AI Firms Expand as Government Regulation Lags

AI companies are pressing ahead with expansion as safety researchers warn that the technology could escape human control and government regulation may come too late. The New York Times convened five experts to assess the risks and possible government tools; researchers who left OpenAI and Anthropic warned that the companies could be courting disaster.

Warnings from inside the industry

The companies were reportedly gaming out scenarios involving a catastrophic cyberattack or another event in the next six to 12 months. According to the reports, corporate executives and government officials still make the decisions shaping AI, while people exposed to the consequences have no direct role in them.

The New York Times said regulation would pose serious challenges for any Congress and White House. President Trump undermined prospects for bipartisan regulation when he called predictions of an AI catastrophe a “hoax” in September. Yoshua Bengio, described as a scientist and godfather of AI, wrote in October: “I do not believe we are on track to see sufficient government regulation in time.”

The experts discussed the issue by video conference. The New York Times edited and condensed the discussion for clarity, and reordered or added some material from follow-up interviews. The reports offer no grassroots, self-organized or mutual-aid response. Government regulation remains the proposed answer, even as the reporting describes a political barrier to passing it.

Expansion, borrowing and market nerves

The Financial Times headline said Meta planned a $95 billion stock offering to fund AI expansion. The article text available in the fetch consisted of the headline and subscription material, with no further details about the proposed offering. The headline makes the scale clear, but the available reporting provides no additional account of the plan.

CNN reported that markets appeared stable on the surface: stocks sat near record highs, and bonds hovered around levels seen over the previous couple of months. But investors were showing signs of nervousness, largely over AI. On Oct. 22, the market-value-weighted S&P 500 fell 1.2%, while the equal-weighted index gained 0.5%.

James Reilly, senior markets economist at Capital Economics, said that divergence had happened only twice before: April 14, 1999, and March 30, 2000. That same Thursday, the 10-year Treasury yield fell below 5% in its largest one-day decline since March 10, 2023, the day Silicon Valley Bank collapsed. Investors shifted money out of technology and into Treasuries and nontechnology stocks.

CNN attributed rising bond yields to Iran, U.S. debt and AI. The Iran war disrupted global oil and fuel supplies, driving prices higher and reviving inflation; bond investors demanded higher yields to compensate. U.S. debt had climbed to $40 trillion, and the budget deficit reached $2 trillion in the previous fiscal year. Politicians across the political spectrum continued to advocate lower taxes and higher spending, keeping Treasury borrowing needs high.

Who carries the financial risk

AI investment supported U.S. economic growth, but the infrastructure needed to power AI cost about $1 trillion a year and kept growing. Technology companies borrowed heavily through bond issuance, competing with demand for other debt, including Treasuries. Oil prices stood at four-year highs, fuel prices were rising sharply in parts of the United States, and President Donald Trump had indicated the costly Iran war could intensify after the midterms.

CoreWeave, a neocloud company that rents computing power to AI companies, emerged as a possible turning point. The debt-funded company saw its borrowing costs reach 15% earlier in October, exceeding the return on its investments, as investors questioned its finances. CNN said CoreWeave was losing substantial amounts of money and might not repay its debt.

A Wall Street Journal report on Anthropic's draft initial public offering filing showed a loss of nearly $81 billion in the first half of 2026. If the company went public that year, it could exceed the largest annual loss reported by a public company. Meta said it spent $2 billion more cash than it generated in the third quarter, and its stock fell 7% on Thursday.

Joe Brusuelas, chief economist for RSM US, said: “What we're seeing is elevated volatility around something that looks and feels and smells like a bubble.” CNN said the Nasdaq was in a 6% slide, while investors had worried during previous earnings seasons about AI spending before returning to AI stocks when some companies reported sustained demand. Reilly said an AI bubble bursting wouldn’t necessarily bring down the economy. In recent years, economists had described the economy as a “rolling recession,” with booming and declining sectors offsetting one another. CNN said it remained unclear whether the market moves marked a broader shift or a temporary setback.