works when multiple hidden assumptions hold true:

* The ARR is actually real (not inflated by aggressive accounting)

* Net retention accounts for real churn (fast growth can hide massive customer losses)

* Gross margins are sustainable (hyper-growth can mask deteriorating unit economics)

* No massive capex requirements (true for most SaaS, decidedly not true for AI infrastructure companies)

Jason Lemkin emphasized the existential risk everyone’s ignoring: “David Sacks coined the Burn Rate Multiple and I think when everything was the same in SaaS and B2B in 2021, it made a lot of sense. All the companies were the same… Then AI breaks it because we’ve never seen growth like this. But the margins are lower.”

The team agreed that while burn multiples remain useful for comparing companies, they’re no longer sufficient as the primary valuation framework. As Rory put it: “We are not in Kansas. There’s a whole bunch of implied assumptions in there which is why even though we love those kind of metrics… we’ve actually come back to saying there’s a real advantage in seeing the GAAP revenue accounting also to make sure all the money is for lack of a better word just showing going up for real, right?”

The Brutal Reality: Good Companies Can’t Get Funded

Harry Stebbings raised what may be the most troubling question for founders: “I have many companies with good burn multiples and they are not getting love. They are not getting attention and they are going, ‘Harry, I don’t get it. I’ve been brought up to understand burn multiples, to understand growth, like what is going on?’”

Rory’s response was unsparing: “There’s only two ways of pricing a deal. You price a deal on hope or you price a deal on the multiples. When you price a deal on hope and growth, you can lean in on anything, right? And you can get prices that quote unquote make no sense because the growth ultimately comes and it all pays off. Once you start valuing things on quote unquote the fundamentals today, right? Then you can value a public company because at 400 million, it’s not nothing. But… a $15 million revenue company that’s perfectly good and has reasonable growth is actually of zero value to a VC because we’re in the upside option game, right?”

This crystallizes the existential crisis for non-AI-native companies: even with objectively strong metrics, if the path to a massive IPO isn’t crystal clear, VCs simply won’t engage. The bar isn’t just high—for certain categories, it functionally doesn’t exist.

Jason was even more direct about the advice gap: “I hear too many folks, they’re like, ‘Oh, you’re triple triple double double or better. You’re you’re golden. Don’t worry, kids.’ And I think that’s terrible advice in 2025. Terrible advice.”

He described portfolio companies at $15M ARR growing 100% with good burn ratios that can’t get funded unless they’re ultra-breakout AI stories. His advice? “If Scale wants to put in money at 250 on that deal… take that deal now.” Even if it means accepting a lower valuation than 2021-era thinking would suggest.

Rory concurred: “Those companies, Jason, you’re right, is that… you should just get the deal done. Raise at a reasonable price. Continue to grow, but be capital efficient. Don’t get lost in your just your burn multiple. Focus on your cash. You know, if you’re right about your business, you’ll be right in the end.”

The Kingmaker Effect: Why Proximity to Harvey or Abridge Matters

Harry introduced a phenomenon that’s reshaping competitive dynamics: the power of King-Made Startups: “The mimetic—and this sounds obvious given the sheep plate analogies applied to venture—but it’s how concerned investors are by going against a kingmaker whether it’s Harvey or Abridge or any of the king-made companies. We have a couple of companies which are the second or the third and going against the kingmaker in the valley is the most unpopular thing in the world.”

The strategy is working: “Founders listening, when you’re raising, raising to deter others from raising is a really working strategy right now.”

Rory acknowledged the effect but with important caveats: “Do I believe that thinking exists in venture? Yes, I do. I don’t fully share it but I acknowledge that you know you have to factor it into your decision-making and your risk analysis… If your first customers are also VC-backed companies, then you get this dual loop. But look, the reality is… if you’re selling to oil and gas companies, they barely can tell the Sequoias from their KPs.”

The takeaway: the kingmaker effect is real and powerful in Silicon Valley-centric markets, but less dispositive in traditional enterprise segments where buyers don’t track VC pedigree.

Jason noted what makes AI different from previous cycles: “What is a little different in AI is that there are in many cases there’s not an established brand and there’s so much change and so much new budget and so much confusion that so many buyers are under pressure and have a desire to make a purchase… Being number one is so powerful when people know they want to do something—they’ve got to do an LLM for legal.”

He cited the death match between Lovable and Replit as an example: “They’re both at nine figures in revenue. They won’t kill each other, right? But… you want to be that brand that nervous folks don’t know who to buy.”

Valuation Insanity or Justified Vision? The $1 Trillion Question

Harry posed the fundamental question: “Do you think we are near peak madness guys or do you think we’ll look back and laugh at ourselves for having this conversation given the might and the size of the markets that we’re entering?”

Rory’s answer was unequivocal: “One of two things has to be true because that’s actually the interesting insight… If that massive transfer of labor doesn’t happen, then all these valuations are wrong by an order of magnitude. One of two things is going to happen in the next five or seven years, right? Either… you are going to see pretty profound productivity changes. You’re going to see companies like OpenAI hitting two to 300 billion in revenue super quickly. That’s option A. Option B… AI is still going to be wonderful, but you’re going to have a readjustment period that’s going to make your head hurt.”

Jason worried about a different risk—loss ratios: “What I wonder is, are our loss ratios correct? If as venture VC firms, especially with larger funds… as long as we get it, as long as we’re cool, if 80% of our unicorns implode or blow up or make no money, more importantly, just don’t make money for venture… But I’m pretty sure we didn’t model it right in 2020 2021.”

He continued: “It feels almost risk-free again like it did in 2021. It feels risk-free. Valuation and loss ratios both matter.”

Rory distinguished between two types of fund-killing mistakes: “You can imagine people failing because their picking ratio was bad and they just put too much of the money in bad companies, right? Or you can also imagine people failing where all their companies were great, were good companies. But the prices were so bad that they only made a subpar return even though they only had a 30% loss ratio, right? So you both—I mean both things can—I mean the sad thing about venture is you actually have to get both things right to make money.”

On whether this is sustainable, Rory offered a sobering forecast: “I personally think that the rate of adoption forecast over the next four or five that’s implicit in all these data center assumptions will in retrospect prove to be too optimistic.”

Sam Altman’s Energy Crisis: Willing a Trillion Into Existence

The discussion turned to OpenAI’s staggering infrastructure requirements: planning to 125x energy capacity in 8 years, requiring more energy supply than India’s current total capacity.

Jason framed the scale in visceral terms: “Just the little hundred billion that he’s doing with Nvidia… needs more power than all of New York City. So it will only be a couple years where the cities of the future don’t even have humans in them… Our country will be dotted with these Stargates that are larger than New York City with only hundreds of people working in them and the equivalent of millions of digital minds.”

Harry expressed fundamental skepticism about the financing: “I feel very stupid honestly because I look at a trillion dollars required to fund data centers for Open AI alone. I’m like I don’t know where that money comes from… Sovereigns don’t have that. I mean, well, actually, funny enough, they do. I mean, they’d have to put it all in.”

Jason offered a nuanced view: “I think Sam is just willing as much of this into existence as possible because of the future. And if we come up short, if it’s $400 billion or $600… we don’t have to buy all the GPUs. It would be okay if they had to last 6 years instead of 3 years or whatever the depreciation schedule is.”

Rory distinguished between directional vision and operational reality: “The sheer act of willing something into existence like this has been just amazing… Whatever the prize is for being the best company in AI, OpenAI is going to get that prize, right? And that’s his job. And he’s doing it better than any other CEO of this decade. That’s true. It is also equally true… just because a CEO says we’re going to treble next year doesn’t mean we should buy real estate and hire people as if we’re going to treble.”

His concern: “Five years from now… we said, ‘Oh, I get it. OpenAI is still the best company on the planet for AI.’ Its growth rate has slowed to a shockingly small 50 60%. It’s freaking amazing. It’s doing 30 billion dollars growing at 50%. It’s astonishing. But maybe they don’t need a trillion dollars of capex this week.”

Zuck’s AI Crisis: When Even Loyalists Lose Faith

Harry made a rare public admission about his largest public holding: “I always thought that Zuck earned the right to do the next thing. He earned the right to do the next thing. And I have to say my faith in Meta’s AI strategy has just dwindled and dwindled and dwindled… I’m looking at Alex Wang. I’m looking at the treatment of Yann LeCun. I’m looking at how they structure teams and I’m going this is not wellrun. OpenAI, Anthropic are coming for you.”

Jason was characteristically blunt: “He is not the communicator that Sam Altman is… If I don’t understand where the hell he’s going and Harry doesn’t understand… if we don’t understand he’s not one of the great communicators at the moment… He’s a crappy communicator right because none of us—I’m not saying his AI strategy isn’t S tier but I don’t think any of us understand that not for the life of us where the hell it’s going.”

Rory offered both defense and critique: “You could also say it’s just a harder bet… The bet for the last 15 years was riding this brilliant invention that you had in 2003 and just optimizing it. And that’s hard but it’s a lot easier than now you got to invent a whole new thing a second time.”

He continued with the existential framing: “The big picture comment is if you spend two hours a day on ChatGPT that’s two hours a day that you are not spending on Facebook and we live and die on our attention. So we’re just going to make s**t until somehow we get people to come back and play with us.”

The panel agreed that while Zuck was “clear just this week he’d rather burn the 20 billion in operating income and fail than become irrelevant,” the specific strategy remains opaque even to sophisticated investors.

Rory’s prediction: “If there was a Couch bet… that you know some version of this AI strategy will not produce meaningful revenue despite a $20 billion burn and will look more like Meta VR and less like Instagram and WhatsApp… I’d take that bet.”

