AI stocks face $800 billion capex test on guardrail calls

AI stocks came under pressure after executives urged slower development, testing a rally built on nearly $800 billion in 2026 capex.

The caution followed weekend remarks from AI leaders, including Anthropic CEO Dario Amodei, who called for a slower pace of development to give companies and policymakers more time to manage safety risks. The comments landed after warnings from industry figures about the technology’s potential danger to humanity.

The timing matters for a market rally closely tied to AI spending. The S&P 500 has more than doubled since the bull market that began in October 2022, and the broad index is up more than 11% this year, with investors pointing to profit growth and earlier AI outlays as support.

$795 billion capex test

Wall Street is watching whether safety talk turns into canceled orders. “This becomes a problem if in fact you see orders being cancelled, you see data centers, construction deals being cancelled,” Chuck Carlson, chief executive officer at Horizon Investment Services, said, adding that he needed evidence of a slowdown “versus just talk.”

The spending base is large enough to move more than one sector. Microsoft, Alphabet, Amazon, Meta Platforms and Oracle are expected to spend about $795 billion on capital expenditures in 2026 and nearly $1.08 trillion in 2027, according to BofA Global Research.

That would put the 2027 estimate about $285 billion above the 2026 figure. Much of the money is directed toward data centers, chips, power equipment and networking infrastructure, making suppliers more exposed if hyperscaler plans are delayed.

Semiconductor shares bore more of Monday’s selling than the hyperscalers, according to the market moves cited by investors. The Philadelphia SE Semiconductor Index is still up nearly 60% in 2026, leaving the group with a large gain even after the latest decline.

Chip suppliers take pressure

Erik Kratz, chief investment officer and co-head of wealth at Arena Private Wealth, attributed the sharper move in supplier shares to their dependence on continued capability gains. “Markets are punishing the picks-and-shovels layer harder than the hyperscalers because it's the layer most exposed to a slowdown in the rate of capability improvement,” he said.

Kratz also said safety rules could make long-term spending easier to defend if they give investors clearer limits around AI development. In that scenario, guardrails would not end the buildout; they would change the terms under which boards, lenders and public shareholders assess it.

A voluntary pause by OpenAI or Anthropic would carry a different market test. If the leading developers eventually sell shares to the public, investors are likely to press them for growth, Carlson said, making any self-imposed speed limit relevant to future valuation debates.

Regulation joins Fed risk

The latest AI doubts recalled the early 2025 market reaction to China’s DeepSeek model, which briefly challenged assumptions about infrastructure spending. That earlier selloff proved short-lived for the AI trade, but investors are now weighing regulation, geopolitics and possible new technology models at the same time.

President Trump on Monday called AI safety threats “a hoax” and played down the need for regulation. Kratz said the larger market risk was not slower development itself but “regulatory overreaction,” a concern that would put policy timing beside corporate capex as a driver of the trade.

Michael Bruun, global co-head of private equity at Goldman Sachs Alternatives, said investors were assessing whether regulation, geopolitical tension and technology changes would become more or less favorable. He also said he still saw openings to deploy capital “into the epicenter of AI.”

The macro backdrop gives the debate less room for error. The S&P 500 and Nasdaq Composite are both about 2% below all-time highs, while investors also face higher bond yields, rising oil prices and the possibility that the Federal Reserve raises interest rates this week to restrain inflation.

Orders decide the path

If the $795 billion 2026 spending plan holds and guardrails remain largely voluntary, the global effect would be continued capital demand for chips, data centers and power infrastructure. For the largest AI developers and hyperscalers, the mechanism would be order continuity; for suppliers, it would preserve revenue visibility.

If safety calls translate into canceled construction deals or chip orders, the pressure would move first through semiconductor and infrastructure stocks. The broader market would then lose part of the growth engine that has offset rate and inflation concerns, while OpenAI and Anthropic would face harder questions before any future public listing.

If governments impose wider restrictions, the key mechanism would shift from technology performance to compliance cost and approval timing. The main open question is whether industry guardrails stay inside company planning or become binding rules that slow spending across the AI supply chain.