AI to ROI

Token prices have fallen 98 percent since GPT-4, but enterprise AI bills are up 320 percent. In this week's Big Story episode, Ray Rike and Peter Buchanan trace where that money is actually coming from, and the answer is not the software budget.

Drawing on Gartner, Oxford Economics, Zylo, Challenger Gray and Christmas, and Goldman Sachs data, the two lay out why labor, not IT, is becoming the primary funding source for AI at scale, and why almost no company has the measurement infrastructure to manage it.

• The price paradox. Per token costs have collapsed, but usage has grown faster than costs have fallen. Ray walks through the math behind average enterprise AI budgets rising from $1.2 million to $7 million in two years.

• Three cautionary tales. Uber consumed its entire annual Claude Code budget in under four months, Microsoft revoked thousands of Claude Code licenses over cost, and one unnamed enterprise ran up a $500 million bill in a single month. Ray and Peter break down why each was a governance failure rather than a technology failure.

• Only two budget pools are big enough. The IT and software budget represents just 3 to 4 percent of revenue, while labor represents 25 to 40 percent depending on industry. Ray makes the case that labor is the only pool large enough to absorb the AI spending trajectory Gartner and Oxford Economics are projecting.

• The attrition lever. Ray and Peter unpack how not backfilling open roles has quietly become the primary way enterprises are funding AI investment, supported by data showing over 113,000 tech layoffs in 2026 with 48 percent explicitly attributed to AI.

• Revenue per FTE as the tell. Ray shares benchmark data showing SaaS company revenue per employee up 25 to 35 percent over the last twelve quarters, and explains why this metric will be the clearest signal of whether the AI budget transfer is actually working.

• Five metrics every CFO needs now. The episode closes with a practical starting list: AI spend as a percent of revenue, AI spend per employee, inference spend as a percent of opex, inference cost as a percent of COGS for AI enabled products, and revenue per FTE tracked against labor cost and agent cost as a percent of OPEX.

AI to ROI is always looking for guests with real-world examples of measuring AI budget impact. Reach out Ray on LinkedIn (@rayrike).

Subscribe to AI to ROI at ai2roi.substack.com for the full June 9th edition, and leave a review wherever you listen

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Professional services firms have billed by the hour for over 200 years. That model is now under direct attack from a new category of company: the AI Native Services firm. Ray Rike and Peter Buchanan break down exactly what makes these companies structurally different from traditional professional services firms and AI-augmented incumbents, profile four companies proving the model at scale, and lay out the six critical success factors that will separate the winners from the well-funded failures.

Episode Highlights:

• Defining the category. Drawing on the Emergence Capital AI Native Services Playbook, Ray and Peter establish a clear three-part taxonomy: AI Native Services companies (AI does 80 to 90% of the work, a licensed human reviews and is accountable for the outcome, and the client pays for results), AI Augmented Services companies (humans still do most of the work with AI as a productivity layer), and traditional SaaS tools (the customer's team does the work, and the vendor takes no accountability for the outcome). The distinction matters enormously for enterprise buyers evaluating contracts and liability.

• How the operating model actually works. The AI Native Services delivery model runs in four phases: client intake and data ingestion, AI-driven execution of the primary service work, licensed human review and approval, and outcome delivery back to the client. Critically, that fourth phase is where the model compounds, because every accepted output becomes training data that makes the system smarter and harder to displace over time.

• Four companies are proving the model. Ray and Peter profile four AI Native Services companies at different stages of scale, each dominating a regulated vertical wedge. Top AI-Native Service companies covered include: 1) Field Guide is automating audit workflows for nearly half of the top 100 US accounting firms, including KPMG and RSM, and recently raised a $75 million Series C at a $700 million valuation; 2) Even Up has built a proprietary PI AI model trained on hundreds of thousands of personal injury cases, processing 10,000 cases per week for over 2,000 law firm clients, following a $385 million funding round; 3) A-Bridge converts physician-patient conversations into structured clinical notes integrated with Epic and other EMR platforms, serving over 150 enterprise health systems including Kaiser Permanente, Mayo Clinic, and Johns Hopkins, with $100 million in ARR and a $5.3 billion valuation; 4) Harper is a licensed commercial insurance broker, not a software tool, that processes applications across 160+ carriers simultaneously and delivers final coverage in 24 to 48 hours versus the industry standard of five to seven days.

