The Venture Capital Show

"It doesn't matter how great the deal is." Matthew Tarver learned that at 27, losing nearly $100,000 on an oil and gas deal bought out of bankruptcy and run by the wrong people. Today he leads seed rounds in nuclear power and battery storage, and about 90 percent of every bet is still the people.Matthew Tarver is founder and general partner of Foundt Capital, a Dallas seed firm he started in 2022 after three years structuring deals at Sohan Capital and an engineering career at Albemarle. Fund I was about $2.5 million across five companies. Foundt led four of those rounds; four companies are still running and one was a complete loss. Two follow-on SPVs took the firm to just under 15 percent of a small modular reactor company where he sits on the board.This is an episode about underwriting hard tech with a small fund. Matthew explains the reactor plant that can be built and run on packaged natural gas boilers today, then switched to nuclear with about 80 percent of the equipment staying in place, and why a roughly $700 million build does not scare a seed investor when hyperscaler customers and project finance can carry the back end. He walks through the NRC clock (a license in 18 months at best, then about four years to build), two battery bets on opposite ends of the market, and the molten salt battery that was carved out of a reactor design and became its own company.He also covers the math he sells his LPs: discount every founder's revenue projection by about 90 percent, then look for $5 back for every $1. Why a $2.5 million position that returns $45 million is a home run worth waiting for. Why he turned down an LP who wanted to be chairman of the fund. And why he and his partner Miller Maestri will not raise Fund II on paper gains: DPI, money actually returned to investors, before TVPI."I might not be a 180 IQ guy, but I'm relentless."Hosted by Ryan Harper. Produced by Harper Belmont Media.CHAPTERS0:00 Cold open: the $100K oil deal1:07 What venture capital actually is2:42 Check sizes, follow-ons, and a $2.5M Fund I4:02 One in 10 vs one in four: startup survival odds5:22 An engineer who invests in people8:03 The oil and gas deal that went wrong10:42 12 of 16 still running. How many pay dividends?12:19 Backing a small modular reactor13:05 A health-tech AI that cuts data center power14:50 The first box: do I like these people?18:03 The nuclear plant that starts on natural gas19:45 Just under 15% of a reactor company20:10 Two battery bets, opposite ends of the market21:54 The molten salt battery carved out of a reactor23:46 Exit timing, the NRC clock, and $5 for every $127:03 Raising Fund I: one mentor, six LPs in Louisiana29:39 Not living off the management fee33:27 Should founders pay themselves?38:25 AI tools: learn to do it yourself first40:30 Leading rounds and a 90-day diligence cycle44:06 The checklist: people, product, can we help46:39 LP expectations and the unicorn myth47:22 Turning down an LP who wanted to be chairman48:23 $2.5M in, $45M out: what a home run looks like52:12 The automatic no: a purple pitch deck57:21 Fund II: why DPI comes before TVPI59:01 Still raising: capital for the reactor company1:01:08 Getting paid for raising: shares over cash1:03:44 Boots on the ground: $50K a ton graphene1:05:36 Advice for LPs, founders, and future VCs1:10:49 Who he wants to work with nextSPONSORED BYISTA Consulting | http://www.istaconsultingllc.com/ISTA Consulting is a one-stop shop for emerging VC funds and their portfolio companies, offering financial operations, investor reporting and controller/CFO advisory services - all under one roof.Foundt Capital: https://www.foundtcapital.comMatthew Tarver: https://www.linkedin.com/in/matthewtarver/The Venture Capital Show: https://www.theventurecapitalshow.comRyan Harper: https://www.linkedin.com/in/ryansharper/

