Featuring
Flexport founder Ryan Petersen joins the besties for a Red Sea stress test, then the room turns to blocked tech deals and AI's effect on startup economics. The spiciest exchange asks whether Gulf prosperity plans could tame the Houthis or whether Washington was heading for a shooting war; hindsight was not kind to the optimists. Sacks has the strongest episode, pairing the best Red Sea risk model with the more durable forecast on AI and venture capital.
Spice rack
Would aligned regional interests end the Red Sea crisis quickly, or pull the United States into escalation?
Original point: The Gulf states' long-term development plans and the shared interests of China and Western powers would create enough pressure to resolve the Gaza and Yemen conflicts quickly.
What everyone argued
Chamath Palihapitiya
Saudi Arabia, the UAE, and Qatar would not let regional proxy wars derail decades of investment, while China and Western governments shared an interest in open trade. He expected that alignment to pressure Iran and bring both wars to a quick close.
Jason Calacanis
Regional leaders were prioritizing modernization, technology, tourism, and commercial ties, so those larger forces would contain the conflict and deal with spoilers. He later narrowed the claim from resolution to containment.
David Sacks
Saudi Arabia wanted stability but had just emerged from a costly war in Yemen and could not easily force the Houthis to stop. With Israel committed to its Gaza campaign and the Houthis tying attacks to that war, U.S. military action and a prolonged shipping crisis were plausible.
Ryan Petersen
Petersen agreed that Gulf governments hated the disruption but warned that history does not guarantee progress merely because progress is economically useful. He emphasized the shipping network's fragility and the unfavorable economics of defending expensive vessels from cheap attacks.
Winner circle
David Sacks and Ryan Petersen had the better model. They identified the veto players, insurance constraints, and coercion problem that the optimistic side skipped. Hindsight delivered direct U.S.-UK strikes and sustained trade disruption, so shared prosperity goals were a brake on escalation, not a settlement mechanism.
Commentary
Chamath Palihapitiya
Assumptions and fact checks
Shared economic interests would translate quickly into enough coercive or diplomatic leverage to end both conflicts.
Why it mattersThe incentives were real, but the armed actors faced political and military incentives that commerce alone did not override. The forecast skipped the mechanism between alignment and settlement.
The Red Sea disruption was likely to create more inflation pressure in Europe than in the United States.
CheckUNCTAD reported Shanghai-Europe spot rates up 256% by February 2024 versus 162% to the U.S. West Coast, and said European energy exposure was especially acute.
Jason Calacanis
Jason improved his case when he switched from 'resolution' to 'containment,' but that concession also weakened the original optimistic forecast.
Assumptions and fact checks
Economic modernization gave Gulf governments enough leverage to contain Houthi attacks without a wider military campaign.
Why it mattersThe wider conflict was partly contained, but navigation still required military defense and rerouting. The economic agenda constrained governments without controlling the Houthis.
A broad international coalition was organizing to protect Red Sea navigation.
CheckCENTCOM described a defensive coalition of more than 20 countries and a U.S.-UK strike supported by Australia, Canada, the Netherlands, and Bahrain.
David Sacks
Sacks won by asking the unglamorous but necessary question: who can make each armed party stop? His downside framing stayed conditional instead of turning risk into prophecy.
Assumptions and fact checks
Without a Gaza ceasefire, neither diplomacy nor commercial pressure alone would restore normal Red Sea shipping quickly.
Why it mattersThe immediate hindsight strongly supports this. Military strikes followed, and UNCTAD documented severe continuing diversion and freight-rate effects.
The United States was likely to conduct military action against Houthi targets if attacks continued.
CheckWithin three weeks, U.S. and UK forces struck Houthi radar, air-defense, storage, and launch sites after 27 attempted attacks on ships.
Ryan Petersen
Petersen's best move was separating what states wanted from what ship operators could rationally risk. His one miss was underestimating how long the disruption could survive those incentives.
Assumptions and fact checks
Insurance and asymmetric attack costs could keep commercial vessels away even when all major trading powers wanted the route open.
Why it mattersThat is exactly the coordination failure the later diversions exposed: shared preference for open trade did not make individual voyages insurable or safe.
The attacks could cause large shipping effects even without reaching COVID-era disruption.
CheckUNCTAD found Suez transits down 42% from peak, weekly container-ship transits down 67%, and Europe-bound spot rates up 256%, while noting the combined shock had not yet matched the pandemic crisis.
Would AI radically shrink startup capital needs and the size of venture funds?
Original point: AI could be more deflationary than AWS, letting tiny teams reach profitability and making today's large venture funds several times too big.
What everyone argued
Chamath Palihapitiya
AI coding tools and offshore teams could give a four-person startup the output of a ten-person company, extend runway, preserve founder ownership, and support smaller exits. He projected that $200 million to $1 billion venture funds might need to shrink to roughly one-fifth their size.
Jason Calacanis
His 2023 seed portfolio was already raising less, hiring globally, and targeting break-even without assuming another round. He agreed that founders were replacing prestige fundraising with capital efficiency and larger ownership stakes.
David Sacks
AI was restocking the opportunity set after cloud, social, and mobile had matured. Product development might get cheaper, but startups still needed capital to build sales capacity, fund customer acquisition, and absorb execution misses; AI had not yet raised sales quotas enough to erase those costs.