The Great Consolidation: Fivetran + dbt Labs and What It Means

The panel discussed unicorn Fivetran’s talks to acquire unicorn dbt Labs as a template for the massive consolidation wave that must happen with 6,700 unicorns and only 15 IPOs year-to-date.

Rory framed the math: “We’ve processed 15 out the IPO gate year to date. So 20 for the year. That implies we got 30 years of this stuff to get through. What they’ve done… every time two companies combine… it takes two midsize companies… and makes a bigger… This is part of the job venture is going to have to do to whip their portfolios into shape to be IPOable.”

Jason identified the key enabler: “The fact that Andreessen is the lead or close to it in both deals makes it much simpler on many levels… If Andreessen owns 20% of Fivetran and 20% of dbt and you combine them, it does kind of suck when you own 20% of a portfolio company, combine it with another leader… and now I own 8% after the deal.”

But he emphasized why it still makes sense: “20% of something that’s not going public is not nearly as interesting as 8% of something that is going public, right? And if you believe that it’s not a continuum of value… but rather it’s like electron states, there’s just a gap and then you got to go to the next state.”

Rory stressed the risk: “The fatal mistake that always scares me is not the delusion… The thing that scares me is you go from a decent deal that’s wellrun where you know everything about it to merging with something else and then the combined entity screws it up. That to me is the really shitty outcome where you took your 20% bet and turned it into 8% of a disaster.”

The PE Reckoning: Why Tech Private Equity’s Model Is Under Threat

Harry asked whether tech PE firms are questioning their business model: “When they see the multiples that you can get on the money that’s being moved by your Kushner and your big firms, combined by the increased loss ratio that will happen from an increasingly volatile new AI world. Are you suddenly going, ‘S**t, I’m not getting paid for the risk that I’m taking on the multiple on the upside given the displacement on the downside.’”

Jason identified the core problem: “These products didn’t change from about 2008 until 2023. They’re the same products… It wasn’t just that we had high NRR which was the spreadsheet glue for the PE model. It was the fact… that the products can’t be static and AI is the accelerant there… The rate of change is unprecedented in business software. Unprecedented.”

He gave the example of his first venture investment, Pipedrive: “It took him four years to launch a mobile app… Could you imagine today waiting four years to launch your AI co-pilot? Like you’re dead in the water.”

Rory expanded on the implications: “There was 15 years where we made the same product and you didn’t have to think that much about product direction at the macro level… You look at Salesforce 2002 and 2022, it’s the same thing, right? And that’s now changed… On the PE side, you’re exactly right. You look at that and you go, I might get away with next 10 years making the same thing… And so maybe I’m entering into some kind of tech risk that I’ve never before internalized. A new tech risk.”

Jason added the seat model crisis: “Now that we’re running 12 AI agents, you know, we only need two seats of Salesforce. Because you don’t have the people. We just don’t… It makes the PE model worse and our jobs harder. It just literally… makes it even tougher for PE to buy these seat models.”

Key Takeaways

* Burn multiples are useful but insufficient: AI companies break traditional SaaS assumptions around gross margins, churn visibility, and capex requirements. VCs must look at GAAP revenue and absolute cash positions, not just efficiency ratios.

* Non-AI-native companies face a binary funding environment: Even with triple-triple-double-double growth and good burn multiples, companies perceived as “legacy” struggle to raise. Founders should take reasonable deals now rather than optimizing for price.

* The kingmaker effect is real but not dispositive: In Valley-centric markets, being the #2 or #3 behind a heavily-funded leader significantly impacts fundraising. But in traditional enterprise segments, customer buying decisions care less about VC pedigree.

* Massive consolidation is inevitable: With 6,700 unicorns and 15 IPOs year-to-date, the math demands hundreds of mergers. Firms like Andreessen that own both sides of deals have structural advantages in making these combinations happen.

* Tech PE’s classic SaaS funding model faces existential threats: The shift from static products (unchanged for 10+ years) to rapidly evolving AI-powered solutions introduces unprecedented tech risk. Meanwhile, AI agents reducing seat-based revenue attacks the core cash flow assumptions.

* OpenAI’s ambitions may be both directionally correct and operationally overstated: Sam Altman is “willing a trillion into existence,” but even if the vision is right, the actual capital deployed and timeline may be far more modest than the rhetoric suggests.

* Meta’s AI strategy lacks clarity even to sophisticated investors: Despite a $20B annual commitment, investors including long-term bulls struggle to understand the strategic endgame beyond “avoid irrelevance.”

* Loss ratios matter as much as entry valuations: Even if VCs pick the right companies, paying prices that require 100x returns to generate acceptable fund returns means most funds will fail even with good picking.

Quotable Moments

On burn multiples in the AI era:

* “A $15 million revenue company that’s perfectly good and has reasonable growth is actually of zero value to a VC cuz we’re in the upside option game.” — Rory O’Driscoll

On outdated advice:

* “I hear too many folks, they’re like, ‘Oh, you’re triple triple double double or better. You’re golden. Don’t worry, kids.’ And I think that’s terrible advice in 2025. Terrible advice.” — Jason Lemkin

On the kingmaker phenomenon:

* “Going against the kingmaker in the valley is the most unpopular thing in the world. Founders listening, when you’re raising, raising to deter others from raising is a really working strategy right now.” — Harry Stebbings

On OpenAI’s prize:

* “Whatever the prize is for being the best company in AI, OpenAI is going to get that prize.” — Rory O’Driscoll

On Sam Altman’s energy vision:

* “Our country will be dotted with these Stargates that are larger than New York City with only hundreds of people working in them and the equivalent of billions of digital minds.” — Jason Lemkin

On financing the impossible:

* “I feel very stupid honestly because I look at a trillion dollars required to fund data centers for Open AI alone. I’m like I don’t know where that money comes from.” — Harry Stebbings

On Meta’s communication problem:

* “He’s a crappy communicator right because none of us—I’m not saying his AI strategy isn’t S tier but I don’t think any of us understand that not for the life of us where the hell it’s going.” — Jason Lemkin

On tech risk in PE:

* “Maybe I’m entering into some kind of tech risk that I’ve never before internalized. A new tech risk.” — Rory O’Driscoll

On consolidation math:

* “20% of something that’s not going public is not nearly as interesting as 8% of something that is going public.” — Jason Lemkin

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

Tomasz Tunguz of Theory Ventures brought together a strong panel at SaaStr Annual and AI Summit for a deep dive on enterprise partnerships. An Enterprise Partnerships Bootcamp: How to Land, Scale, and Win with Linear, Omni, Theory Ventures, and Vesey Ventures.

* Christina Cordova – Chief Operating Officer at Linear, the purpose-built tool for planning and building products. Previously spent 7.5 years at Stripe where she started their partnerships organization, and later led partnerships and platform initiatives at Notion.

* Colin Zima – CEO of Omni, rebuilding BI and analytics to create a tool accessible to both developers and end users with enterprise readiness and AI agility. Previously spent nearly 10 years at Looker, bringing deep data ecosystem expertise to his current venture.

* Julia Huang – Founding Partner at Vesey Ventures, a fintech fund based in New York and Israel. Specializes in brokering partnerships between portfolio companies and financial incumbents, with deep expertise in enterprise sales motions within regulated industries.

* Tomasz Tunguz – Founder and General Partner at Theory Ventures, focusing on early-stage enterprise software investments with particular emphasis on go-to-market strategy and partnership development.

Top Take-Aways:

Christina Cordova (Linear): Start with partnerships that are instrumental to your product experience first, not distribution. Build credibility by talking to 10+ users within an enterprise before approaching executives.

Colin Zima (Omni): Partner with companies at your growth stage rather than chasing Fortune 50 whales early. Strategic investments from key ecosystem players can bootstrap credibility faster than organic growth.

Julia Huang (Vesey Ventures): In enterprise partnerships, you’re controlling for reputational risk, not financial loss. Find the sponsor who can tell your story back to you—they’re your true advocate.

Tomasz Tunguz (Theory Ventures): The cost of building integrations has become trivial with AI. Companies should have 25-100 integrations by Series A, not 5-10 like in the past decade.

The Partnership vs. Sales Decision: Where to Start and Why

The fundamental question facing early-stage B2B companies isn’t whether to pursue partnerships or direct sales—it’s understanding when partnerships become instrumental to your product experience and business growth. Linear’s approach exemplifies this strategic thinking.

“For us on the partnership side, we really started with partnerships that we felt were instrumental to the product experience first and foremost,” explains Christina Cordova. “At a certain point we were a very small company. We viewed some integrations as hyper-critical to the core product—things like integrations with Slack and GitHub.”

Quotable Moment – Christina Cordova: “We were able to have 10 conversations for a very large account and then we go back to the CTO and say, ‘Actually, we’ve talked to 10 people on your team and here’s what they say about the products that you’re using today.’ And then they’re kind of like, ‘Oh, you know more about my team and my org than I do.’”

This product-first approach to partnerships reveals a crucial insight: the most successful early partnerships aren’t just distribution plays—they’re core to delivering your value proposition. Linear recognized that without seamless integrations to the developer tool stack, their product experience would be fundamentally incomplete.

The Evolution of Partnership Strategy: From Product to Platform

Omni’s partnership evolution illustrates how dependency relationships in the enterprise ecosystem have fundamentally shifted over the past decade. Colin Zema’s perspective, shaped by his experience building partnerships at both Looker and Omni, reveals how the technology stack has matured.

