Iran War, Oil Shock, Off Ramps, AI's Revenue Explosion and PR Nightmare
Speakers
Jason Calacanis — Host
— angel investor, All-In host
David Sacks — Host
— White House AI/crypto czar
Chamath Palihapitiya — Host
— Social Capital CEO
Brad Gerstner — Guest / Investor
— Altimeter Capital CEO
Summary
- Oil prices experienced extreme volatility, spiking to $119 per barrel due to the Iran conflict before retreating to $90, signaling the market expects a short-lived war.
- Goldman Sachs raised its PCE inflation forecast to 2.9% and lowered GDP expectations due to the oil shock, causing the S&P 500 multiple to compress from 24x to 21x.
- AI revenue is scaling at an unprecedented historical rate: OpenAI reached a $20B annualized run rate, and Anthropic hit a $14B run rate (including a $6B month in February).
- Enterprise AI adoption is currently driven by coding assistants, which are bypassing traditional IT budgets and directly augmenting labor budgets due to a structural shortage of software engineers.
- Data center infrastructure faces massive capital requirements ($50B per 1 gigawatt) and severe NIMBY headwinds, with 40% of protested data centers canceled in 2024, though states like Texas remain highly accommodating.
- State-level wealth and millionaire taxes (e.g., Washington, California) are actively driving high-net-worth individuals to relocate, resulting in net-negative tax revenues and structural budget deficits.
Trade Ideas
"Just a few months ago, the S&P peaked at 24 times. Now we're at 21 times. But I think the market may be getting it a little bit wrong... I think Trump has a very limited set of goals. He wants to destroy and degrade threats... He doesn't want to spread democracy." The broader market has sold off and multiples have compressed due to "post-traumatic stress" fears of a prolonged Middle East quagmire and resulting inflation. If the administration secures a quick exit and avoids a long-term occupation, these macroeconomic fears will subside, allowing equity multiples to expand back to previous highs. LONG the broader market indices to capture the multiple expansion that will occur once the geopolitical overhang and associated inflation fears are resolved. If the conflict drags on, forces boots on the ground, or causes sustained inflation above 3%, multiple compression could worsen.
"The market literally took oil from 120 a barrel to 90 a barrel almost in a nanosecond... there is a coordinated release of about 400 million barrels of petroleum. That's going to dampen the effect of any price spike." The geopolitical risk premium in oil is being rapidly priced out because the market anticipates a swift diplomatic off-ramp rather than a prolonged conflict. Furthermore, massive coordinated releases from strategic petroleum reserves will artificially cap any near-term supply shocks. AVOID chasing the oil spike, as the combination of strategic reserve releases and pragmatic political de-escalation will likely suppress sustained price breakouts. Escalation by rogue actors targeting Gulf desalination plants or broader energy infrastructure could force a structural repricing of oil regardless of US policy.
"Anthropic unquestionably has a lot of financial momentum... we had a $6 billion month out of Anthropic in February... OpenAI ended 2025 at 20 billion annualized run rate." Frontier AI models are generating historic, unprecedented revenue growth by augmenting labor rather than just competing for fixed IT budgets. Because OpenAI and Anthropic are private, the most direct public market beneficiaries are their primary hyperscaler partners and investors, who will capture massive cloud compute and API distribution revenues. LONG the major cloud providers as they are the foundational toll roads capturing the explosive, proven enterprise spend on frontier AI models. A significant portion of current AI spend may be experimental; if enterprises fail to see long-term ROI, test budgets could dry up, leading to a sharp revenue contraction.
"They need cheap access to money to continue to build out the compute they need to support. There is more compute constraint in these businesses this very day than they've had any time in the last 3 years." The staggering revenue figures from OpenAI and Anthropic validate the massive capital expenditures required to train and run these models. Because the models are constrained by compute and are still in the "early innings," hyperscalers and AI labs will be forced to continue buying next-generation GPUs at an accelerating pace to maintain market share. LONG the dominant AI chip provider, as the proven end-user revenue from AI labs justifies and guarantees continued hyper-scale infrastructure spending. Open-source models running on cheaper, alternative silicon or local devices could eventually commoditize the need for massive centralized GPU clusters.
Up Next
This All-In Podcast video, published March 13, 2026, features Brad Gerstner, Chamath Palihapitiya discussing SPY, QQQ, USO, MSFT, AMZN, GOOGL, NVDA. 4 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Brad Gerstner, Chamath Palihapitiya · Tickers: SPY, QQQ, USO, MSFT, AMZN, GOOGL, NVDA