Winner circle
David Sacks wins the system-level forecast. Chamath and Jason correctly anticipated leaner AI-native startups, but Sacks separated cheaper product development from the rest of the capital stack. The result was bifurcation—not a 5x shrink—with tiny teams at one end and historically concentrated mega-financing at the other.
Commentary
Chamath Palihapitiya
Chamath saw the founder-ownership upside early. He made a category error when he converted cheaper coding into a system-wide forecast for fund size.
Assumptions and fact checks
AI's labor savings would justify shrinking established venture funds by about 5x.
Why it mattersThe forecast captured software-building efficiency but missed the capital intensity and concentration created by frontier models, compute, distribution, and aggressive competition.
AI would become a dominant force in venture investment.
CheckBy 2025, AI accounted for 65.4% of U.S. venture deal value according to the NVCA 2026 Yearbook.
Jason Calacanis
Jason offered useful field notes, not market proof. The right conclusion was 'a viable lean path exists,' not 'the whole financing stack will shrink.'
Assumptions and fact checks
Observed capital efficiency in one seed portfolio would generalize across the venture market.
Why it mattersIt generalized well to many application startups, but not to capital-intensive model, robotics, energy, biotech, or infrastructure companies.
Venture fundraising became more difficult for new managers even as AI investment surged.
CheckNVCA reported only 101 first-time funds in 2025, the lowest since 2011, while total U.S. venture deal value reached $320 billion and AI took 65.4% of it.
David Sacks
Sacks had the more complete cost model. He treated coding leverage as one input to startup economics, not as the entire balance sheet.
Assumptions and fact checks
Go-to-market and execution costs would keep meaningful startup capital needs even as AI reduced product-development labor.
Why it mattersHindsight supports the decomposition. AI created cheaper building blocks and much larger markets, while sales, compute, infrastructure, and competitive spending absorbed vast capital.
AI restocked venture opportunities rather than simply shrinking the market.
CheckU.S. VC firms deployed $320 billion across 15,352 deals in 2025, the second-highest deal value on record, and 65.4% of that value went to AI.
Was aggressive antitrust enforcement the main cause of the venture M&A drought?
Original point: The blocked Adobe-Figma and Illumina-GRAIL deals showed that venture-backed companies no longer had a viable M&A path, forcing an unprepared IPO market to carry nearly all liquidity.
What everyone argued
Chamath Palihapitiya
Large blocked or abandoned deals signaled that investors could not underwrite M&A exits with confidence. With banks and research coverage also unable to support young public companies, startups would remain private for 15 to 20 years and risk capital would shrink.
David Sacks
A 15-month review and a billion-dollar breakup fee would chill future deals because targets and buyers could not trust the approval process. Removing acquisition as one of only two successful startup outcomes would reduce venture returns and risk investment.
David Friedberg
The blocked deals were not numerous enough to make antitrust the core M&A constraint. Large buyers were instead facing higher interest rates and stricter return-on-invested-capital tests, while smaller acquisitions could still produce excellent venture outcomes because startups needed less capital.
Winner circle
David Friedberg wins narrowly. The Adobe-Figma facts validate the regulatory-chill argument at the deal level, but Chamath and Sacks stretched that evidence into a venture-wide diagnosis. Friedberg's rates-and-ROIC explanation had broader reach and required fewer unsupported leaps, though the overall cause remains mixed.
Commentary
Chamath Palihapitiya
Chamath was strongest on the liquidity feedback loop and weakest when 'harder to underwrite' became 'no viable path.' The first is supported; the second outran his evidence.
Assumptions and fact checks
A few blocked mega-deals showed there was no viable M&A path for early-stage venture-backed businesses.
Why it mattersThose cases proved meaningful regulatory risk for large or strategically adjacent combinations, not the disappearance of smaller acquisitions. The claim lacked a venture-wide denominator.
David Sacks
Sacks explained why this particular review could deter similar deals. He did not establish that the deterrent dominated the cost-of-capital reset across the broader exit market.
Assumptions and fact checks
Regulatory uncertainty over large acquisitions would materially depress venture investment across Silicon Valley.
Why it mattersIt plainly raises the required return and deal risk, but the magnitude cannot be isolated from interest rates, valuations, and buyer discipline. Later AI investment surged despite continued liquidity pressure.
Adobe owed Figma a $1 billion termination fee when the merger ended.
CheckAdobe's SEC filing states that it would pay Figma the previously agreed $1 billion cash termination fee within three business days.
David Friedberg
Friedberg won by matching the scope of his conclusion to the evidence. He could have made it decisive with venture-specific deal counts instead of relying on intuition about smaller exits.
Assumptions and fact checks
Higher rates and buyer ROIC discipline mattered more broadly than antitrust enforcement to the M&A slowdown.
Why it mattersThe evidence supports a multi-cause slowdown, and this mechanism applies across buyers and deal sizes. Regulation was dispositive in specific cases but was not shown to dominate venture M&A as a whole.
Large transactions still occurred despite the tougher enforcement environment.
CheckFTC and DOJ recorded 2,031 HSR-reportable transactions in fiscal 2024, about one quarter valued above $1 billion. That does not measure venture M&A alone, but it disproves a market-wide disappearance of large deals.
This was a clean map of who preferred stability, but not a theory of how stability would be imposed. The missing enforcement step proved decisive.