Dems Call for Massive Regulation on AI | The Tom Bilyeu Show

Tom Bilyeu said Democrats’ election gains increase the risk of nationwide data-center moratoriums and tighter AI regulation. He pointed to Texas and Ohio, major data-center development areas that he said flipped convincingly blue, and argued that expansion was an election issue in both states. Texas governor-elect Gina Hinajosa promised to halt new projects until lawmakers enact stronger protections, while Sherrod Brown campaigned against Ohio expansion. Bilyeu said that, “if memory serves,” 56% of voters identified data centers as a major or crucial reason for their vote.

The election, in Bilyeu’s view, gave investors already looking for a reason to reduce exposure a catalyst, rather than creating the AI sector’s underlying problems. The S&P 500 fell 4.36%, its largest one-day loss in raw dollars but not an unprecedented decline by percentage; the Nasdaq dropped 5.52% into correction territory, and the Dow fell more than 1,800 points. Nvidia lost $447 billion in market value and fell 9%, while Broadcom fell 10% and Oracle 7%. Six data-center builders dropped 17% to 23%, and the Magnificent Seven lost more than $1 trillion combined. Bilyeu described AI as real and likely to become more integrated into daily life, but said valuations had outrun the growth businesses could support.

He identified debt, delayed revenue, and power shortages as additional pressures on the buildout. Infrastructure companies and “neoclouds” are borrowing against revenue that has not yet arrived, and some firms, he said, rely on new borrowing to service existing debt. If expected revenue comes too slowly, companies may have to sell assets, putting further pressure on prices and making new investment harder to attract. Citing Ray Dalio’s view of bubbles, Bilyeu said a break becomes more likely when companies can no longer service debt or raise capital and must sell assets to meet obligations. He also said an AI energy constraint is expected in 2027, with chips reportedly being stored because power or data-center capacity is unavailable; Oracle, he said, had not secured power for a New Mexico project.

Bilyeu said Anthropic’s expected IPO valuation had reportedly fallen from about $2 trillion to $1.5 trillion and might reach roughly $1 trillion. Its reported quarterly revenue was $16.8 billion, up 45%, after more than doubling in the prior quarter. He said investors interpreted the slower growth as deceleration and questioned the previous valuation. In his view, an IPO below $1.5 trillion would signal that the AI boom may be over, while one below $1 trillion would indicate a worse-than-expected downturn. He also recalled that OpenAI and SoftBank announced a $500 billion infrastructure rollout a year earlier, and said Elon Musk had retweeted that they did not have the money. Bilyeu said he had “good authority” that SoftBank had well under $10 billion secured. A mystery seller was reportedly trying to sell $500 million in data-center debt; he speculated it could be SoftBank but said the sale mattered more than the seller’s identity.

A slower buildout could temporarily benefit Amazon, Microsoft, and Apple by limiting spending and supporting margins, Bilyeu said. But he warned that worsening debt stress could lead to company failures and consolidation, market declines, layoffs, and scrutiny of exposure in pensions, insurance, and private credit. A rapid unwind could trigger a recession, though he said its severity was uncertain; a private-credit implosion would make the downturn especially dire. He also contrasted China’s current disadvantage in chips with its long-standing advantage in power, underscoring that infrastructure constraints are not only financial.

Bilyeu treated rapid AI takeoff as speculative: a self-improving system making a major leap in hours seemed unrealistic to him, while substantial progress over days, weeks, or months was conceivable. He argued that the prospect of one country gaining a major investment and cyber advantage helps explain why neither the United States nor China will stop pursuing advanced AI. If the United States reached artificial superintelligence first, he speculated it might seek better negotiating terms rather than launch a “cyber nuke,” and said he hoped it would not impose draconian rule. He also argued that Republicans could blame a recession on Democratic policies, including slowing AI, if they could credibly claim the economy had been doing well beforehand.