“We can’t deliver our product without a whole bunch of other products tightly coupled to us,” Zema notes. “We need to get data into a database. We need a database to run the queries and then we need to present them out to the user. At every level to be successful, we have these contingencies on these other tools and they need to want to work with us also.”

Quotable Moment – Colin Zema: “I can go call all of you in the audience and ask you to come try my tool. It’s a lot more successful if Snowflake is doing that alongside us and if DataBricks is doing that alongside us.”

But the partnership landscape has evolved dramatically. When Looker was scaling, companies often had to sell both ETL solutions and their core product to be successful. Today, that infrastructure is assumed. “Everyone sets up a database, buys a database, sets up Fivetran, and now how we partner with them is more important than whether they exist,” Zema explains.

This shift has profound implications for partnership strategy. Modern partnerships are less about filling infrastructure gaps and more about creating seamless experiences within established ecosystems.

The Credibility Framework: Building Trust at Enterprise Scale

Establishing credibility emerges as the central challenge—and opportunity—in enterprise partnerships. Julia Huang’s experience working with financial incumbents provides a masterclass in managing reputational risk.

“The risk that we were controlling for isn’t financial loss. It was reputational risk,” Huang explains. “If their reputational brand and image gets tarnished, then they don’t have the ability to sell their customers anything else.”

Quotable Moment – Julia Huang: “My practical advice on finding your sponsor is the person who can tell your story back to you in that meeting. Somebody who’s starting to get it—the wheels are turning. That person will at a minimum give you feedback, the true feedback of where you sit within the product and how long it’s going to take.”

This insight fundamentally changes how startups should approach enterprise partnerships. Success isn’t just about proving product-market fit—it’s about demonstrating that your company can enhance, rather than threaten, your partner’s reputation with their customers.

The credibility challenge manifests differently across organizational levels. When selling to individual contributors, you’re proving quarterly value. When selling to VPs, you’re demonstrating annual impact. When engaging C-level executives, you’re articulating a three-year vision that aligns with their strategic objectives.

The API-First Partnership Playbook

Linear’s approach to building their partnership ecosystem offers a replicable framework for developer-focused products. They started with a GraphQL API and “shipped the API and didn’t do a whole lot after that to bring on partners outside of just having really good technology.”

This hands-off approach generated over 150 integrations organically. But Linear’s strategy evolved with market conditions, particularly around AI partnerships. “More recently, we’ve thought more about partnerships in the context of AI in particular because it’s such a fast-moving space,” Cordova explains. “There we’re going much more the partnership route where we’re going to all the companies and spending time with those companies.”

The strategic shift reveals when to move from passive platform provision to active partnership development: when the technology landscape is moving so quickly that organic adoption can’t keep pace with market opportunities.

Strategic Investment as Partnership Accelerant

The relationship between strategic investment and partnership success has become increasingly important. Omni’s experience securing investment from both Snowflake and DataBricks illustrates how financial alignment can accelerate go-to-market execution.

“Both of those in isolation have been huge legitimizers for us as a company,” Zema notes. “It gives us opportunity to build more product over time and lets us bootstrap credibility with our customer base.”

But the investment decision requires careful consideration. “We’re way more concentrated in terms of our partnership,” Zema explains. “The databases are the most important to us. So it’s Snowflake, BigQuery, RedShift, DataBricks… We can’t move the needle for Snowflake’s earnings, but Looker did. And so we want them to trace that curve out and see that forward.”

The Timing Revolution: When to Hire Your First BD Person

Traditional wisdom suggested waiting until $5-10M ARR before engaging seriously in partnerships. Tomasz Tunguz challenges this conventional thinking: “I think that’s really changed. One of the earliest employees at Looker, Kenan was BD. He was responsible for 10 to 15% of bookings from the very first.”

The catalyst for this change? Technology has dramatically reduced integration costs. “The cost to produce an integration, an API integration, is really trivial,” Tunguz explains. “One of our companies had a hack day where in 90 minutes an engineer spun up an integrations factory.”

Quotable Moment – Tomasz Tunguz: “You kind of expect as a Series A investor a company having 25, 50, maybe even 100 integrations by the time of the Series A because it’s so simple. Whereas 10 years ago, that would be almost unimaginable because the engineering work required to do that would have taken years or hundreds of thousands of person-hours.”

This technological shift means companies can now reasonably be expected to have 25, 50, maybe even 100 integrations by Series A—a scale that would have been unimaginable a decade ago.

Christina’s hiring philosophy provides practical guidance: “When you’ve tested it with your current team, what are the kinds of partnerships, what do they look like? Once you know you want to do more of these and it’s going to be really a full-time job, then it makes sense to hire someone to do these in a dedicated fashion.”

Channel Strategy: Bottom-Up vs. Top-Down Approaches

The choice between bottom-up and top-down partnership approaches depends fundamentally on your product’s natural adoption pattern. Products that win on user experience often benefit from grassroots adoption, while products that solve C-level problems may require executive-first approaches.

Linear’s bottom-up strategy leverages product enthusiasm to build executive credibility. “We’re able to have 10 conversations for a very large account and then we go back to the CTO and say, ‘Actually, we’ve talked to 10 people on your team and here’s what they say about the products that you’re using today,’” Cordova explains.

This approach works because it builds internal advocacy before asking for budget authority. The CTO realizes, “You know more about my team and my org than I do,” creating credibility that transcends traditional sales processes.

The Enterprise Breakthrough: Acquisition-Driven Expansion

Some of the most successful enterprise partnerships emerge through acquisition scenarios. Both Linear and Omni have experienced rapid expansion when their startup customers get acquired by larger enterprises.

“We started off with this company of like 50 people. It was ultimately acquired by a Fortune 20 company and that Fortune 20 company started using us,” Cordova shares. “We were able to take that one customer and expand it to like a 500 seat account very quickly.”

This pattern reveals a crucial insight: startup CEOs often retain significant influence post-acquisition and can advocate for tools they’ve grown to depend on. It’s a channel strategy that requires no additional sales effort—just product excellence that creates passionate advocates.

The AI Partnership Acceleration

The current AI wave has created new urgency around partnership development. Both Linear and Omni are adapting their partnership strategies to capture AI-driven opportunities.

Linear is building new APIs specifically for AI agents and taking a much more proactive approach with AI partnerships. “There’s now this weird discretionary pool of cash that people can use for tools if you check the AI box,” Cordova notes.

This budget reality means partnerships that can credibly position AI capabilities have accelerated paths to enterprise adoption. The challenge is ensuring the AI positioning represents genuine value rather than checkbox marketing.

Measuring Partnership Success: Beyond Revenue Attribution

Traditional partnership metrics often focus on directly attributable revenue, but the most successful partnerships create value that’s difficult to measure directly. Brand credibility, ecosystem positioning, and customer trust building may be more valuable than immediate revenue in early-stage partnerships.

Colin’s experience illustrates this: “It’s a lot more successful if Snowflake is doing that alongside us and if DataBricks is doing that alongside us.” The value isn’t just in the leads generated—it’s in the credibility transfer that happens when established players validate your solution.

The Partnership Portfolio Strategy

As companies scale, partnership strategy must become more portfolio-driven. Not every partnership will generate immediate results, but a balanced approach can create multiple paths to success.

“I think as long as you place the right ones, it can be very fruitful,” Cordova explains about Linear’s approach to partnering with both established AI companies and emerging startups. “Some of those bets aren’t going to work out, but I think as long as you place the right ones, it can be very fruitful.”

This portfolio thinking requires discipline about resource allocation and clear criteria for partnership investment levels.

The Procurement Reality Check

Enterprise partnerships ultimately face the procurement reality of large organizations. Colin’s experience provides sobering perspective: “You realize that the conversations are actually completely different. The person may not have any idea what your company does. They have no ability to buy.”

This reality requires partnership professionals to develop multiple messaging frameworks for different organizational levels and decision-making authorities. The technical value proposition that resonates with practitioners may be irrelevant to budget holders focused on strategic outcomes.

Building Partnership Infrastructure for Scale

Successful partnership programs require infrastructure that can scale beyond individual relationships. APIs, documentation, partner portals, and co-marketing resources become essential as partnership volume increases.

Linear’s evolution from building individual integrations to providing platform capabilities illustrates this maturation. The infrastructure investment pays dividends when partners can self-serve many of their integration needs.

The Strategic Investment Decision Framework

When considering strategic investments from partners, companies must balance accelerated credibility against potential constraints on future partnerships. The decision requires careful analysis of how strategic alignment affects future partnership optionality.

Omni’s concentrated approach—focusing strategic investment on their most critical database partners—provides a model for how to think about these tradeoffs. “You can’t do a hundred of them either,” Zema notes.

Top Mistakes the Speakers Made (And What We Can Learn)

Christina Cordova’s Mistake: Being too hesitant to hire dedicated BD resources. “I’m always very hesitant to make that hire because I’ve seen it so many times how it can so easily go wrong.” While caution is warranted, this conservative approach may have delayed partnership scaling at Linear compared to more aggressive competitors.

Colin Zima’s Mistake: Admitting to poor executive communication. “I did a conversation last week with a CIO of a Fortune 50 and at the end he was pretty much like I had no idea what you were talking about.” This reveals insufficient preparation for senior-level conversations and highlights the critical importance of adapting technical messaging for different audiences.

Julia Huang’s Mistake: Allowing strategic partnerships to drag on too long without clear milestones. She describes partnerships with banks that “take years to come into fruition” and mentions going through reorgs “three times or four times” during one strategic partnership. Better partnership governance and clearer exit criteria could prevent resource drain.

