Iran War, Oil Shock, Off Ramps, AI's Revenue Explosion and PR Nightmare

Episode Summary

Executive Summary: The episode centered on two major debates: the geopolitical and economic fallout from the Iran war, and whether AI revenue is real, durable, and socially beneficial. The hosts argued for a limited U.S. objective and an off-ramp in Iran, while warning that escalation could trigger oil shocks and political blowback. They then pivoted to explosive AI growth at OpenAI and Anthropic, debating whether it reflects true production demand or still-heavy experimentation, alongside concerns about PR, regulation, and social disruption.

Main Topics: Iran war and oil-market shock (Priority: 5/5): The hosts discussed Brent crude volatility, the risk of prolonged conflict, and how oil spikes are already affecting inflation, GDP, confidence, and market valuation. Trump doctrine, off-ramp, and regime-change debate (Priority: 5/5): Brad, Sachs, and others argued Trump’s approach is more pragmatic than neocon interventionism and that the U.S. should declare victory and exit rather than expand the war. China as the hidden strategic variable (Priority: 4/5): Several speakers framed Iran, Venezuela, and energy insecurity as ultimately a China problem, arguing Xi has incentives to strike a grand bargain with Trump. AI revenue acceleration at OpenAI and Anthropic (Priority: 5/5): The group debated extraordinary revenue growth, with both companies seemingly scaling from experimentation into major commercial traction, especially via coding assistants and agents. Experimental vs production AI revenue (Priority: 5/5): There was sharp disagreement over how much enterprise AI usage is real production value versus pilot budgets and temporary experimentation, especially in large firms. AI messaging, fear, and public backlash (Priority: 4/5): Speakers criticized industry leaders for apocalyptic messaging that scares the public, fuels regulatory backlash, and slows adoption in the U.S. compared with China. Taxes, wealth migration, and state policy (Priority: 3/5): The hosts criticized Washington’s millionaire tax and broader wealth-seizure proposals, arguing high earners can and will relocate, shrinking tax bases.

Key Arguments: Oil spikes from the Iran conflict create direct inflation pressure and indirect confidence/GDP damage, with Goldman reportedly raising PCE inflation and lowering growth forecasts. The Trump doctrine is presented as limited and pragmatic: destroy threats to U.S. interests, declare victory, and avoid long occupation or democracy-promotion. Escalation could produce worse outcomes than a temporary Strait of Hormuz closure, including Gulf infrastructure attacks, desalination disruption, and broader regional catastrophe. China is the true pressure point because it depends heavily on Middle East and Venezuelan oil, making the summit with Xi strategically critical. Anthropic and OpenAI are showing unprecedented revenue growth, implying AI demand is real and the market for intelligence tools is much larger than skeptics thought. A large share of AI revenue is still experimental, especially in enterprise, but coding assistance is the clearest breakout use case and is already in production at many startups. Industry leaders are hurting adoption by alternately claiming superintelligence doom and commercial utility, creating fear, policy backlash, and confusion. Open source is a major competitive force, but frontier labs are still growing rapidly, suggesting the total addressable market is expanding rather than shrinking. State wealth taxes may accelerate the exit of high earners and cause negative fiscal outcomes rather than solving budget problems. The long-term opportunity for AI is to solve core consumer problems in healthcare, education, housing, and access to expertise, if regulation allows it.

Data Points: Brent crude price spike: $84 to $119 to $84 to $100; currently around $99 - Discussed as the key market signal during the Iran conflict Goldman Sachs PCE inflation forecast: 2.1% to 2.9% - Updated forecast cited as a knock-on effect of higher oil prices and conflict uncertainty Goldman Sachs core PCE forecast: 2.2% to 2.4% - Excluding direct oil effects, still expected to rise due to second-order impacts Goldman Sachs GDP forecast change: -30 basis points - Lowered annual growth forecast because of the war and oil shock Goldman Sachs unemployment outlook: higher unemployment - Expected labor-market softening tied to conflict-driven uncertainty PCE inflation peak valuation context: SP at 24x, now 21x - Used to explain market multiple compression during geopolitical stress Polymarket U.S. boots-on-ground odds: 27% by end of March; 57% by end of year - Used as a measure of market expectations for escalation Trump account signups: 100,000 kids a day - Invest America/Trump accounts claimed to be signing up rapidly Eligible children for Trump accounts: nearly 30 million - Children in America eligible for at least $250 if they claim the account Oil market history: $100 in the Gulf War; $216 in 2008 dollars; $115 during Russia-Ukraine (133 in today’s dollars) - Historical comparison of current oil spike with prior shocks Anthropic revenue run rate: $14 billion annualized - Claimed for February; used to illustrate rapid scale-up Anthropic monthly revenue: $6 billion in February - Cited as a ‘nuclear moment’ for AI monetization OpenAI revenue run rate: $20 billion annualized - Presented as ending 2025 at this level OpenAI revenue growth: $2 billion to $20 billion in 25 months - Used to show extraordinary ramp Anthropic valuation: $380 billion - Discussed alongside its rapid revenue growth OpenAI valuation: $840 billion - Compared against Anthropic in valuation debate Compute investment / data center example: $50 billion for a 1-gigawatt data center shell in Arizona - Used to explain the depth of the AI capital-expenditure J-curve Sarah Fryer revenue heuristic: $10 billion annual revenue per gigawatt - Used to estimate payback and profitability timing for AI infrastructure Open source token share: ~85% of tokens - Brad said their portfolio/startups are increasingly running open-source models locally Data center cancellations: ~25 canceled in 2025, about 5 GW lost - Used to argue AI/data-center backlash is creating large revenue losses Potential lost revenue: $50 billion/year from 5 GW - Estimated revenue tied to canceled data-center projects Additional protested data centers: ~100 protested by end of February 2026 - Projected to imply another wave of cancellations Potential additional lost revenue: $70 billion/year from 7 GW - Forward-looking estimate from current protests California billionaire tax polling: 25% support - Used to show public resistance to wealth-seizure proposals California tax proposal impact: 30,000 households; $4 billion for general fund - Washington-style millionaires tax comparison and state budget framing California tax simulation: 71% negative NPV; ~$25 billion hole - Hoover Institution Monte Carlo analysis cited to criticize wealth taxes China consumer reliance on oil: 20% of domestic consumption from Venezuela and Iran - Used to argue China is more exposed than the U.S. to the conflict China youth unemployment: 25% - Mentioned as a vulnerability if oil flows are disrupted AI optimism survey: ~80% positive in China; 30s in the U.S. - Used to explain cross-country differences in AI sentiment

Pivotal Quotes: "I think the Trump doctrine is far more pragmatic than the neocon doctrine." — Brad Gerstner: He argued the U.S. objective is limited destruction of threats, not nation-building or democracy promotion. "The only justification you could possibly have would be that if we don't do it, our adversaries will do it and we will be subject to their rule of law." — Sachs: He framed AI and military power as necessary defensive tools, while criticizing reckless escalation and social disruption. "I think the industry is doing a horrible job with PR. I think they are scaring the bejezas out of the public." — Sachs: He explained why U.S. sentiment on AI is negative and argued that doomer messaging is fueling backlash.

Implications: Listeners should expect continued volatility in oil, policy risk, and markets if the Iran conflict drags on. For AI, the takeaway is that revenue is real but still mixed with experimentation, while messaging and regulation may become as important as model quality.