• Six critical success factors. The hosts lay out what separates durable AI Native Services companies from those that will stall: genuine domain expertise on day one, a proprietary data flywheel that compounds with every case resolved, a clear migration path from labor-based to outcome-based pricing, honest gross margin accounting that properly classifies LLM inference and human labor as cost of goods sold, narrow vertical focus on a specific wedge rather than broad horizontal expansion, and distribution through regulated industry incumbents who provide both credibility and enterprise access.

• Gross margin as the early warning signal. If gross margins are declining as an AI Native Services company scales, that is a signal the human-in-the-loop is becoming the bottleneck rather than the leverage point. The financial goal is to continuously increase gross margins as AI does more of the work, moving from a 35 to 45% gross margin profile toward 50 to 60% over time.

• What enterprise buyers should ask. Ray closes with a direct call to action for executive buyers: if an AI vendor is pricing by the seat, by the partial FTE, or by the hour, push hard on how much of that is human supervision of AI versus a truly AI-native delivery model. You are no longer contracting a resource to do the work. You are contracting an organization to deliver the outcome you need.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Token spend is exploding across the enterprise, but the value it creates remains largely invisible on corporate dashboards. In this week's Big Story episode, Ray Rike and Peter Buchanan unpack why AI investment and ROI visibility are moving in opposite directions, and what enterprises need to do about it.

Drawing on Ramp data, Exponential View, Semianalysis, and Ray's recent conversation with Russ Frayden, CEO of Lariden, the two dig into the measurement infrastructure gap that is turning individual AI productivity gains into an unmeasured expense line.

• The productivity paradox. Individual output is up across engineering, sales, and research functions, but those gains are not translating into company-level financial impact. Ray connects this to Parkinson's Law and explains why more productive workers do not automatically produce more profitable companies.

• AI dark output. Peter introduces the concept from Semianalysis: real economic value created by AI that never registers on a P&L, using the example of a legal document that drops from $400 to $5 to produce, where the savings disappear while the token expense shows up in plain sight.

• The cost to compensation shift. Ray walks through why token spend approaching 50 to 100 percent of engineering compensation changes the entire calculus for measurement, contrasted against IT's historical 3.5 to 6 percent share of revenue.

• Case studies in good and bad. The episode breaks down three real examples: Uber's Claude Code rollout that ran out of budget without measurable output gains, Lowe's cross-functional agent deployment that built proper context tracking, and Petrobras's narrow tax compliance pilot that identified $120 million in savings and is scaling toward $1 billion.

• A four-stage framework for real measurement. Ray and Peter lay out the progression from cost visibility to utilization, proficiency, and business impact, and explain why almost every enterprise is stuck at stage one.

• Six tactical takeaways. The episode closes with concrete actions: measure outcomes not activity, design for the middle 70 percent of users rather than power users, give CFOs real budget ownership, start narrow and expand from proof points, redesign decision rights as AI scales, and treat organizational role changes as a planned program rather than a side effect.

Subscribe to the AI to ROI newsletter at ai2roi.substack.com for the full breakdown, and leave a review wherever you listen

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

The April 2026 announcement that SpaceX may acquire Cursor for $60 billion, or alternatively pay $10 billion for a compute partnership, stopped the enterprise tech world in its tracks. A four-year-old company founded by four MIT students with $2.7 billion in annualized revenue but nearly $900 million in losses on $700 million in actual revenue. This deal is not primarily a valuation story. It is a signal and a cautionary tale about the economics of the AI coding tool market.

In this Big Story edition, Ray Rike and Peter Buchanan break down what is really happening in the AI coding wars, why Cursor ended up at SpaceX's door, and where this market goes from here.

Key topics covered in this episode:

• From copilot to autonomous agent: how the AI coding market structurally shifted. Four years ago, AI coding tools suggested your next line of code. Today, they read entire codebases, plan multi-step tasks, edit files across a project, run tests, and submit pull requests with minimal human direction. Claude Code reached $1 billion in annualized revenue six months after launch, the fastest of any enterprise software product in history, and crossed $2.5 billion by February 2026. Meanwhile, 90% of enterprise developers now use at least one AI coding tool, and nearly half of all GitHub code is AI-generated or AI-assisted.

• The productivity gains are real but uneven, and the risks are underappreciated. JPMorgan deployed AI coding agents to 40,000 engineers and reported 10 to 20% productivity gains in code creation and conversion, along with a 70% increase in code deployments. But CodeRabbit's research found 1.7 times as many defects in AI-authored pull requests as in human-authored code. Meta's brief "token maxing" leaderboard experiment, designed to spotlight power users, had to be taken down within two weeks after producing high token consumption and limited usable code. Senior developers are shifting toward architecture and review roles while junior developer pipelines are shrinking, even as total software developer job postings are up 5 to 10% year over year.