Tomasz Tunguz’s Mistake: Overestimating the speed of integration development. While he correctly identifies that integration costs have dropped dramatically, his “90 minutes” integration factory example may set unrealistic expectations for quality, scalable integrations that require proper testing, documentation, and ongoing maintenance.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

This week:

* How the new Replit v3 is the Future: Agents Managing Agents, For Real

* How our 4th AI “BDR” helps close deals 24x7, and is much better than a human at it

* How our new SaaStr.ai Startup Valuation Calculator processed 300,000+ AI startup valuations in less than 30 days. And how disruptive (and just plain cool) the new Replit v3 is.

At SaaStr, we’ve gone from having essentially zero AI agents at the start of 2025 to now having over 12 AI agents in production. We have AI SDRs and BDRs handling both inbound and outbound, a digital assistant answering 150,000+ chats on our websites, and we’ve built AI tools that have processed over 300,000 startup valuations and graded nearly 1,000 VC pitch decks - all in less than 30 days.

I think we’ve learned something here about how to effectively build, deploy, and scale AI agents in a real B2B business. And more importantly, what breaks, what doesn’t scale, and where the real ROI lives.

Here’s what we learned across three major areas: the future of AI development platforms (Replit V3), why you absolutely need AI BDRs now, and how we built tools that generated real revenue without a development team.

Replit V3: The First Glimpse of AI Managing AI

The Big Question: Can you build real B2B applications without a team of developers? The answer is increasingly yes, but it’s more complicated than the marketing suggests.

I started this journey a couple months ago looking at the leaders: Replit, Lovable, Bolt, and hot newcomers like Wix Studio and Base 44 (which may have 10% market share already). Many are doing nine figures in revenue very quickly.

Why I Chose Replit: When I asked Twitter which platform to pick, everyone said they were similar — at time, a few months back. But Replit was the only one where you could go end-to-end - build, test, prototype, and push to production without configuring databases, moving to different hosting, or dealing with infrastructure headaches. The white-labeled Neon database made it plug-and-play with other tools.

Agents Managing Agents: It’s Already Here

But here’s what blew my mind with V3: Replit now has agents that manage other agents.

Most of us are struggling to get one AI SDR working, let alone having AI manage 20 agents. But Replit can do it today. When I hit a complex problem building our pitch deck grader, Replit autonomously brought in:

* An architect for really tough problems

* Specialists for specific issues

* Senior and junior agents with different capabilities

I watched these agents debate each other (in English) for almost 3 hours while conducting a security audit. They went through every line of code, every function, every page, debating how much to lock down our SaaStr AI app for security.

The Reality Check: When Even The Best AI Perhaps Is Too Powerful

The agents did incredible work autonomously, but in some ways, they did go too far. By the time they finished, the app was so locked down for security that almost nothing worked. You couldn’t upload PDFs, couldn’t upload anything, analytics were blocked - it was all blocked as security risks.

The result: I spent over 10 hours undoing the security audit, retesting every page and link. Everything interactive had been locked down to be “super secure” but became in many cases non-functional.

The Learning: This level of autonomous capability is incredibly cool, but it’s different. Many Replit users revolted when V3 launched because:

* Costs went up (smarter AI uses more tokens)

* It’s slower (smarter AI takes more time)

* No option to stay on V2

* Required learning new workflows and autonomy settings

The Business Process Change Problem: We’re moving at AI speed because we have to - competition is fierce. But as we go mainstream, we’re colliding with the reality that most users only want so much business process change. V3 is incredibly powerful, and it’s so cool. But if V2 was good enough, many users didn’t want to invest the learning curve.

This is a pattern we’ll see across all AI tools as they rapidly evolve.

Why You Need AI BDRs Now (And How to Train Them Right)

We now have four AI agents handling different parts of our revenue funnel:

Three AI SDRs (using Artisan):

* Outbound to potential sponsors and partners for events

* Outbound to prospects for event attendance

* Recirculation to previous attendees

One AI BDR (using Qualified):

* Manages inbound on SaaStrAnnual.com and SaaStrLondon.com

* Integrates with all our Salesforce and Marketo data from the past decade

* Qualifies prospects without them feeling like they’re being qualified

* Sets up meetings in real-time with our sales team

What Makes This AI BDR Different

Go to SaaStrAnnual.com or SaaStrLondon.com and try the chat. Superficially, it looks like every other chat bubble (Intercom, Finn, Drift - they all look the same). But here’s what’s clever:

It knows everything about you already. If you’ve sponsored before or been to a SaaStr event over the last decade (50,000+ people have), we already know about you. If you’re a potential sponsor, it’ll tell you about similar companies that sponsored last year and their results. If another AI SDR company visits, it’ll mention how Artisan got $500,000 of pipeline from last year’s event.

It does three things that are game-changing for tiny teams:

* Qualifies prospects better than humans - without the soul-crushing experience of talking to some junior person who knows nothing and is just deciding if you’re worth an AE’s time

* Solves problems and answers questions using a decade of integrated data

* Automatically sets up meetings with the right sales rep in real-time

It can even create new opportunities. Someone reached out wanting to bring 30 CEOs to SaaStr AI London. The AI created a whole package and suggested the idea to us.

The Training Reality: It’s Not What You Think

You absolutely must train these agents relentlessly. But here’s what we learned that most people get wrong:

You don’t need as much data as you think upfront. Our first agent was trained on 20 million words of SaaStr content. But when the CEO of Deli compared our agent to Brian Halligan’s, Lenny Rachitsky’s, and Keith Rabois’s agents (all on the same platform), we learned that massive historical content isn’t that valuable.

What you actually need:

* A solid base of good content (your website, sales pitch, collateral, reference customers)

* Relentless daily iteration and AB testing

* Watching every interaction in real-time

* Proofing and correcting anything wrong

* Constantly adding improvements

Set and forget does not work. You have to relentlessly iterate on these AI platforms.

Two Examples of Terrible AI “Non-Training” (And Why You Can Do Better)

Example 1: Brex’s AI Support Disaster

Brex kept sending alarming emails to us demanding we “immediately deposit $5 million at Brex” to maintain credit limits. When I tried their AI support (using Decagon), it wouldn’t even let me chat with AI or humans - it just closed the window.

The Fix: This needed zero complex training. Just one human sampling issues and uploading content explaining: “That email sounds alarming, but you don’t need to deposit $5 million. Here’s how Brex credit actually works.” Thousands of people got this email. No excuse for not training the AI.

Example 2: Generic Zendesk AI Support

When I had a specific infrastructure issue, Replit’s Zendesk AI kept asking for screenshots and giving generic answers. It wasn’t trained on the exact issue and couldn’t solve something that would take minutes with proper training. It wasn’t that hard to fix — a great human actually help me fix it in minutes. But the AI was utterly lost, and kept asking me basic questions again and again. All day long.

The Lesson: Unlike humans, AI agents won’t keep making the same mistakes once properly trained. But you have to do the work upfront and continuously.

Building Revenue-Generating Tools: 300,000 Valuations and 1,000 Pitch Decks

At SaaStr.ai, we built two tools that generated real user engagement:

AI Startup Valuation Calculator: 300,000+ free valuations in 30 days

VC Pitch Deck Analyzer: Almost 1,000+ founders uploaded decks in 10 days

Both are free, fast, simple, and don’t require registration. But here’s the key learning for anyone building complex AI applications:

The Secret: Offload the Complexity

My first vibe coded app failed because I tried to build matching algorithms for 10,000 SaaStr attendees (10,000 x 10,000 potential matches) in real-time within Replit. Way too complex. But my next 7 got into production.

For the valuation calculator:

* I took thousands of funding rounds and data points

* Distilled them in Claude into four simple tables instead of asking Replit to crunch all the data each time

* Simplified 5,000+ funding rounds into manageable datasets

* Let OpenAI API handle the calculations

For the pitch deck analyzer:

* Combined learnings from 800+ VCs who spoke at SaaStr Annual

* Simplified complex analysis into a manageable prompt for Anthropic API

* Ran 100+ test decks (including recent funded SaaStr Fund companies) to tune the weights

* Let Anthropic do 90% of the heavy lifting

The Pattern: Have AI do as much of the complex work as possible, but pre-process and simplify the inputs.

The Two-Domain Problem: A Glimpse of What’s Coming

We now run SaaStr.com (WordPress, 35 million SEO views/year) and SaaStr.ai (Replit-built, dynamic, modern) simultaneously. Why?

The Dilemma: Do you want state-of-the-art dynamic platforms that ca.n do things WordPress can’t, or do you want proven SEO and stability?

Moving off SaaStr.com’s domain, static URLs, and content to a new platform is super risky. The new SaaStr.ai is shiny and cool, but doesn’t have the same SEO benefits.

The Reality: We’re probably not the only ones dealing with this. Many companies will face this choice in coming years. Exporting AI code to run in WordPress isn’t elegant - it’s slow, clunky, and you lose the ability to rapidly iterate and add features.

We’re at the bleeding edge here, with about 30,000 monthly users on SaaStr.ai (10-20% of our traffic), but it’s a preview of infrastructure decisions every B2B company will face.