"We had the opportunity to invest in OpenAI in the $30 billion round... But we couldn't do the right due diligence." David Evans passed. By his own count it would have been a 100x. He explains the rule that made him pass, and why it still governs every check he writes.David Evans is Managing Partner of Sentiero Ventures, a Dallas seed fund that has invested only in AI-native companies since 2019. Fund I was about $5 million across 12 investments. Fund II is $12 million, nine companies in, about five to go. First checks run $250,000 to $750,000, up to roughly $1.75 million per position over time.This is a fund math episode. Sentiero underwrites every company to a potential 25 to 40x because half the portfolio will fail, and it sells LPs a 3 to 5x target. David walks through how a venture portfolio actually distributes, why the power law has become an excuse to forget the other nine companies, and why an emerging manager's fundraising cycle is offset from when the fund realizes. That gap is exactly what "I'm in Anthropic, I'm in Thinking Machines" fills when a GP goes out to raise the next fund. He calls it a broken incentive, and says the personal relationships with his LPs are the governor that keeps him out of it.He also lays out the "bagel or a trillion" market: by his count close to 60 percent of 2025 venture dollars went to about ten companies, leaving good, default-alive founders growing 30 to 50 percent a year without capital. And he explains the thesis behind Sentiero, the Ubers of AI: companies that cease to exist if you take the AI away, built for vertical markets too small for OpenAI, Anthropic, or Google to bother with."I'm not selling my soul. I'm selling the asset."Also in this episode: default alive versus default dead, what a post-mortem is for on a ten-year bet, the unknown unknowns no diligence can catch, founder salaries, why every data center gets used and the peak buyers still lose, the 1,000x Anthropic story and the lockup math nobody runs, and why "why wouldn't Google do that?" is the laziest question in venture capital.Hosted by Ryan Harper. Produced by Harper Belmont Media.CHAPTERS0:00 Intro1:17 What venture capital is actually for2:31 Rounds are named, but the names are misleading4:32 Why companies stay private longer7:03 Default alive vs default dead10:34 His own post-mortem: not getting creative about financing14:37 "No bank will ever own George Evans"16:08 Why every deal gets a post-mortem20:40 Unknown unknowns: what no diligence can catch23:41 How many software companies is Claude killing?25:41 Early seed: the market votes with its checkbook first27:12 Who the LPs are, and why Fund I is "trust me"29:44 Fund I: $5M, 12 investments, two exits, one failure30:44 Founder salaries: why ramen is the worst possible thing34:16 The math: 25 to 40x per company to return 3 to 5x37:17 The power law is not an excuse to forget the other nine38:48 Bagel or a trillion: 60% of VC to ten companies40:49 Hype cycles, GPT-2, and open-source Chinese models44:51 The thesis: the Ubers of AI46:52 Amazon Basics and the laziest question in venture48:23 Check sizes and Fund II: $12M, nine in, five to go50:54 The 1,000x Anthropic story and the lockup math53:25 Every data center gets used; peak buyers still lose55:57 The broken incentive: "I'm in Anthropic" sells a fund56:57 Passing on OpenAI at $30B58:28 Who Sentiero wants to work with nextSPONSORED BYISTA Consulting | http://www.istaconsultingllc.com/ISTA Consulting is a one-stop shop for emerging VC funds and their portfolio companies, offering financial operations, investor reporting and controller/CFO advisory services - all under one roof.Sentiero Ventures: https://sentiero.vcDavid Evans: https://www.linkedin.com/in/davidevansaivc/The Venture Capital Show: https://www.theventurecapitalshow.comRyan Harper: https://www.linkedin.com/in/ryansharper/

Trey Bowles takes 50% of a company on day one, and argues the founder still walks away with more than venture capital would leave them.Trey Bowles, co-founder and managing partner of 1845 Ventures, joins Ryan Harper on The Venture Capital Show to explain what a venture studio actually is, how the cap table works when the studio is a co-founder, and why capping a company at $2 to $3 million for life is the point rather than a limitation.1845 co-founds companies at inception on a 50/50 split, puts in $2 to $3 million total, and engineers toward a $50 to $150 million exit in three to four years. Trey walks through the math: at exit the table is roughly a third to investors, a third to 1845, a third to the founder, against the 5 to 10 percent a founder typically holds after a traditional venture path. He explains why they call potential acquirers during diligence and ask whether they would actually buy the company, why the strategy is "go get 217 customers" instead of "someday be worth a billion," and why Texas produces a steady stream of $100 to $300 million acquisitions too small for the buyer to bother writing a press release about.

Sponsored by: http://www.istaconsultingllc.com/ISTA Consulting is a one-stop shop for emerging VC funds and their portfolio companies, offering financial operations, investor reporting and controller/CFO advisory services - all under one roof.

Cybersecurity's next $30B company won't protect your network. It will protect your AI. Delta-v Capital's Dan Williams explains why.Dan Williams, Partner at Delta-v Capital, joins Ryan Harper on The Venture Capital Show to break down the AI security thesis: why large language models and AI agents create an entirely new attack surface, why enterprises are racing to wrap their models in an "AI firewall," and why he believes this category produces the next Wiz, Zscaler, or Cloudflare.Dan is a former software engineer turned growth-stage investor who leads Delta-v's infrastructure software practice across cybersecurity, AI, and DevOps. He walks through a real prompt-injection attack, why an AI security company in his pipeline is growing 1,000% year over year, how that rewrites the old "triple triple double double" growth playbook, and how a small operator-led firm competes against the multi-billion-dollar names.IN THIS EPISODE- Why AI models and agents are a brand-new attack surface for cyber- How prompt injection works (the "white text in a webpage" bank-account example)- What an "AI firewall" actually does: watching every prompt, response, and agent action- Security FOR AI vs. AI FOR security, and why both are exploding- The 1,000% growth company, and how AI is bending venture growth benchmarks- Why new tech waves always mint new cybersecurity giants (Wiz to Google for $32B, Zscaler, Cloudflare)- How Delta-v avoids competitive conflicts across a concentrated portfolio- Operators vs. financiers: how a growth firm differentiates from the mega-funds- Why AI makes the 10x engineer a 100x engineer, and why it is not dystopian for labor- Naval Ravikant on the three forms of leverage: capital, code, and media- Lifelong learning, stoicism, and the mindset behind the work