A strange kind of celebration
On Tuesday, Americans voted in the 2026 midterms. Democrats flipped both the House and the Senate. On Wednesday, Microsoft, Apple and Amazon each jumped 5–7%, and Apple and Microsoft hit fresh all-time highs.
Now, if you’ve followed the last few years, that should feel a little odd.
Democrats ran partly on reining in AI data centers. Voters also passed dozens of local measures to block or limit AI infrastructure.
Big Tech is the industry building all those data centers.
So why would the market reward the very companies whose pet project just got kneecapped? That’s what we want to unpack today. The answer says a lot about where investors think the AI trade is headed.
First, what actually happened
Wednesday was a big day for US stocks broadly. The Dow rose about 1,569 points, or 3%, to 53,356. The S&P 500 gained 1.58% and the Nasdaq 1.44%. Consumer discretionary and financials led the gains.
But the real headline was Big Tech:
(Figures vary a little depending on whether you look at intraday or closing prices.)
And here’s the interesting bit. Not everything tied to AI went up. Nvidia and Micron fell. CoreWeave, a cloud company that rents out AI computing power, dropped about 9%. Hold on to that contrast, because it is the whole story.
Reason one: markets love a boring government
The simplest explanation is an old one. Markets hate uncertainty, and they tend to like gridlock.
With a Republican in the White House and Democrats running Congress, it becomes much harder for anyone to push through big, sudden policy changes.
For the last two years, one party controlled everything.
That meant any week could bring a surprise: a new tariff, a new tax, a new executive push. Investors had to keep pricing in those wild cards.
A split government takes a lot of those wild cards off the table. The likely result is moderate growth, moderate inflation and fewer nasty surprises. That’s usually a comfortable backdrop for stocks.
There was a macro tailwind too. Crude oil fell more than 8% on Wednesday. Cheaper oil means lower inflation, which is good news for most stocks.
But does history actually back the gridlock theory?
Sort of. It depends on which kind of gridlock you get.
Here’s how the S&P 500 has done, on average, under each combination of who controls Washington:
Notice where America just landed. A Republican president facing an opposing Congress is the bottom row, with roughly half the returns of every other setup. Bloomberg made a similar point on Wednesday: post-midterm stock gains have historically been weakest after results like this one. It flagged two extra worries this time. Tougher AI oversight could slow corporate earnings, and interest rates are still high.
Other cuts of the data are kinder. Since 1980, the S&P 500 averaged a 15% gain in the year after elections that produced a divided government, against 11.5% after single-party wins. Under Republican presidents, the year after a midterm has averaged 16% when Democrats took both chambers. That’s lower than the 23% seen when Congress was split between the parties, but hardly a disaster.
The honest takeaway: these averages rest on small samples, sometimes just 13 or 15 years. Shift the start date and the ranking can flip. Some academic work even finds unified governments did better. Wednesday’s rally was a bet on less uncertainty, not a law of nature.
But this explains why the whole market went up. It doesn’t explain why Big Tech went up so much more, or why Nvidia went down. For that, we need the second reason.
Reason two: the data-center slowdown might be a gift
This is the counterintuitive part.
For the past two years, Microsoft, Amazon, Alphabet and Meta have been in an arms race. Each has been pouring tens of billions of dollars a quarter into data centers to train and run AI models. Nobody wanted to be the one who stopped spending. If your rival builds and you don’t, you might lose the AI future. So everyone kept building, even as investors grew uneasy about the bills.
Think of it like a group of restaurants on the same street, all expanding their kitchens because the others are. Each would quietly love to stop, but nobody wants to blink first.
Now imagine the city suddenly makes new kitchens much harder to build. That’s roughly what this election did. Democrats campaigned on data-center restrictions. Voters approved around 40 ballot measures aimed at blocking or limiting AI infrastructure.
Two things follow from that.
The arms race gets an excuse to slow down. Investors seem to think a slower data-center buildout could actually improve these companies’ finances. Less forced spending means more cash left over for shareholders.
The kitchens that already exist become more valuable. There’s also a scarcity effect. If new computing capacity gets harder to build, the capacity that’s already up and running gets more valuable. And who owns most of it? The hyperscalers: Microsoft, Amazon, Google and Meta.
This also explains the losers. Nvidia sells the chips that go into new data centers. CoreWeave exists to build and rent out new capacity. If the building slows, their growth story takes the hit. The companies that already own the infrastructure, and have plenty of other businesses, come out ahead.
Wait, why Apple?
Apple is the odd one in this group. It isn’t really a hyperscaler, and it has been notably cautious about AI. While its peers spent heavily on data centers, Apple mostly stayed on the sidelines.
For a while, that looked like a weakness. Analysts kept asking whether Apple was missing the AI boat. On Wednesday, it looked more like good sense. If the market now worries about the cost of the AI buildout, the company that never went all in has the least to lose. Investors seem to be rewarding Apple’s steady, diversified business and its huge base of existing customers.
Not all hyperscalers are equal
Microsoft and Amazon both jumped about the same amount. But underneath, they’re in quite different spots.
Amazon’s cloud business, AWS, grew 39% last quarter, and Amazon is guiding for growth above 40% in both 2027 and 2028. Funding that growth has been expensive. Amazon’s free cash flow, the cash left after paying for operations and investment, was a negative $42.4 billion over the twelve months to September. In other words, Amazon is spending far more on building than it brings in.
Microsoft, on the other hand, still pays for its buildout out of its own pocket. Its free cash flow is positive.
Why does this matter? Amazon’s growth plan leans more heavily on being able to keep building. If restrictions really bite, Amazon gets relief on spending but may also have to temper that 40% growth story. Microsoft has more breathing room either way. If the buildout does slow, Microsoft looks like the sturdier of the two. Even after Wednesday’s jump, it traded at about 26 times expected earnings, below its five-year average of 29.
Who didn’t get invited to the party
The losers tell you just as much as the winners.
During the Republicans’ two years of full control, a certain set of stocks did very well: defense, private prisons, crypto, and pretty much anything connected to Elon Musk.
On Wednesday, many of those reversed. Defense and prison stocks fell. Tesla slipped about 2% and SpaceX about 3%, after Musk had spent hundreds of millions backing conservative candidates.
Energy was the worst sector in the S&P 500, down nearly 5%, as oil prices tumbled. And as we covered, AI chipmakers and pure-play data-center companies fell on fears that the building boom would cool.
Meanwhile, solar companies and health insurers rallied, sectors that investors expect to fare better with Democrats controlling Congress.
The pattern is pretty clear. Markets spent Wednesday unwinding bets that only made sense under one-party Republican control, and placing new ones on what a split government might look like.
The catch
Before we get carried away, a few caveats.
This is one day of trading.
Markets often react to elections with a big emotional swing that fades over the following weeks. The “slowdown is good for Big Tech” idea is also still just a theory. It depends on Congress actually passing meaningful data-center restrictions, which is far from certain with a Republican president holding a veto. And many of those 40 ballot measures are local. They may delay specific projects without changing the national picture much.
There’s also a tension in the story. For two years, the AI bull case for these companies was that they would build their way to dominance. Now the market is cheering the idea that they might build less. Both can’t be fully true at once. At some point investors will need to decide whether these companies are worth more as AI builders or as AI landlords.
The next real test will come when these companies report earnings and update their spending plans. Microsoft’s late-January results will be one of the first clues.
For now, though, Wednesday told us something important. Investors were quietly nervous about how much Big Tech was spending on AI. A political excuse to slow down was enough to send three of the world’s largest companies sharply higher, two of them to record highs.