The Bottom Line: What Actually Works

For AI Development:

* For non-developers, choose platforms that let you go end-to-end without infrastructure headaches

* Agents managing agents is here today and super powerful, but prepare for cleanup time

* Business process change resistance is real - factor it into rollouts

For AI Revenue Operations:

* You need AI BDRs now - your competitors will have them soon

* Training is everything, but you need less data than you think upfront

* Daily iteration and monitoring is non-negotiable

* The experience should feel human, not like talking to a bot

For Building AI Tools:

* Offload complexity to AI APIs rather than building everything in your platform

* Pre-process and simplify inputs rather than asking AI to handle raw complexity

* Focus on solving real problems people will pay for or engage with repeatedly

The Meta-Learning: We now have more AIs than humans at SaaStr. The technology works, but success comes down to relentless training, iteration, and understanding that AI agents are tools that amplify good processes - they don’t fix bad ones.

Try our tools at SaaStr.ai (hit “AI VC” at the top), and let me know what you build with your own AI agents.

What AI agents are you putting into production? What’s working and what’s breaking? Share your experiences in the comments.

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

Harry, Rory and Jason are back!

We’re witnessing an unprecedented capital concentration in AI with NVIDIA’s $100B OpenAI investment creating a fascinating circular money machine, while new H-1B visa fees threaten startup talent acquisition and the venture funding landscape shifts dramatically toward mega-rounds for a tiny number of companies. The era of “founder friendly” has become somewhat hollow rhetoric, and traditional B2B growth metrics like “triple triple double double” are becoming irrelevant as the market polarizes between AI unicorns and fundamentals-driven businesses.

Key Numbers That Matter:

* 75% of 2025 VC dollars went to just 19 companies

* NVIDIA’s $4.5T market cap relies on only 6 customers for 83% of revenue

* New H-1B visa fees of $100K will impact 440,000 annual applications

* Navan filing for IPO at $8B valuation with $613M revenue, 32% growth

The $100B AI Money Machine: When Six Customers Drive a $4.5T Market Cap

NVIDIA’s massive investment in OpenAI represents more than just capital deployment—it’s the creation of what could be an infinite money printing loop. OpenAI commits $300B to Oracle, Oracle buys NVIDIA chips, and NVIDIA invests back into OpenAI. As one observer noted: “Sam’s gonna get to make the bet he wants to make which is apply infinite amount of capital and see how long these scaling laws last.”

The most striking aspect? NVIDIA, now the world’s largest company by market cap at $4.5 trillion, has only six meaningful customers accounting for 83% of revenue. Compare this to Apple’s 2 billion customers or Microsoft’s hundreds of thousands of enterprise clients. It’s “this really weird dynamic where you’ve got this company with only six customers, but the good news is all six of them are determined to spend themselves into oblivion to win the prize.”

The Scaling Laws Gamble

Sam Altman’s recent comments suggest this is just the beginning: “We need three orders of magnitude more compute than this.” The market is essentially allowing OpenAI to test whether massive capital can break through current AI limitations. Whether the marginal $300 billion will earn a return on capital remains questionable, but as Rory put it: “We will find out because no one’s going to call timeout along the way.”

The H-1B Shock: $100K Fee Creates New Startup Reality

The new $100,000 fee for H-1B visas represents a significant shift for the startup ecosystem. With 440,000 applications generating $19-120 billion in GDP annually, this policy change will have material impact on early-stage companies.

“Anyone that has been doing this for a while that isn’t just three kids working 24/7 in SF has had H-1B folks on their team,” noted one investor. “My first startup wouldn’t have been possible without H-1B. I had two on my first team of 10.” notes Jason.

While larger tech companies will simply absorb the cost, startups face a more complex reality. The workaround? Most founders are now pursuing O-1 visas, though these come with their own complications and ongoing maintenance requirements.

Venture Capital’s Great Concentration: 75% of Dollars to 19 Companies

The venture landscape has fundamentally shifted. In 2025, 75% of VC dollars went to just 19 companies—a stunning concentration that reflects the bifurcation between traditional venture and ultra-late-stage private public investing.

“What’s really happened is on top of that business has emerged this totally separate business called ultra late stage,” explains Harry. “It just means there’s another business that you can choose to be in or not that exists one or two orders of magnitude above you in the valuation world.”

The Death of “Triple Triple Double Double”?

Traditional SaaS growth metrics are becoming obsolete for many companies. While “triple triple double double” (3x-3x-2x-2x growth pattern) remains relevant at early stages, it’s insufficient for late-stage rounds in today’s market.

“There’s only so many folks growing beyond triple triple double double at 50 to 100 million—VCs will take the meeting,” notes Jason. But for companies in traditionally unloved verticals or without AI narratives, even strong growth isn’t guaranteeing funding notes Harry.

The Founder Friendly Facade

The concept of “founder friendly” has become meaningless in today’s hyper-competitive environment. “Founder friendly has become b******t,” argue both Harry and Jason. “Any hot AI deal, there is no diligence provided, nor is any done. It’s just done on Saturday.”

Real founder-friendly behavior shows up in difficult times: “Founder friendly is writing the check when no one else does. That’s founder friendly. Founder friendly is when no one else is there at the board meeting anymore and you’re there and you still have a W on the other side of it.”

Market Concentration Creates New Dynamics

The concentration extends beyond just funding to data labeling and infrastructure. Multiple data labeling companies report that the same two AI giants make up 55% of their revenue across all providers. This concentration creates a unique dynamic where “none of those six customers is trying to optimize their cost basis. They’re just trying to build as fast as they can.”

IPO Strategy in a Concentrated World

Companies like Navan (filing at $8B with $613M revenue, 32% growth) are timing their IPOs strategically. Rather than waiting for profitability, they’re moving quickly to avoid being the “third player” in their category after competitors like Brex and Ramp potentially go public.

“When people perceive it as a direct comp, even if it isn’t, it becomes troubling to get out because every public investor says, ‘I already have Brex and Ramp. Why would I buy you?’” notes Rory.

The New Venture Reality Check

For investors, the current environment requires recalibrating expectations. Jason admits: “I’m 0% cash, just like 2008. I remember feeling I could literally not have enough cash to fix the roof on my house—nothing was more fun than selling my stock at a 70% loss to fix that roof.”

The frothiness is evident: LPs are now bragging about returns on LinkedIn, historically a sign of market peaks. Yet for companies with genuine traction, funding remains available. The key is having clarity on realistic next-round pricing rather than extrapolating long-term returns.

Portfolio Concentration Lessons

Success breeds more risk tolerance. “Having an early success is highly correlated with future success. Partly you get the referral effect, but partly I think it’s that you just have the stomach to roll the dice and you get braver.” notes Rory.

This explains why firms like Sequoia can hold positions like Klarna while emerging managers face different portfolio pressures around concentration limits.

Looking Ahead: The End of 2021 Valuations

There’s growing consensus that 2021 valuations should be written off entirely. “I think we can’t talk about 2021 valuations anymore. It’s time to just flush them down the toilet,” suggests Jason proposing a deadline: “You’ve got 90 days left to caveat and complain about your 2021 valuations. And on January 1, 2026, no one is allowed to talk about their 2021 valuations.”

Companies like Notion, now at $500M ARR with re–\acceleration, show that B2B companies can recover by embracing AI while maintaining their core business. The path forward isn’t becoming “something totally different” but rather intelligent integration of AI capabilities.

The Monopoly Question

OpenAI’s consumer dominance raises antitrust questions, but it’s complicated by their massive subsidization of users. “Never have there been a less successful monopoly than ChatGPT because they’re subsidizing the consumer surplus delivered by ChatGPT in the tens of billions of dollars” notes Rory.

The real monopoly question may be more relevant to NVIDIA, which has true market share dominance in AI chips, making their equity investment in OpenAI particularly strategic.

Key Takeaways

* Capital concentration is unprecedented: 75% of 2025 VC dollars went to just 19 companies, creating a new tier of ultra-late-stage private investing distinct from traditional venture

* H-1B changes will impact startups materially: The $100K fee creates real barriers for early-stage companies, though workarounds and for founders use of O-1 visas are and will emerge

* “Founder friendly” has become meaningless: In hot AI deals, there’s no diligence and term sheets are signed on Saturdays, making the concept hollow

* Traditional SaaS metrics are obsolete: “Triple triple double double” no longer guarantees late-stage funding except for the hottest AI companies

* Timing matters for IPOs: Companies are racing to go public before competitors in their category, as being third creates comparison problems

* 2021 valuations need to be forgotten: It’s time to write down frothy valuations and move forward with fundamental-based pricing

Quotable Moments

“This is an epic monopoly like we’ve never seen. Think how much ChatGPT already dominates our lives. It’s the Standard Oil of tech.” — Jason discussing OpenAI’s market dominance

“You’ve got this company with $4.5 trillion market cap with only six customers. That’s bad news. But the good news is all six of them are determined to spend themselves into oblivion to win the prize.” — Rory discussing NVIDIA’s customer concentration risk

“Any hot AI deal, there is almost no diligence provided, nor is any done. It’s just done on Saturday. All you can lose is one extra money.” — Jason Lemkin on current venture funding practices

“Having an early success is highly correlated with future success. Partly you get the referral effect, but partly I think it’s that you just have the stomach to roll the dice and you get braver.” — Rory on venture capital success patterns

“I’m 0% cash, just like 2008. Nothing was more fun than selling my stock at a 70% loss to fix the roof on my house.” — Jason reflecting on his current market position

“I’m seeing less diligence now than I did in 2021.” — Harry discussing the current state of venture due diligence

“The hardest thing I find is honestly we are getting it wrong. A lot of our startups at 20VC Fund now are are growing nicely and I cannot predict what the next round wants because it seems to be moving so much and I’m having real problem predicting financing markets.” — Harry on the unpredictability of current funding markets

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

How to scale from startup to $300,000,000+ ARR by mastering the fundamentals of go-to-market strategy

Building a billion-dollar B2B business isn’t about finding secret hacks or silver bullets. They don’t last or scale. It’s about mastering fundamental principles and being willing to reinvent yourself every six months to a year as you scale. Gaurav Agarwal, COO of ClickUp came to SaaStr Annual + AI Summit to share how they did it — and keep doing it.