• A tour of the seven major players and where the structural tension lives. Ray and Peter profile Anthropic Claude Code, GitHub Copilot, Cursor, OpenAI Codex, Google Gemini Code Assist, Replit Agent, Lovable, and Cognition's Devin across revenue, differentiation, and risk. The common thread: most point-solution coding agents run on Anthropic or OpenAI models, and those same model companies have now launched their own competing coding products. The Oracle database-to-applications parallel is not subtle.

• Why the Cursor-SpaceX deal happened and what it actually reveals. Cursor had $2.7 billion in annualized revenue, negative 23% gross margins, and was losing money faster than it was growing. Even with a $2 billion funding round in process from Andreessen Horowitz, Thrive Capital, NVIDIA, and Battery Ventures, Cursor's leadership concluded they would need to raise billions more by year-end to fund compute costs. SpaceX's acquisition offer, or the $10 billion partnership payment that Ray reads as a very generous breakup fee, solved that problem while giving XAI a revenue base three times its current size ahead of a $1.75 trillion IPO valuation push.

• The Chinese open-source threat and three scenarios for where this market goes. Kimi, DeepSeek, and Qwen models are improving rapidly and are significantly cheaper. They are, as Peter puts it, lurkers haunting every company on the list. Ray and Peter then lay out three scenarios: model makers consolidate, and IDE players get marginalized; a durable multi-tool ecosystem persists because different tools serve different workflow stages and buyer profiles; or a compute-native player builds a fully autonomous coding agent that eliminates the need for an IDE entirely. Claude Code already resolves 64.3% of real-world GitHub issues, and full autonomy for defined-scope tasks may be 18 to 36 months away.

The AI coding market is projected to reach $49-$50 billion by 2030, at a 38% CAGR. The speed gains are real. So are the defect rates, the governance gaps, the model-dependency risks, and the token-budget surprises landing on CFO desks.

If you are a CIO, CFO, engineering leader, or investor trying to make sense of who wins and what it costs, this is the episode to start with.

Read the full story at ai2roi.substack.com.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Most companies jumped straight to headcount reduction when AI arrived. AMCS Group went the other direction, starting with governance, ethics, and a question most companies never ask: not what can AI do, but what should it do? In this episode, Ray talks with Evan Schwartz, Chief Innovation Officer at AMCS Group, a platform serving resource-intensive industries across 80 countries, about how that single question reframed their entire AI strategy and produced results measured in multiples rather than percentage points.

Topics we discussed include"

Why "person plus AI" beats "AI replaces person" from day one

Companies that moved quickly on headcount reduction before AI left the lab found themselves rehiring at a higher cost when the technology did not perform in the wild as it did in controlled conditions. AMCS took a different path. Rather than treating headcount reduction as the goal (a finite game with a ceiling of zero), they pursued asymmetric growth: amplifying the capabilities of experienced employees so the business could scale revenue without incurring proportional costs. The result was output multiples, not efficiency percentage points.

The governance and ethics framework that drove better AI decisions

Operating across 80 countries with GDPR, SOC 1, SOC 2, and a range of regional regulatory requirements, AMCS could not afford to move fast and fix things later. They codified existing governance frameworks (including the EU AI Act and NIST standards) into a use-case design framework that forced a structured question before any deployment: what should this AI do? That question filtered out low-value applications, surfaced the high-impact ones, and created the foundation for what Evan calls the stewardship model.

What an AI steward actually does, and why the role is human

As AMCS built out orchestrator agents and sub-agents, they needed a clear accountability structure. The steward is always a human. Effective AI stewards share three skills: they communicate tasks clearly to orchestrators, they understand what data context the agent needs to do the job well, and they know what good output looks like even without knowing how the system produced it. That last skill, the ability to look at a result and say "that number is wrong," is what keeps agentic systems on the rails and prevents AI sprawl from becoming unmanageable.

Two external agentic AI use cases with hard ROI numbers

The dispatch management agent now monitors 700,000+ trucks globally, dynamically reroutes based on real-time events (blocked containers, missed pickups), and automatically notifies customers through their preferred channel, including rescheduling VIP accounts before they can call in a complaint. The result: 17 gallons of diesel saved per truck per month in fuel optimization, plus a $650,000 pull-forward of aged receivables (from 90-day to 30-day collection cycles) in just the first month at one customer. The customer service agent enables CSRs to double or triple their customer-touch volume by having AI handle all post-call documentation, action items, scheduling, and follow-up. That increased coverage cut AMCS's own churn rate from 6% to 3%.

