LIVE 24/7
Próxima actualización en --:-- EN中文ES

Midterm Blood Bath, $4T AI Sell Off, Data Center Freeze, What's Next for Trump

All-In Podcast · 2026-11-07 · 1:33:03

Presentadores: Chamath Palihapitiya · Jason Calacanis · David Sacks · David Friedberg

De qué trata este episodio

Recorded the Friday after the 2026 midterms, the four hosts first dissect a Democratic sweep of both chambers of Congress, then give the second half to the AI trade: a single-day selloff that erased close to $4 trillion in market value, a growing local backlash against data centers, fears of state-level AI bans — and a closing Science Corner on OpenAI's claimed disproof of the Exponential Time Hypothesis.

Puntos clave

A midterm rout the hosts blame on war and affordability

The panel's consensus: the GOP's generational loss was driven less by ideology than by the Iran war and its economic fallout — gasoline up from about $3 to over $4, inflation climbing back from 2.4% to 3.7%, expected rate cuts turning into a hike that pushed mortgage rates to roughly 7.5%. Friedberg also flagged a fractured base: the anti-war podcast hosts who helped win 2024 stayed home or cast protest votes.

Trump's comeback playbook, as pitched on the show

Calacanis proposed a deal-making reset: pardon broadly, tax margin loans above $10M, phase the minimum wage up to $15, ban stock trading by officials, and build housing — and above all, end the war. Sacks added wiping federal income tax on incomes under $125K. The group also reached for history: Clinton (1994) and Obama (2010) recovered from midterm routs, while Bush doubled down after 2006 and his party lost again in 2008.

Close to $4T wiped out in a day — and IPO math changes

Calacanis called it the worst single day in stock-market history measured in total dollars: NVIDIA, SpaceX, Tesla, AMD and Micron all took a massive haircut. The chill reached the IPO pipeline — Polymarket odds of Anthropic going public this year dropped from 95% to 71%, and the hosts debated AI IPO valuations resetting from the $1.5–2 trillion range toward roughly $1 trillion.

Data-center freeze: hyperscalers win, neoclouds and frontier labs lose

Sacks described SpaceX's pitch for a 6 GW data center in Cochise County, Arizona, being shouted down by locals, and Chamath predicted the era of grand tax-break deals is over. Their game theory: constrained supply turns the hyperscalers — Amazon, Meta, Microsoft, Google, SpaceX — into kingmakers. Sacks pointed to the SpaceX–Anthropic compute deal's mutual 90-day termination clause as a possible repricing trigger worth 2–3×, and Chamath expects permitted projects to keep 2027 capex meaningful but the pipeline to be closed or significantly reduced by 2028. Friedberg warned the backlash could go further: hundreds of state AI laws are already forming, and Chamath put a major Democratic push for AI legislation in the next Congress at 100%.

Science Corner: OpenAI's claimed disproof of the Exponential Time Hypothesis

Friedberg walked through OpenAI's 175-page paper claiming a sub-exponential-time algorithm for 3-SAT — which, if it holds, disproves the Exponential Time Hypothesis proposed in 1999 and reopens the P vs NP question. The hosts stressed the proof is still being evaluated and has not been verified by third parties, and marveled that the work reportedly took days rather than decades.

En sus palabras

“Close to $4 trillion in market cap wiped out.”

— Jason Calacanis · 53:13

La lectura: The through-line of this episode: AI's bottleneck is no longer chips or models — it is politics, from county zoning hearings to state legislatures, and this week the market priced that in.

Escuchar el episodio original ↗

Esta columna es un resumen y comentario. Todos los derechos pertenecen al programa original.