The biggest names in America's AI boom gathered at the San Francisco opera house on election night for a show. Over wine, executives in tuxedos buzzed about an emerging threat, people there said. The results coming in ran against the massive build-out of computing capacity that Silicon Valley is counting on. By Wednesday morning, senior leaders at one AI company were discussing whether they'd have to reset their data-center plans for next year.
Across the country, voters backed candidates and local measures to slow or pause data-center development. More than a dozen Ohio towns voted to halt or restrict new projects. Similar bans passed in towns in Kansas, Florida and California. About 56% of voters nationally said data centers were a crucial or major factor in their vote for Congress. The concern is concrete: the power and water these centers draw, and the strain on local grids. Texas's incoming governor campaigned against the state becoming a "wild west" for data centers. "It stops when I'm governor," she said.
Why it matters to the market
This is more than a local-zoning story because the build-out has been carrying the whole economy. AI investment is projected at about 1.9% of GDP this year, the biggest share of growth for a new industry since the railroad boom of the 1800s, Goldman Sachs estimates. LinkedIn data ties some 750,000 U.S. jobs since 2023 to AI. Texas, Ohio and Georgia alone account for more than 40% of planned U.S. data-center capacity. All three elected data-center critics. One venture investor estimated that 25% to 35% of next year's build-out could be at risk. Shares of data-center companies fell Wednesday even as the broad market soared: CoreWeave slid nearly 10%, and another builder fell almost 7%.
There's a flip side some investors noted. If less gets built, the giant companies that already own data centers can charge more for the computing they have, and the fear of reckless overspending eases. That's part of why Microsoft, Alphabet and Amazon rose while the suppliers fell.
Our read
This is an Investments and Risk story, module IN04, and it names a real risk in plain terms. For two years, a handful of AI and chip names and the companies that feed them have driven the market. Now voters have put a brake on the thing those companies were counting on. That doesn't mean the AI trade is over. It means the easy, unchecked version of it just got harder, and the stocks priced for unchecked growth are the ones at risk.
Here's why it matters for your own money even if you've never bought a single tech stock. A regular S&P 500 index fund is heavily weighted toward these names. When they fall, as they did Wednesday while almost everything else rose, your fund feels it. The check this week is the same one we keep coming back to, because this week keeps proving why. Pull your largest fund's fact sheet and read the top-ten weight. If a big share of your retirement rides on the AI build-out continuing at full speed, you now know the political weather turned against that. The fix isn't to sell in a panic. It's to know your exposure and decide, calmly, whether it's more than you meant to have. In our model portfolios we hold these names at their normal weight. This week we trimmed CoreWeave to 10% from 16% in its one sleeve and added nothing. If you want to see how concentrated your own funds are, that's fifteen minutes with your statement.