What the Midterms Mean for the Data Center Build-Out

Democrats won control of both the House and Senate last night and flipped governorships in states including Texas, Georgia, and Ohio. For AI, it means the data center build-out just got harder.

We estimate 2028 earnings could fall about 6% to 12% for contractors, but only about 2% to 4% for chipmakers. Contractors are already down nearly 15% today on average, so the market has moved quickly to price in that hit, and then some.

Then there’s Big Tech. Amazon (NASDAQ:AMZN) is up about 7%, Microsoft (NASDAQ:MSFT) about 6%, Alphabet (NASDAQ:GOOGL) about 4%, and Meta Platforms (NASDAQ:META) about 3%. On the surface, that makes no sense. These are the companies building the AI future, and the election supposedly just made that harder.

But look closer, and the logic is pretty simple. Investors have been punishing these companies for their massive AI spending. Amazon, Alphabet, and Meta have all seen free cash flow turn negative, and they’re increasingly tapping debt and equity markets to fund their plans. So,the Street’s bet today is that Washington and the states just gave them a reason to slow down.

We call this the “capex cover” trade, and we think it goes too far, for three reasons.

First, a delay isn’t a cut. Say the big four push $50 billion of next year’s roughly $1 trillion of planned spending out by a year. At a 10% discount rate, that’s worth only about $5 billion in today’s dollars, by our math. Yet these four stocks added about $620 billion of market value today. For cash flow alone to justify that, the market would need to believe in a lasting cut. Our base case is a delay. Much of today’s move also looks like plain old relief after a brutal few weeks; Alphabet, for instance, is still about 14% below its 200-day moving average.

Second, nobody has actually said they’re cutting. Our rule here hasn’t changed: follow what they do, not what they say, and certainly not what the market guesses. Last week, Amazon projected AWS growth above 40% in both 2027 and 2028, with capacity effectively sold out. Alphabet said 2027 capex will “increase significantly.” You can’t deliver that kind of cloud growth without building the capacity to support it.

Third, there’s a better explanation for the rally. If it gets harder to build new data centers, the ones already built, permitted, and powered become more valuable. That’s a real tailwind for the hyperscalers, and it doesn’t require anyone to cut a dollar of spending.

That said, we aren’t dismissing the risk. The market just showed corporate America that it will reward spending restraint, and incentives matter. A delay pushes supplier revenue out, but a real capex cut would remove it. So, this has gone from a remote risk to a live one, and we’ll be watching every hyperscaler comment on 2027 and 2028 budgets very closely. So far, though, the evidence still says the spending is coming.