As someone responsible for “all things money” at ClickUp – sales, marketing, growth, pre-sales, and post-sales – Gaurav has lived through the reality that what gets you to $1M ARR is completely different from what gets you to $10M, $50M, $100M, $300M and beyond. Nothing scales infinitely, and every stage requires its own playbook.

Here are the key principles that have driven ClickUp’s remarkable growth:

1. Know Where You Win: The LTV vs. TAM Matrix

Most companies fail because they try to adapt everyone else’s strategies without understanding their own fundamental positioning. Before you copy anyone’s playbook, you need to map your business on a simple 2×2 matrix:

* X-axis: Customer Lifetime Value (LTV) – How much can you make from your customers?

* Y-axis: Total Addressable Market (TAM) – How many customers are out there?

This creates four distinct quadrants, each requiring completely different go-to-market strategies:

High LTV, Small TAM: Whale Hunting You’re selling to Fortune 500 companies with limited prospects. Your channels must be high-touch: field marketing, trade shows, conferences, and business development. You can afford expensive customer acquisition because deal sizes justify the investment.

Low LTV, Large TAM: Cast a Wide Net You can’t afford expensive acquisition channels. Focus on organic growth: content marketing, SEO, social strategies, and community building. You need LTV-to-CAC positive channels that scale efficiently.

Low LTV, Small TAM: Exit Strategy If you have few customers who don’t pay much, you shouldn’t be in this business. Run away and find a better opportunity.

High LTV, Large TAM: The Sweet Spot This is where ClickUp operates, and it’s the most exciting quadrant. You can make almost any channel work – enterprise sales teams, billboards, TV ads, digital marketing. The world is your oyster for distribution strategies.

2. Learn From the Best-in-Class Across Industries

Don’t limit yourself to studying companies in your vertical. The best growth strategies often come from unexpected sources:

* For SEO: Study HubSpot, but also look at NerdWallet, Canva, and Zapier

* For brand building: Don’t just look at B2B companies – examine what Liquid Death and Beats by Dr. Dre accomplished

* For viral growth: Understand how consumer companies create shareability

At ClickUp, teams obsess over these best-in-class companies and adapt their learnings regardless of industry. A B2B company can absolutely learn from B2C growth tactics, because all customers are ultimately humans with similar psychological triggers.

3. Avoid Zero-Sum Thinking

The biggest mistake scaling companies make is creating false either/or narratives:

* Product-led growth vs. sales-led growth – Why not both?

* Brand building vs. demand generation – They should work together

* B2B vs. B2C tactics – Customers are humans regardless of context

ClickUp runs a dual-engine growth model that proves these approaches can be complementary:

Product-Led Growth Engine: Focuses on users – acquire, activate, monetize, and expand. This gives ClickUp distribution at scale and catches a wide net of prospects.

Sales-Led Growth Engine: Focuses on accounts where PLG has already landed users. The goal is reaching out to existing customers to drive expansion and deeper penetration.

The results speak for themselves: when a customer moves from self-serve/PLG to sales-assisted, ClickUp sees an 11x lift in LTV. Product-led growth provides the distribution; sales-led growth maximizes the lifetime value.

Most companies kill one motion in favor of the other, missing the massive opportunity of letting them feed into each other.

4. Master Your Growth Portfolio

As you scale, you need to think like a portfolio manager. Your growth strategy should diversify across multiple bets – channels, segments, products, geographies – each with different risk/reward profiles.

The 70-20-10 Resource Allocation Framework:

* 70% on proven channels with high probability of success

* 20% on small bets that deliver incremental gains

* 10% on big bets like viral content or breakthrough strategies that could create inflection points

This approach delivers predictable growth with asymmetric upside. The 70% gives you financial predictability, while the 20% and 10% create opportunities for breakthrough growth moments.

Key Portfolio Principles:

Understand Risk vs. Reward: Higher risk channels might extend payback from 20 to 50 months, while optimizing for quick payback could limit top-line growth. Find the efficient frontier.

Not All Channels Are Equally Repeatable: ClickUp can predict which SDR will hit their number and the ROI of Google branded search, but can’t predict which social media clip will get 50 million views. Plan accordingly.

Channels Have Capacity Limits: After hiring 100 account executives, productivity per rep declines. Most paid channels hit diminishing returns, but some organic channels compound infinitely.

5. Identify Compounding vs. Diminishing Channels

Understanding which channels scale versus which hit walls is crucial for long-term success:

Diminishing Return Channels:

* Paid advertising (Google, Facebook, etc.)

* Outbound sales beyond optimal team size

* Most performance marketing channels

Compounding Channels:

* Community: Takes time to build but becomes self-sustaining

* SEO/Content: High domain authority compounds with each new piece

* Viral Product Features: Like Slack or Dropbox, usage drives sharing

* Network Effects: Value increases with each new user

While you can’t scale Google paid search infinitely, you can scale Google organic content indefinitely with the right content engine.

6. Never Assume You’ve “Made It”

At $10M ARR, getting to $15M feels relatively easy. At $500M ARR, growing 50% to $750M is “extremely, extremely hard.” The only way to beat this challenge is maintaining hunger and assuming there are always better opportunities you haven’t discovered yet.

Combat the Local Maximum Trap: Teams often get stuck optimizing their current approach instead of finding breakthrough strategies. Always assume better distribution tactics and motions exist.

Apply Constant Pressure: Challenge yourself and your team to keep uncovering new ground. The moment you get comfortable is when competitors start gaining ground.

7. Leverage AI for 10x Velocity Gains

If you view revenue generation as a complex machine, AI creates opportunities to optimize efficiency at multiple steps simultaneously. Companies not using AI to accelerate their go-to-market motion will be left behind.

ClickUp’s AI-First Approach Across 80% of Revenue Functions:

* AI SDR: Automatically reaches out to form submissions, generated $1M in pipeline in the first month

* Content at Scale: Produces 40,000 pages monthly using AI, localized in 15 languages, resulting in 10x search impression volume

* Customer Support: AI chat agents resolve issues faster, freeing human agents for proactive account management

The key is systematically identifying every step in your revenue machine and asking: “How can AI increase throughput here?”

The Top Scaling Mistakes to Avoid (According to Gaurav)

Based on his experience scaling ClickUp to nine-figure ARR, Gaurav identified the most common and costly mistakes he sees companies make:

1. Misunderstanding Your LTV/TAM Position and Copying Wrong Strategies

The Mistake: Companies blindly copy successful strategies from other businesses without understanding whether they operate in the same LTV/TAM quadrant. A high-LTV, small-TAM business trying to use Canva’s mass-market SEO strategy will waste massive resources.

The Fix: Map your business honestly on the LTV/TAM matrix first, then only study companies in your quadrant. A whale hunting business should learn from enterprise software companies, not consumer apps.

2. Creating False Either/Or Choices

The Mistake: Forcing zero-sum decisions between complementary strategies. “We’re PLG, so we can’t do sales-led.” “We’re B2B, so we can’t use B2C tactics.” “We focus on demand gen, not brand building.”

The Fix: Look for ways strategies can compound each other. ClickUp’s 11x LTV lift from combining PLG and sales-led growth proves these approaches can be synergistic, not competitive.

3. Getting Stuck at Local Maximum and Stopping Innovation

The Mistake: Once companies find what works, they optimize that single approach to death instead of continuing to explore new opportunities. They become comfortable and stop pressure-testing for better distribution tactics.

The Fix: Always assume there are better strategies you haven’t discovered. Maintain the 70-20-10 resource allocation where 30% of resources go toward new opportunities, even when current channels are performing well.

4. Abandoning Compounding Channels Too Early

The Mistake: Companies give up on long-term compounding channels (community, SEO, viral features) because they don’t see immediate results, instead over-investing in diminishing return channels like paid ads because they’re more predictable.

The Fix: Understand which channels have infinite scaling potential versus those with natural limits. Invest in building compounding channels even when they take longer to show results, while using diminishing return channels for predictable short-term growth.

The Bottom Line: No Hacks, Just Fundamentals

There are no secret growth hacks for building a sustainable, ever-growing revenue machine. Success comes from:

* Understanding your positioning in the LTV/TAM matrix and choosing appropriate strategies

* Learning from the best-in-class regardless of industry

* Avoiding zero-sum thinking and finding ways to combine approaches

* Building a diversified growth portfolio with the right risk/reward balance

* Investing in compounding channels while managing diminishing return channels

* Maintaining hunger and constantly seeking better opportunities

* Using AI to 10x velocity across every step of your revenue machine

The competition is getting fiercer every day. ClickUp is now accelerating at $300,000,000+ ARR. How about you?

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

The technical and product insights from Dara Ladjevardian’s AI cloning experiment at SaaStr Annual + AI Summit.

The Clone Performance Reality Check

When Dara Ladjevardian, CEO of Delphi.ai, ran 1 million simulated conversations with digital versions of Brian Halligan (Chairman and founding CEO HubSpot), Lenny Rachitsky, Keith Rabois and Jason Lemkin, the most interesting findings weren’t about the business advice the clones gave—they were about how AI clones actually behave, fail, and succeed.