How AMCS justifies AI investments internally, and why it starts with board-level metrics

AMCS is targeting ISO 42001 compliance (the AI management system standard) by year-end, which requires registering every AI tool, documenting bias risks and mitigations, and tying each use case to measurable outcomes. Evan's framework for approval is straightforward: identify your current baseline, set a target, and trace the expected return all the way to a board-level financial metric, EBITDA, free cash flow, or SG&A. Stopping at "we saved three hours" is what he calls lazy intellectualism. The real question is what those three hours produce when redirected to high-value work.

Career advice for the AI era: stop valuing yourself by the output.

Evan's message to early-career professionals is direct. If AI can produce the output, the output itself has an approaching-zero value. What has value is the ability to get AI to produce it, to steward the system, to know what good looks like, and to course-correct when it does not. The leaders of the next decade will be those who can direct a digital workforce of agents toward outcomes that matter, not those who were best at producing the deliverables themselves.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

The AI to ROI team, Ray Rike and Peter Buchanan, mark the official launch of The Big Book of AI Metrics, an 180-page, 81-metric operator's reference guide built to close the gap between AI adoption and AI ROI. Twenty-seven percent of executives say AI has met their ROI expectations, enterprise AI token spend is up 13x since last year, and most companies still can't explain what they got for the investment. Ray and Peter break down why that gap exists and what to do about it.

Topics covered:

• Why adoption, utilization, and outcomes are three different things, and why most companies stop measuring at adoption
• The five layer causal chain framework: input signals, leading indicators, operational KPIs, financial outcomes, and strategic value
• Why establishing a baseline before deployment is the single most skipped step, and why skipping it turns results into opinion instead of evidence
• Four real world case studies: Petrobras ($120M in tax savings), Stocks Insurance (83% reduction in claims processing time), Uber's cautionary token budget blowout, and Klarna's revenue per employee gains
• Three actions operators should take this week: define the outcome metric, establish a baseline, and build a measurement cadence before and after deployment

Key quote: "Adoption still is not ROI. Outcomes are ROI. And outcomes that translate into better financial performance, that's true ROI that a CFO, investor, and a board of directors can get behind." - Ray Rike

The Big Book of AI Metrics is organized into 13 functional roles, covering both operating executives investing in AI to improve their functions and B2B software executives whose product economics now depend on token consumption, inference costs, and gross margin impact.

Get the Big Book of AI Metrics: https://www.benchmarkit.ai/ai-big-book-of-metrics

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Most enterprises have deployed AI broadly. Far fewer know what they are actually getting from it.

Russ Fradin, Co-Founder and CEO of Larridin, has spent his career building measurement infrastructure at inflection points in technology adoption, from early days at ComScore measuring internet advertising to founding Larridin with backing from Andreessen Horowitz and Google's Gradient fund.

In this episode, Russ makes the case that AI spend is on a trajectory to become the number-one or number-two driver of enterprise OpEx, and that most organizations still lack the basic visibility needed to manage it.

Topics covered:

• The AI visibility gap: Why AI adoption moved faster than measurement infrastructure, and why enterprises are only now scrambling to answer fundamental questions about what they are spending, where, and by whom

• Utilization vs. proficiency vs. business impact :Why these three dimensions require separate measurement, and why the 1,800 heavy users at a 30,000-person company are not a success story on their own

• Token spend as a new category of OpEx risk: How consumption-based pricing turns every employee into a cost endpoint, with real examples of runaway agent spend and blown budgets that no one turned off

• CFO ownership of AI investment: Why AI spend is the first technology cost category large enough to pull the CFO into governance conversations that historically belonged to the CIO and department heads

• Change management as the bottleneck: Why the hard work is not experimentation but operationalizing what works, scaling proven behaviors from the top 5% of users to the full organization
• 
  
• Career advice for AI-era professionals: Work harder than the room, achieve deep tool mastery, and invest in relationships, the same fundamentals that applied before AI, now with higher stakes for the people who act on them

Russ closes with a memorable framing: "Companies have committed to a fitness journey but have not yet bought a scale; Larridin is building that scale."

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Most B2B software companies are sitting on one of the most powerful and underutilized data assets in their business: product telemetry. Every click, API call, and feature interaction is a signal. The question is whether your go-to-market organization knows how to read it.

