PodBrowser
All In

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

Saturday, 7 November 2026 · 4 min read · Listen to the episode ↗

Democrats’ sweep of Congress and at least 29 governorships sparked debate over whether affordability pressures or the Iran war drove voters away from Trump, and what policy shifts could restore support. The AI-linked selloff erased nearly $4 trillion from Nvidia, SpaceX, Tesla, AMD and Micron, while local opposition and permitting uncertainty clouded data-center expansion. Speakers warned a construction slowdown could favor hyperscalers and constrain compute for outside developers and AI model companies.

Democrats swept both chambers of Congress and won at least 29 governorships, with Alaska and Mississippi unresolved. They were projected to hold 54–56 Senate seats and at least 250 House seats, possibly 260, after Republicans lost ground across voter groups. Texas, Kansas, Iowa and South Carolina shifted blue despite backing Trump by double digits in 2024. Chamath called the result a broad rejection of economic conditions and the gap between asset owners and wage earners, not an ideological mandate; he noted Democrats won 11 of 15 redrawn Republican seats while Republicans won one. Sacks instead argued that the Iran war and its economic effects were the main drivers, a view Chamath disputed.

Affordability was the common thread in the debate. Sacks said inflation fell to 2.4% in January from 9% under Biden, then rose to 3.7%, largely as energy and transport costs climbed. He cited gasoline rising from about $3 to more than $4, higher diesel prices and mortgage costs around 7.5%, even as post-tax household income rose about 3.1% in Trump’s first year. Chamath said growth had not broadly reached wage earners, while Friedberg argued that inflation lifted asset values and squeezed people without assets. Sacks said the war was eroding purchasing power and made ending it Trump’s top priority; he predicted Brent crude could fall from about $110 to $70 if the conflict ended, but warned Republicans would keep losing support if it continued.

Speakers differed on what Trump should do to recover politically. Sacks said survey respondents who considered the war crucial or a major factor in their congressional vote broke heavily for Democrats, and warned that escalation could repeat the pattern of George W. Bush’s response to the 2006 midterms, which preceded further Republican losses. He pointed to Bill Clinton and Barack Obama as presidents who adjusted course after midterm defeats. Others cited a longer sequence behind Trump’s approval decline, from about +5% at the start of his term to −28%, including tariffs, ICE enforcement, the Epstein files, inflation and grifting allegations; Sacks disputed tariffs as the trigger for the Senate shift. Proposed economic steps included health-care and housing deals, a $15 minimum wage phased in over four years, incentives to build homes, and eliminating federal income taxes for people earning $125,000 or less. One proposal would tax loans against equity holdings above $10 million at 50% of the capital-gains rate, with a credit if the loan is repaid. The speakers also discussed barring officials from stock trading or foreign business dealings.

The AI-linked selloff wiped close to $4 trillion in market value from Nvidia, SpaceX, Tesla, AMD and Micron. Speakers called it the largest single-day loss in total dollars, but not by percentage. Markets had rebounded somewhat, but the decline clouded prospective Anthropic and OpenAI IPOs; Polymarket’s odds of an OpenAI public offering this year fell to 71% from 95% before the midterms. The discussion treated data-center construction as pivotal to whether the downturn deepens. At a presentation for a proposed 6-gigawatt SpaceX data center in Cochise County, Arizona, representatives promised 40% lower electricity costs, no water use, 20% lower property taxes and enough revenue to repay a county bond; audience members shouted over the presentation. The dispute reflected concerns that communities bear costs while technology leaders capture gains, and speakers said governors may demand measurable local benefits rather than approve tax incentives without conditions.

A slowdown in construction could advantage hyperscalers with existing capacity, including Amazon, Meta, Microsoft, Google and SpaceX, while hurting neoclouds that rent capacity and model companies that need more compute. Hyperscalers could reserve new capacity for their own products, raising costs for outside developers and potentially degrading user experience. Speakers said Anthropic and OpenAI improve models on a 60-to-90-day cycle, so compute constraints could give competitors time to catch up. Sacks put gigawatt-scale data centers at about $50 billion each and said uncertainty over permits, labor actions and local opposition makes investment difficult. He expects meaningful spending in 2027, mainly on fully permitted projects, but a sharply smaller pipeline by 2028. Restrictions could shift construction to Canada, Mexico, Singapore, Malaysia or the Middle East, taking tax revenue and development with it; speakers warned the United States could become a net importer of tokens.

The scale of possible AI restrictions remained disputed: one speaker put the chance of extreme restrictions above 10%, citing state laws, while Jake put it near 1%. Sacks argued that restricting software raises free-speech questions, while Freeberg warned that state control of AI could concentrate power and weaken individual liberty. Sacks expected a major Democratic push for federal AI legislation, possibly including a department empowered to review frontier models; Chamath predicted a regulatory bargain that could give government more power while protecting closed-model companies. Jake said generative AI adoption could drive compute demand to 500–1,000 times current levels, with only 10–20% of corporate America using the technology regularly today. Freeberg emphasized that the risks he described were possibilities, not certainties, and that policy should account for them as they develop.

This summary was generated from the episode transcript and can contain mistakes.