And you can try them all yourself here:

* Digital Jason, try it here

* Digital Lenny, try it here

* Digital Brian, try it here

* Digital Keith, try it here

Key Technical Learnings

1. Context Dimensions Drive Dramatically Different Outputs

The discovery: The same clone gives fundamentally different advice based on just four input variables:

* Company stage (0-1M, 1-10M, 10-100M ARR)

* Market context (emerging vs. established, crowded vs. uncrowded)

* Team dynamics (solo vs. co-founder, data-driven vs. visionary)

* AI adoption position

What this means technically: Current LLM approaches that treat context as simple “system prompts” miss the nuanced way human experts actually adjust their thinking. The clones needed sophisticated context weighting to perform authentically.

The failure mode: Without proper context handling, AI clones default to generic advice that sounds like the person but lacks their actual decision-making sophistication.

2. Temporal Knowledge Graphs Beat Static Training

Dara’s architecture insight: “The best way to represent a network of ideas that changes over time is a temporal knowledge graph.”

Why this matters: A static knowledge graph might show Keith Rabois believed X in 2015, but his 2024 graph shows he believes Y. The temporal system tracks belief evolution to predict future responses.

The technical challenge: Most AI clones train on a person’s entire corpus as if their views never changed. This creates internally inconsistent outputs that feel “off” to people who know the subject well.

Real-world impact: Dara’s grandfather’s clone could apply 1970s Iranian business principles to 2024 AI startup decisions—something impossible with static training.

3. Model-Agnostic Architecture Outperforms Single-Model Training

Strategic decision: Delphi doesn’t train custom models—they use multiple existing models with sophisticated mind mapping on top.

The reasoning: “We could train our own model right now, but why spend all that money? The product works really well by mapping out the mind and leveraging multiple models.”

Performance insight: The hard problem isn’t the LLM—it’s accurately representing someone’s decision-making patterns and worldview. Once you solve that, you can ride the commodity curve of improving foundation models.

4. Two-Mode Architecture Solves the Accuracy vs. Utility Trade-off

Static Mode: Only answers questions the person has explicitly answered before. Higher accuracy, limited utility.

Adaptive Mode: Can predict responses to novel situations based on learned patterns. Higher utility, requires stronger guardrails.

The professional insight: For doctors and lawyers, wrong answers create lawsuit risk. For creators and advisors, novel insights create value even if occasionally wrong.

Product learning: Users need explicit control over this trade-off rather than a single “accuracy” dial.

5. Identity Verification is Critical for Trust (And Scaling Pain)

Current process: Every user submits photo holding ID. Dara manually verifies each one.

The scaling problem: Manual verification obviously doesn’t scale, but automated systems miss edge cases that matter for trust.

Why it matters: Creating clones of others without permission isn’t just unethical—it destroys platform credibility when discovered.

The unsolved challenge: How to verify identity at scale while preventing abuse and maintaining trust.

6. Guardrails Need Domain-Specific Tuning

The discovery: Generic content filters don’t work for professional AI clones. A doctor clone needs different guardrails than a business advisor clone.

Technical challenge: Building “pretty strict guardrails” requires understanding not just what the person would say, but what they’re legally/ethically allowed to say in their professional capacity.

Performance impact: Over-aggressive guardrails make clones feel robotic. Under-aggressive ones create liability risks.

What Actually Works in Clone Architecture

The Ray Kurzweil Method

Based on “How to Create a Mind” (2014): The brain is “a hierarchy of pattern recognizers.” Since LLMs are pattern recognizers, you can recreate minds by mapping patterns correctly.

Focus on Representation, Not Training

The breakthrough insight: Spend engineering effort on accurately modeling someone’s thinking patterns, not on training custom language models.

Multi-Model Redundancy

Use multiple foundation models rather than relying on a single custom-trained model. The mind representation layer handles consistency.

The Unsolved Problems

Belief Evolution Tracking

How do you automatically detect when someone’s views have changed on a topic without manual annotation?

Context Sensitivity at Scale

How do you handle the exponential complexity of context combinations as you add more dimensions?

Authenticity Validation

How do you measure whether a clone “sounds like” the original person beyond basic accuracy metrics?

Dynamic Guardrail Adjustment

How do you automatically tune safety constraints based on the professional context and risk tolerance of different use cases?

Additional Learnings from the Clones:

* Speed vs. Accuracy Trade-off: The clones performed best when optimized for speed of response rather than perfect accuracy. Users preferred fast, “good enough” responses over slow, perfect ones.

* Personality Quirks Matter More Than Content: The clones that felt most authentic weren’t the ones with the most training data, but those that captured subtle personality traits—Keith’s contrarianism, Jason’s directness, Brian’s optimism.

* Context Switching is Expensive: When conversations jumped between topics (stage advice to pricing strategy to hiring), clone performance dropped significantly. Single-topic conversations maintained much higher quality.

* The “Not Always Right” Disclaimer Actually Improves Trust: Dara noted that clones explicitly saying “I’m not always right” made users more comfortable engaging deeply, similar to how GitHub Copilot’s 35% acceptance rate doesn’t hurt adoption.

* Users Test Clones Immediately: The first questions people ask are always attempts to “break” or test the clone with edge cases, not genuine advice-seeking. The clones needed to handle skeptical users before helpful ones.

* Memory Across Conversations Doesn’t Scale: While individual conversations stayed coherent, maintaining context across multiple sessions created exponential complexity that hurt performance more than it helped.

The biggest technical insight: Building authentic AI clones isn’t about better language models—it’s about better models of how humans actually think, decide, and evolve their beliefs over time.

Thanks for reading The Secrets To Scaling in The Age of AI! Subscribe for free to receive new posts and support my work.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com

We're back on 20VC + SasStr with Harry Stebbings, Jason Lemkin, Rory O'Driscoll, and special guest Jeff Lawson (Founder & Former CEO, Twilio)

Bottom Line Up Front

Rory O'Driscoll: Tesla's trillion-dollar pay package for Elon is a board betting everything on doubling down - they believe without him, the stock drops 75% overnight. It's intellectually coherent but terrifying risk concentration.

Jason Lemkin: We're in the greatest wealth hunt in venture history. Orders of magnitude larger deals are the new normal. A $10 billion company feels "niche" today when we're discussing $100+ billion valuations.

Jeff Lawson: The AI wave creates unprecedented opportunities for infrastructure companies like Twilio that aren't selling seats - no innovator's dilemma. SaaS companies selling seats face existential disruption as AI eliminates 75% of human roles.

Harry Stebbings: Late-stage AI investing has become the rational play for VCs - when only valuation risk remains, even $100M checks into $13B rounds make mathematical sense for portfolio construction.

The Trillion-Dollar Elon Bet: Rational or Reckless?

Tesla's board just approved what could become the first trillion-dollar executive compensation package in history. But is this visionary leadership investment or a high-stakes gamble gone wrong?

The Board's Logic: Double or Nothing

Rory O'Driscoll dove deep into Tesla's 332-page proxy filing to decode the board's thinking. "Compensation is how boards reveal their real priorities," he explained. "Nothing else matters as much. The board wants the Elon bet - they believe they owe him the past and they're betting on him for the future."

The package's operational metrics tell the story: Tesla needs to hit $400 billion in EBITDA (four times Google's current profitability), manufacture 20 million cars, deploy 10 million Full Self-Driving systems, and produce 1 million Optimus robots. It's essentially asking Elon to double the existing business while building entirely new categories.

The Downside Protection Theory

Jeff Lawson raised the critical counterpoint: "Maybe this isn't about upside - it's about the downside case. Tesla is overvalued as a car company. If valued purely on automotive fundamentals, it's worth 25% of current market cap. The other 75% is Elon's special sauce."

This creates a prisoner's dilemma for the board. As Rory noted, "If you try to demonstrate resolve and he threatens to walk, you're down 75% next morning. The individual shareholders who voted for this compensation twice want to make this bet, even though it makes my head hurt."

The New Benchmark for Founder Compensation

Jason Lemkin sees this setting a new standard: "This is the new normal for anyone whose board consists of their brother-in-law and other relatives. Almost all my portfolio companies - the founders control the board, not just from a cap table perspective, but from a relationship perspective."

The trend is clear: instead of 2-3% top-ups after years of struggle, founders are now getting 7-8% packages tied to massive outcome targets of $10-100 billion valuations.

Ramp Hits $1B ARR, Brex at $700M: AI Tide Lifting All Boats?

Two major fintech announcements dominated the week: Ramp crossing $1 billion ARR and Brex hitting $700 million ARR with 50% growth. But are these isolated successes or signs of broader market strength?

The AI Money Flow Effect

Jason Lemkin argues it's the latter: "The AI boom is filtering further down the stack. We're seeing it in Broadcom, Cisco - where our grandpa learned to be an engineer. If you're a B2B company seeing nothing from AI, you get an F. There's so much money flowing through this system."

The fish food metaphor resonated: "It's floating down to where it's dark in the ocean now. It's embarrassing if you can't get any of it."

Financial Services vs. Software Valuations

Rory O'Driscoll provided clarity on how to value these hybrid companies: "They have the margin profile and core dynamics of financial services but the growth rate of software companies. Once growth slows, they'll be valued just like AmEx. The growth is what's saving them."

Jeff Lawson offered the infrastructure perspective: "There's money out there, and it's got to go into some bank. Are they winning market share from legacy companies? Part of it. But during the mobile boom at Twilio, we had customers spending millions who didn't make it - that revenue went away and had to be replaced."

Sierra's $10B Valuation: Bubble Territory or Brilliant Bet?

Brett Taylor's Sierra commanding a $10 billion valuation at $100 million ARR (100x multiple) sparked intense debate about AI valuations and whether we're seeing rational investing or complete market detachment.