In this episode, Ray Rike is joined by Dan Harmeson, co-founder and co-CEO of QuadSci, to explore how machine learning applied to telemetry data is changing how software companies predict churn, protect the base, and accelerate expansion revenue.

Key topics covered in this episode:

• Why telemetry data is the largest untapped GTM asset in B2B software. Dan defines telemetry data, from front-end product analytics events to back-end observability metrics, and explains why these trillions of usage signals are the single biggest data set B2B software companies generate but rarely use to make go-to-market smarter. QuadSci deploys AI locally inside the customer environment so sensitive data never moves to a third party.
• How QuadSci builds trust before the sale. Rather than asking customers to take predictions on faith, QuadSci runs a retrospective exercise: predicting churn and growth events that already happened, including data the model never trained on. Customers consistently see 90%+ accuracy, which becomes the foundation for acting on forward-looking risk signals.
• Gross revenue retention is under pressure and the data is clear. Per Benchmarkit's not-yet-published 2026 benchmarking data, GRR has declined four percentage points to 84% as an industry benchmark. For companies above $100M in ARR, roughly 95% of revenue comes from renewals and expansion, which means a two-point GRR drop cannot be offset by new logo acquisition within a 12-month window.
• Expansion revenue is a precision play, not just a CS motion. Dan walks through how QuadSci identifies Goldilocks-zone consumption patterns, surfaces cross-sell opportunities aligned to actual usage behavior, and helps account teams build nine-to-twelve month consumption forecasts that customers can actually plan around. The result is expansion conversations grounded in data, not intuition.
• Token consumption is the next frontier. As agentic AI deployments scale, CIOs and CFOs are facing unpredictable inference costs. Dan explains why the same telemetry-based approach that protects software GRR today is directly applicable to governing AI token spend inside Fortune 5,000 enterprises, a market QuadSci is beginning to address.
• Rapid fire: ROI measurement, ownership, and career advice. Dan ties AI ROI to trust and verifiability rather than vanity metrics, identifies StratOps as the emerging owner of go-to-market performance measurement, and offers practical guidance for early-career professionals on why deep business process expertise paired with AI fluency is the highest-value combination in the market right now.

If your company is facing pressure on retention, trying to build a more systematic expansion motion, or wrestling with unpredictable AI infrastructure costs, this episode delivers both the framework and the evidence behind it. Subscribe to AI to ROI on your favorite podcast app, leave a five-star rating, and connect with Ray at Ray Rike on LinkedIn to suggest a future guest.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

What does it actually look like when a CFO drives the strategic, pricing, and financial decisions behind an AI-first product transformation? Kunal Agarwal, CFO at Gorgias, the leading e-commerce customer experience platform for Shopify merchants, joins our host, Ray Rike to share the unfiltered story of how Gorgias built, priced, and operationalized its AI agent product from the ground up. This episode goes well beyond theory, covering the real decisions, real numbers, and real lessons learned from a company that has roughly half its customer base already using its AI agent product.

Episode Highlights:

• The build decision: re-architect, don't bolt on. In early 2024, Gorgias made the deliberate choice to re-architect its platform around an agentic future rather than layering AI on top of an existing help desk product. The first AI agent focused exclusively on email support, shipped in July/August 2024, and expanded from there into chat and shopping assistance. Kunal explains why starting with a single, high-confidence use case was critical to earning early adoption and trust from merchants.

• The North Star metric: full resolution rate, not deflection. Gorgias intentionally moved away from deflection rate as its primary success metric, which can mask frustrated customers who simply abandon a conversation, and anchored instead on end-to-end AI resolution rate. That metric started with a target of 20 to 25% and has scaled to 60 to 80% for their largest enterprise customers.

• Why outcome-based pricing was the only intellectually honest answer. Seat-based pricing misaligns incentives, and per-ticket pricing creates the wrong incentive to grow ticket volume rather than resolve issues. Gorgias charges per resolution, meaning it only gets paid when the AI agent delivers a measurable outcome. Kunal explains how that pricing model forces the company to stand behind product quality and why keeping it simple, at the cost of short-term revenue maximization, was the right call to accelerate adoption.

• Gross margin reality: AI-native economics are structurally different from SaaS. Kunal is candid that AI agent gross margins are lower than traditional SaaS and that denying that fact is living in an alternate reality. With LLM inference costs running approximately 55 to 60% of fully loaded cost per interaction, and infrastructure as the fastest-growing expense line, Gorgias built real-time cost instrumentation by feature, a rolling 28-day average LLM cost per interaction, and a CFO-led governance model with weekly to bi-weekly engineering check-ins to stay ahead of cost drift.

