The US midterms dominated the week. Democrats won both chambers of Congress and voters in dozens of localities backed limits on data centers, adding siting and financing risk to the AI infrastructure buildout. Elsewhere, Sequoia reshaped its operations, late-stage valuations stayed high, and big tech companies adjusted how they buy and spend on AI.
Midterms raise the stakes for AI infrastructure
Democrats' sweep of Congress puts AI companies and data centers in the spotlight. Democrats won both chambers and signaled hearings for AI executives and stricter data center rules. Yorkville Ives analysts expect slower progress on roughly 800 to 1,000 planned US data centers, so infrastructure backers may need to plan for longer timelines.
Voters in 30-plus US localities back data center bans and moratoriums. Voters in more than 30 US localities across nine states backed data center bans and moratoriums. If lenders price this local resistance into credit decisions, developers could face higher costs or tighter terms.
Midterm results show voter backlash against AI data centers spanning both parties. Candidates opposed to data center expansion won in several states, and the backlash crossed party lines. Goldman Sachs and Morgan Stanley forecast delays, and proposals such as Sen. Martin Heinrich's GRID Savings Act could change project economics.
AI companies court Democrats as midterm results shift power in Washington. AI companies are courting Democrats, with OpenAI hiring a senator's daughter and renaming its data centers. A post-election sell-off and reported hesitation over Anthropic's IPO show how political uncertainty can feed into valuations.
Venture firms and late-stage capital
Sequoia shuts London office, Arc and Scouts while raising about $10 billion. Sequoia is closing its London office and ending Arc and Scouts while raising about $10 billion. Founders outside the Bay Area lose two early entry points, and the firm appears set on fewer, more concentrated bets.
Revolut valued at $115B in secondary sale as it readies a US bank launch. A secondary sale valued Revolut at $115 billion, 53% above last year's sale, as it prepares a US bank launch. The deal gives existing holders liquidity well before any IPO and shows continued appetite for late-stage fintech.
Humanoid robot maker Agility lists on Nasdaq via SPAC, raising $465 million. Humanoid robot maker Agility listed on Nasdaq via a SPAC, raising $465 million. Its $174,000 in first-half revenue against a $362 million loss puts pressure on converting its $300 million Digit 5 pipeline.
Big tech recalibrates its AI bets
Nvidia's Jensen Huang apologizes to staff as board opens independent probe. Nvidia chief Jensen Huang admitted infidelity in a letter to staff and backed a board-led inquiry into national security concerns raised in media reports. The scope and outcome are unknown, leaving open questions for partners building on Nvidia.
AMD guides to about $17 billion in Q4, poised to top Intel's quarterly revenue. AMD guided to about $17 billion in fourth-quarter revenue, putting it on track to top Intel, after data center sales nearly doubled to a record $8.55 billion. The results tie more of AMD's growth to a handful of large AI buyers.
Meta buys Conductor, a startup that runs parallel AI coding agents. Meta bought Conductor, a startup that runs parallel AI coding agents built by Anthropic and OpenAI, and its cloud service shuts on November 20. The deal points to large platforms treating developer orchestration tools as part of their AI push.
Apple trims daily AI spending allowance for engineers, with deeper cuts planned. Apple reportedly cut engineers' daily AI spending allowance to $200 in October, with some limits possibly falling to $100. Founders selling usage-priced AI tools may face tighter enterprise budgets and closer scrutiny of returns.
Uber says third-party AI agents booked just four rides in the third quarter. Uber disclosed that third-party AI agents booked just four rides in the third quarter. The figure suggests agents have not yet come between platforms and their customers, and large platforms may resist ceding that interface.
The takeaway
Political and local opposition is becoming a measurable cost for AI infrastructure, so siting, financing and timelines now carry more risk than a few months ago. At the same time, capital is concentrating in fewer, larger bets, and even well-funded buyers are putting firm limits on AI spending.
The briefing
The week in tech, in one email.
The dispatches worth your time, every week. Unsubscribe anytime.