The Generational Talent Premium

Jason Lemkin made the bull case: "If Scale was for sale for $28 billion, Brett's got to be worth $56 billion. This is a generational guy who turned down being co-CEO of Salesforce to do this. You get everything - the ex-CTO of Salesforce and Facebook, his team, and a potential category leader."

Rory O'Driscoll walked through the investment logic: "Is there a category that can support a big winner? Customer support is the number one AI use case outside coding. Are they going to be a winner? Clearly one of a small number with a great position. The only question left is: are you getting paid for the risk?"

The Portfolio Math

Harry Stebbings questioned the concentration: "Neil Mather is probably putting 10% of his fund into this next check. That's notable given the percent of his fund this represents."

Jeff Lawson provided founder perspective: "When I start a company, that's 100% of my capital allocation - for my life, time, bank account, everything. Don't say you're brave putting 10% of the fund into one deal when the founder is putting 100% with no way out."

Customer Support: The Perfect AI Use Case

The consensus emerged around customer support being ideal for AI disruption. As Rory explained: "You have lots of people answering phone calls and emails. You can do 70-80% with AI. It saves money on a cost center. This is going to happen."

Kleiner Perkins' $100M Anthropic Bet: Logo Deal or Smart Money?

Kleiner Perkins investing $100 million into Anthropic's $13 billion round at a $183 billion valuation raised eyebrows as their first model provider investment.

The "Logo Deal" Debate

Jason Lemkin initially called it a logo deal: "You can't walk into LP meetings without Anthropic or OpenAI on the website." But deeper analysis revealed more complexity.

Rory O'Driscoll pushed back: "I don't think anyone does a $100 million logo deal. When you run the math, one of two things will happen: either the fastest slowdown in history in the next two years, or if this trajectory continues, this round will work."

The Evolution of Venture Capital

The discussion revealed how venture capital has fundamentally changed. Jason noted: "This isn't venture capital as we knew it. Venture capital today is 20% old school and 80% late-stage growth investing that would have been Fidelity."

Jeff Lawson quoted Gloria Swanson: "I didn't leave Hollywood. Hollywood left me. This is venture capital today."

Risk-Adjusted Returns in AI

Harry Stebbings made the mathematical case: "When only valuation risk remains, you're putting $100 million in and potentially getting $300 million back. That feels like an easier way to make money than being in the trenches in the seed round."

OpenAI's $10B Secondary: 1,000 New Millionaires Hit the Market

OpenAI's unprecedented $10 billion secondary sale creates the largest private liquidity event in history, potentially minting over 1,000 new millionaires in San Francisco.

Historical Context and Market Impact

Jeff Lawson recalled similar dynamics: "I remember feeling this way before Twitter's IPO. I was looking for our first house and thinking I had to buy before the IPO because everything's going to go nuts."

Rory O'Driscoll provided perspective: "If this were a public company worth half a trillion dollars, 20% held by management means $100 billion. People selling 10% to diversify is entirely rational. It's only anomalous because it's private."

The Recruiting Arms Race

Jason Lemkin highlighted the competitive implications: "The impact on recruiting is so much bigger this generation. If you're running a boring B2B company doing triple-triple-double-double, even 36 months ago you were S-tier. Today, you're not going to get engineering talent when everyone's making eight figures."

Jeff Lawson offered historical perspective: "It's like asking Jim Farley from Ford how he recruited developers during the teens when they could have worked at Twitter and Facebook. Maybe the key is getting out of Silicon Valley."

Anthropic's $1.5B Author Settlement: Setting Legal Precedent

Anthropic's $1.5 billion payment to authors for copyright infringement established important legal boundaries for AI training data.

The Legal Framework

Rory O'Driscoll analyzed the court ruling: "The judge said if you bought the book and paid $15, using it for training is legal. If you downloaded from pirate sites without paying, you get fined $3,000 per book. It's fairly coherent - 500,000 books at $3,000 each."

The precedent is clear: AI companies must purchase content for training rather than downloading from pirate sites. "This wasn't cutting corners," Jason noted. "This was pure piracy - going to Pirate Bay for books."

Future Implications

The ruling suggests a path forward: legitimate content licensing for AI training. As Rory explained, "There will be a corpus available where someone has bought 500,000 books, scanned them, creating a legally compliant dataset for training."

The Great Corporate AI Investment Wave

From ASML's $14 billion investment in Mistral to Atlassian's $610 million acquisition of Browser Company, corporate America is making aggressive AI bets.

ASML's Mysterious Mistral Investment

The semiconductor equipment giant's investment in the French AI company puzzled observers. Jeff Lawson explained the corporate cash dynamics: "When you're generating massive cash, it's orphaned on your balance sheet. You can't hire 1,000 engineers without hurting EPS. But swapping one asset for another can be essentially free if it doesn't decline."

Rory O'Driscoll suggested European sovereignty concerns: "Non-US regions feel the need for local champions. It's the AI version of defense equipment - Europeans make weapons they have no business making from an economies of scale perspective because they don't want to rely on Americans."

Atlassian's Browser Company Acquisition

The $610 million cash acquisition sparked debate about corporate desperation versus strategic vision.

Jeff Lawson was skeptical: "The thesis of needing a different browser for work didn't resonate with me. I'm not sure there are enough upsides to change behaviors."

Jason Lemkin saw it as symptom of broader urgency: "Everyone feels like they have to make a play. Sometimes when you're itchy, you might not make the ideal investment."

The SaaS Disruption Imperative

Jeff Lawson delivered perhaps the most important insight of the discussion: "Any SaaS company today is primed for disruption because AI will decimate their seat base. AI will do the jobs people are sitting there doing in Atlassian products today."

His advice was direct: "I would skate directly there and say: what is the job humans are doing in Atlassian products, and here's the AI version of that. That's what I would be doing."

The Developer API Evolution

Jeff Lawson shared his framework for successful developer companies, identifying only three categories that achieve breakaway revenue:

Business Development as a Service

"Developers can't open bank accounts or strike deals with AT&T on behalf of their company. But with Twilio, Stripe, AWS, you can engage in business relationships you weren't able to do previously."

Capex as a Service

"A developer can't spend $10 million to build a data center, but they can put it on a credit card. AWS and Google Cloud fit here."

Algorithm as a Service

"The algorithm must be so complicated that developers say 'I'm not smart enough to figure that out.' I used to only put DynamoDB here - infinite scaling databases are so hard. Now I'd put inference in that category, except it's becoming open source."

The framework explains why many developer tools struggle: "Developers take your cool thing as a challenge. You're saying they can't build what you built. Especially when it gets to meaningful revenue - if you're paying $5 million a year, developers think they can save the company money by recreating it."

The Fraud and Ethics Discussion

The arrest of IRL's CEO for fraud prompted discussion about startup ethics in the current environment.

The Enforcement Perspective

Jason Lemkin advocated for stronger consequences: "Venture rounds are getting done on Saturdays. Diligence isn't even being attempted. The best control would be if more founders who committed fraud went to jail."

He continued: "There's no consequence to standing up at a top accelerator saying you have millions in revenue when it's not there. If a few more people went to jail, it would have the proper chilling effect."

The Due Diligence Responsibility

Jeff Lawson offered a more nuanced view: "Whenever I read these fraud stories, I've never heard of the companies. If I've never heard of them as a real human being, maybe there wasn't much real behind them."

Rory O'Driscoll balanced both perspectives: "The commission of crime is on the 22-year-old who lies. But the 40-year-old running money who's sophisticated owes the system a duty of care to check this stuff and not get carried away."

Key Takeaways

On Founder Compensation

* Elon's package sets new benchmark for founder-controlled boards

* Risk concentration reaches extreme levels with $1 trillion packages

* Boards face prisoner's dilemma: pay up or risk 75% stock decline

On AI Market Dynamics

* Infrastructure companies positioned better than SaaS (no innovator's dilemma)

* Customer support emerges as perfect AI use case

* Late-stage AI investing becomes mathematically rational for VCs

On Market Valuations

* 100x ARR multiples justified for generational talent in large markets

* Geographic sovereignty concerns drive irrational investments

* Corporate cash seeking AI exposure creates acquisition opportunities

On Industry Evolution

* Venture capital fundamentally changed: 80% late-stage growth investing

* Developer API success requires business development, capex, or algorithm as service

* SaaS companies face existential threat from AI eliminating seat-based revenue

Most Quotable Moments

Rory O'Driscoll

* "Compensation is how boards reveal their real priorities. Nothing else matters as much."

* "The board wants the Elon bet. If they try to demonstrate resolve and he threatens to walk, you're down 75% next morning."

* "I didn't leave Hollywood. Hollywood left me. This is venture capital today."

Jason Lemkin

* "This is the greatest wealth hunt in venture history. A $10 billion company feels 'niche' today."

* "If you're a B2B company seeing nothing from AI, you get an F. There's so much money flowing through this system."

* "Venture rounds are getting done on Saturdays. Diligence isn't even being attempted."

Jeff Lawson

* "Don't say you're brave putting 10% of the fund into one deal when founders are putting 100% with no way out."

* "AI will decimate SaaS seat bases. AI will do the jobs people are sitting there doing in Atlassian products today."

* "You don't start companies to make money. You start companies because you think the world needs what you're building."

Harry Stebbings

* "When only valuation risk remains, valuation risk expands to fill like a vacuum."

* "We're in a weird world where you can't buy the things that are great. You can't buy Sierra."

* "If you're not in, you can't win. But oh my god, it's hard to know which bets to make."

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cloud.substack.com