• The shopping agent and the attribution problem. Gorgias expanded its AI platform from post-sale support into pre-sale shopping assistance, helping Shopify merchants drive incremental AOV and repeat purchases. The challenge is attribution: when a customer engages with a product recommendation but converts two to three days later, did the AI agent drive that sale? Kunal describes the approach of co-creating attribution logic with customers, which is the only way to make the ROI story believable and defensible.

• The CFO as owner of AI ROI, internally and externally. On measuring the return on internal AI investments, Kunal's view is clear: the Office of the CFO owns AI ROI measurement across every function, including product, marketing, and sales. Product and engineering teams are important stakeholders but have inherent incentives to measure outcomes favorably. Independent, finance-led measurement is what gives the numbers credibility with the board.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

OpenAI built $25 billion in annualized revenue and 910 million weekly active users in three and a half years. It also has 33% gross margins, a projected $14 billion loss, a CFO who was reportedly demoted for saying the company is not ready to go public, and an investor presentation that told its software partners it plans to replace them. In this episode, Ray and Peter work through six documented challenges facing OpenAI, six specific actions that could right the ship, and what enterprise leaders should actually do with their AI strategy given all of it.

What we covered in this episode:

The model is not the moat, and ChatGPT's market share is eroding

Analyst Benedict Evans has noted that the six leading large language model companies are now roughly equivalent in capability, with no proprietary data advantage or network effect allowing any one to pull decisively ahead. ChatGPT's share of enterprise and developer usage has fallen from roughly 80% two and a half years ago to around 60% today, growing at just 4% while Claude grew 14% and Gemini 12%. OpenAI is a consumer-first product trying to pivot to enterprise at a moment when Anthropic is already the preferred first purchase for 73% of enterprise buyers according to Ramp data.

Leadership integrity and financial credibility are both under pressure

A 16,000-word New Yorker profile drawing from over 100 interviews raised serious questions about Sam Altman's management behavior and integrity. The Wall Street Journal followed with reporting on his personal investment conflicts. The CFO, Sarah Friar, was reportedly demoted after privately advising colleagues the company is not ready for an IPO. At a $852 billion valuation (roughly 28x projected 2026 revenue) with 33% gross margins and a $14 billion projected loss, institutional investors interviewed by The Information said they would not buy the stock and some indicated they would short it.

The partner ecosystem problem could be existential

In a February investor presentation, OpenAI stated it intends to build products that replace Salesforce, Workday, Adobe, Slack, and Atlassian, companies with whom it has active revenue-generating partnerships. Every systems integrator and enterprise software company building on top of OpenAI's models is now evaluating whether that is a safe long-term bet. Bill Gates defined a platform as something that creates more value for partners than for itself. OpenAI's current stated strategy is the opposite.

Six actions that could change the trajectory

Ray and Peter walk through a specific set of recommendations: launch a structured enterprise customer evidence program with named deployments and quantifiable outcomes; stop the public sniping at competitors and replace it with product and customer communication; fund an independent AI governance and safety board with real veto authority; impose IPO-grade communications discipline and treat major leaks as firing offenses; commit credibly to a partner ecosystem with defined product boundaries that give integrators a durable business case; and operate as a mature growth company, not a startup, because $30 billion in revenue demands the leadership behaviors that go with it.

What enterprise leaders should watch and do right now

Three signals will tell the real story over the next 12 months: whether Sarah Friar stays or exits, whether the IPO timeline slips to 2027, and whether enterprise case studies with quantifiable outcomes start appearing in volume. In the meantime, the strategic prescription is straightforward. Do not build single-model dependency into your AI architecture. Require the same evidence from OpenAI you would from any other vendor: verified outcomes, clear product roadmap, and accountability. And build API portability into your application design so you can move if you need to.

The closing question: if you had to pick one LLM company to invest a million dollars in, where does it go?

Peter picks Google, citing distribution advantages, DeepMind's research depth, and full control over its own financial destiny. Ray picks Anthropic, citing a lower revenue base with larger upside, near-universal goodwill across hyperscalers and enterprise buyers, and a safety-first positioning that is proving to be a genuine competitive differentiator. They agree on the conclusion: OpenAI is the defining company of the AI generation, but Netscape, Lotus, and BlackBerry were all category leaders too.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.