Dan Ives predicts midterm results could derail some AI data center plans

Dan Ives has spent years as one of Wall Street’s loudest AI cheerleaders. Now he’s flagging a speed bump, and it came from the ballot box.

Ives expects the November 4, 2026 midterm results to slow AI data center construction across the US. He also expects the peak in Big Tech’s capital spending growth to arrive sooner, in 2027.

What changed on election night

Democrats took control of both the House and the Senate in the midterms. In Texas, Democrat Gina Hinojosa defeated incumbent Republican Governor Greg Abbott.

Ives, an analyst at Yorkville Ives & Co., did not mince words about the outcome.

He called it a “seismic shift” for US AI infrastructure.

His estimate is specific. He believes 5% to 10% of the 800 to 1,000 new data centers planned in the US may not be built as expected. That works out to approximately 40 to 100 facilities.

Ives ties the slowdown to heavier regulatory scrutiny and political resistance. Those facilities were meant to serve demand from generative AI workloads, the computing muscle behind chatbots and image generators.

Fewer facilities coming online means less to spend on. Hence his view that Big Tech’s capex growth will hit its high point in 2027 rather than continuing to stretch further out.

Why Texas is the pressure point

The Texas result may sting most. The state accounts for 25-30% of the nation’s anticipated data-center electricity capacity through 2030.

Texas had long been seen as a magnet for these projects, thanks to affordable power and available land.

Hinojosa campaigned against the proliferation of data centers.

The resistance isn’t confined to Texas. Political measures aimed at restricting data center expansion passed in multiple states on November 4, 2026.

A 2026 CBS News poll found that more than twice as many Americans oppose new data centers in their communities as support them.

Ives isn’t alone

Brad Gerstner of Altimeter Capital has voiced similar concerns. He has pointed to 10-15 gigawatts of expected data center capacity now at risk.

Markets reacted quickly. Shares of data-center and AI-related companies declined on November 4-5, 2026, as results came in.

Still bullish, with an asterisk

None of this has turned Ives into a bear. He continues to argue that the economic benefits of AI adoption are underestimated.

His concern centers on the pace of physical buildout, not on whether AI demand is real.

What this means for investors and the industry

A 2027 peak in capex growth also carries a subtle message. Peak growth doesn’t mean spending stops. It means the rate of increase starts to flatten.

A Democratic Congress plus state-level ballot measures creates pressure from both the federal and local directions at once.

The CBS News polling suggests this isn’t a fleeting mood. When opposition outnumbers support by more than two to one, elected officials have little reason to roll out the welcome mat.

Watch Texas closely. How the Hinojosa administration translates campaign rhetoric into policy will shape a large share of the planned electricity buildout through 2030.