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

Welcome back to the number one pod in the world; Freeberg’s off inventing vegetables, so I put Brad Gerstner in the hot seat after his State of the Union shout-out—did you know it was coming or was that a surprise?

Total surprise; the president added it himself, and being in that tiny chamber for a 250‑year tradition was electric regardless of party, just a great civic moment.

On the kids’ ‘Trump accounts,’ we’re onboarding more than one hundred thousand a day with almost thirty million eligible for at least two hundred fifty dollars, going live July fourth to get Main Street owning great American companies.

Idea for you: a voluntary equity pledge where founders pledge a slice of their shares to kids’ accounts over time; has that surfaced?

Yes, stay tuned for big announcements as we approach July fourth.

Pivot to Iran: oil has whipsawed from the mid‑80s to near one hundred with the Strait of Hormuz shut, and prediction markets put real odds on U.S. troops; how does this end and what’s the market impact?

Rising oil hits PCE and confidence, with Goldman lifting PCE to about 2.9 and trimming GDP while multiples slid from 24 times to 21, but I think the Trump doctrine is limited, pragmatic, and shorter‑duration than neocon wars.

The instant oil drop from about 120 to 90 after Trump said the war would end soon showed markets believe this won’t be sustained, and coordinated SPR releases plus IEA barrels can blunt spikes while we find an off‑ramp.

Declare success and step back now; escalation risks Iranian tit‑for‑tat on Gulf oil and desal plants, turning a supply crunch into a humanitarian disaster, while Israel’s defenses could get exhausted and trigger unthinkable responses.

Politically, a long war sinks Republicans in the midterms and fractures MAGA; he needs a fast off‑ramp or the base sees betrayal.

I expect a near‑term off‑ramp consistent with a narrow mission to degrade threats, not nation‑build.

All roads lead to China; Xi and Trump have a high‑stakes summit, and China’s oil exposure and youth unemployment create leverage for a grand bargain that helps reopen the strait.

The U.S. can state objectives met and pull back; if Iran keeps hitting ships, Gulf states, Israel, and heavy importers like China and India have the strongest incentive to pressure Tehran because they bear the biggest energy pain.

The president’s instincts favor swift, decisive actions, so the key is resisting neocons who would expand goals and drag this out.

Back to our lane: Anthropic and OpenAI are posting historic revenue ramps, with Anthropic reportedly at a mid‑teens run rate and OpenAI near twenty billion; Brad, which is the better buy and do they go public soon?

Both are exceptional, and the big February breakthrough was revenue finally tied to labor augmentation via agents, not just IT, with compute now the bottleneck and public listings likely this year to tap cheaper capital and broaden ownership.

Caution on quality: most enterprise spend is still pilots, not mission‑critical workflows, and recent agent‑written code incidents forced human‑in‑the‑loop edicts because regulated domains cannot tolerate hallucinations.

There’s experimentation, sure, but critical users like defense, Palantir, and major enterprises are running production workloads, and sustained month‑over‑month growth will separate hype from durable demand.

The breakout enterprise use case is coding assistance, tapping a chronic shortage of engineers; paying per token to generate code scales easily and sets the stage for agents, which should expand use without immediate layoffs.

Adoption is real in startups right now across legal review, marketing, SDRs, HR, and accounting, where they replace contractors and hires with LLM workflows.

On profitability, the J‑curve is brutal: building a one‑gigawatt AI data center can land near fifty billion in total outlay, with roughly ten billion in annual revenue capacity per gigawatt and a five to six year payback before real profit.

Hype has also warped messaging; doomerism raises money but spooks governments and partners, so the industry needs sober, compliant posture over theatrics.

AI will shift economic power among voter blocs and raises societal risk; without clear national‑security framing and a plan for disrupted workers, you will not sustain public support.

Does AI have a PR problem; people here are anxious about jobs and inequality while China leans into abundance.

Yes, and like past industrial shifts we’ll see anxiety, but this year can prove clear benefits in health, drug discovery, and education if leaders coordinate messaging and show tangible wins.

The U.S. is uniquely pessimistic because dystopian media, CEO scaremongering, and EA‑funded doomer think tanks feed regulatory panic and push licensing schemes that lock in incumbents.

That fear loop empowers bans like New York’s move to restrict medical and legal chatbot advice, which hurts the poorest who most need affordable guidance.

The rhetoric now fuels calls to halt data centers, as we saw from Bernie, which shows how counterproductive the industry’s scare‑first narrative has become.

Doomer groups have billions to seed talking points and even fund journalism, and they are behind many NIMBY attacks on data centers despite companies agreeing to pay for incremental power and water solutions.

Data‑center cancellations are surging, with gigawatts of capacity scrapped that translate to tens of billions in lost annual AI revenue and tax base, while places like Texas stay wide open.

Brad, open source plus fast Apple silicon is cutting token spend; how does that hit frontier labs?

Open source is booming and many run ensembles, but the fact that frontier labs are still scaling revenue in that environment means the total addressable market is far larger than we expected.

Housekeeping: All‑In Summit and Liquidity are at all‑in dot com slash events; please fill out the listener survey.

Last story: Washington’s new millionaires tax kicks in later this decade as Howard Schultz decamps to Miami, raising the question of whether these policies just push creators and capital out.

California already proved state‑level wealth grabs miss targets and blow budget holes when top taxpayers leave, sticking the middle class with the bill.

A federal wealth tax is being floated at levels that would effectively confiscate fortunes over two decades, and political actors are already repositioning to make it party orthodoxy.

We have seen where class‑war rhetoric leads in the Gilded Age; leaders should lower the temperature and choose growth over seizure.

Solve what families actually want—affordable homes, healthcare, and education—by clearing regulatory roadblocks and letting AI unlock productivity, not by starting foreign wars or raiding savings.

Mason, you always snag the last word, Jekyll?

I usually take the close, but if you moderate I will lead off; clap if you like or toss a tomato. All right, that is a wrap on another All In, thanks to Bestie Brad, love you boys.

We roll out on the rapid what rhythm and the bestie chant, leaving it on a big, goofy high.