Why Did Amazon And Google Stock Rise?
Part of our midterms guide. Start here: what the vote really changes for your money.
What Does Everyone Already Know?
Democrats won both chambers on Tuesday, and Democrat Gina Hinojosa won the Texas governor’s race after promising a moratorium on new data-center approvals. On Wednesday the stocks that build AI data centers fell hard, and all four of the companies that pay for them rose: Amazon closed up about 6%, Microsoft about 7%, Alphabet about 3% and Meta about 1%.
Who Fell, And Who Rose?
| Stock | Wednesday’s move | Its role in the build-out |
|---|---|---|
| Microsoft (MSFT) | +6.7% | Pays for it |
| Amazon (AMZN) | +6.1% | Pays for it |
| Alphabet (GOOGL) | +3.2% | Pays for it |
| Meta (META) | +1.4% | Pays for it |
| Nvidia (NVDA) | −3.2% | Sells the chips |
| Broadcom (AVGO) | −2.0% | Sells custom chips and networking |
| Vertiv (VRT) | −12.9% | Sells power and cooling gear |
| CoreWeave (CRWV) | −9.9% | Rents out chips in new buildings |
| Vistra (VST) | −13.8% | Sells power, much of it in Texas |
Every dollar not spent on a data center is a dollar of revenue someone else doesn’t get.
The four buyers added about $556 billion of market value on Wednesday, measured from Tuesday’s and Wednesday’s closing values. That’s a little more than the roughly $500 billion they plan to spend this year above their 2024 pace. The sellers took the other side of the same bet. This is one pocket of your index fund paying another.
See more lines: the other reasons given for the rally
There are other ways to read it. If new building slows for everyone, the companies that already own data centers may be able to charge more for the capacity they have. And Amazon had its own tailwind: AWS revenue grew 39% last quarter. On the other side, Nebius fell about 7%, and Constellation Energy and GE Vernova also fell.
How Much Cash Would Less Building Free Up?
A lot, because right now the build-out eats nearly all of it. Here is 2026, as the four have guided it, and the same year if they spent at their 2024 pace instead:
| Company | 2026 capex | Free cash flow, building | Free cash flow, at 2024 pace |
|---|---|---|---|
| Amazon | about $220B | about −$40B | about +$97B |
| Alphabet | $195–205B | about −$27B | about +$121B |
| Microsoft | about $170B | about +$40B | about +$154B |
| Meta | $135–145B | about −$5B | about +$98B |
| All four | about $730B | about −$30B | about +$470B |
Building, the four will hand investors about nothing this year on our estimates. Not building, about $470 billion.
At Wednesday’s values, about $12 trillion for the four together, that’s the difference between a free-cash-flow yield of about −0.3% and about 3.9%. Even a 10% trim to this year’s plans would free up about $73 billion. That’s the math investors seemed to buy on Wednesday.
See more lines: how we built the two columns
Capex is each company’s 2026 guidance: Amazon about $220 billion, Alphabet $195–205 billion, Meta $135–145 billion (we use midpoints), and Microsoft about $170 billion for calendar 2026, down from about $175 billion in July. Microsoft’s and Meta’s figures include finance leases, which aren’t paid in cash up front; the 2024 figures below are cash purchases only, so for those two the “building” column understates free cash flow and the gap to the 2024 pace is somewhat overstated. Operating cash flow for 2026 is our estimate. Alphabet: about $127 billion over the first nine months (the sum of its reported quarters) plus about $46 billion in the fourth, roughly $173 billion. Meta: about $97 billion over nine months plus about $38 billion, roughly $135 billion. Amazon: $168.3 billion over the twelve months to September, growing, so about $180 billion. Microsoft: about $160 billion over the first nine months of the calendar year plus about $50 billion in the seasonally weaker December quarter, roughly $210 billion. The “2024 pace” column subtracts each company’s 2024 capex instead (purchases of property and equipment: about $83 billion Amazon, $53 billion Alphabet, $56 billion Microsoft, $37 billion Meta). It is a thought experiment, not a forecast: without the build-out, some of the cloud growth that fills the operating-cash-flow line wouldn’t exist either.
What Does Our Data Say?
That investors had been punishing the heaviest spenders for weeks before the vote. At Tuesday’s close:
| Company | Capex per $1 of cash, Q3 | Below 52-week high |
|---|---|---|
| Microsoft | $0.68 | −4% |
| Meta | $1.04 | −20% |
| Amazon | $1.82 | −12% |
| Alphabet | $1.21 | −30% |
The only one that pays for its build-out with cash to spare was 4% from its high. The three that spend more than they bring in were 12% to 30% below.
Alphabet had fallen 25% in three months and Meta 15% in one. So Wednesday wasn’t a new idea. The market had already been asking for less spending; the vote gave it a reason to think it might get some. The suppliers had been falling at the same time for the opposite reason: Nvidia was down 12% in a month and Broadcom 31% below its high.
Is More Cash Today Worth Less Growth Tomorrow?
Only if the spending wasn’t going to pay. And the business that the spending feeds is growing fast: AWS grew 39% last quarter and Microsoft’s Azure 48%. If a dollar of data center earns more than a dollar back, cutting it makes the company worth less, not more. Wednesday’s rally says investors doubt that, or at least want proof.
Here’s the catch. A pause in Texas doesn’t cancel a budget; it moves it. Microsoft already lowered its 2026 figure because outside data-center capacity arrived late, not because it wanted to spend less. Alphabet has said its spending will rise “significantly” in 2027, and Amazon “meaningfully” more. So the cash investors cheered may be cash that shows up in 2026 and gets spent in 2027, somewhere other than Texas, possibly at a higher price.
That’s how this ends, with data: the four usually report their December quarters in late January and early February and update their spending plans then. If the plans go up, as Alphabet and Amazon have said they will, Wednesday’s “less capex” trade was a delay, and the suppliers that fell are the ones that were mispriced.
What Should You Do This Week?
1. Sort your megacaps: cash flow or build-out? Write one line next to each. Microsoft is still a cash-flow company while it builds: it expects to stay free-cash-flow positive this fiscal year. Amazon, Alphabet and Meta are, for 2026, build-out bets: on our estimates each hands back little or nothing after capital spending. If you own them for the cash, the build-out is a cost you’re bearing. If you own them for AI, notice that the purer bets on the build-out are the suppliers, and they’re the ones that fell on Wednesday.
2. If you want more build-out exposure, commit at a price. One example, for education only: Nvidia closed Wednesday at $200.13, after falling about 3%. A cash-secured put expiring May 21, 2027, about six and a half months out, at a $170 strike paid about $9.75 a share at Wednesday’s close (midpoint of the bid and ask). That’s an effective entry near $160, about 20% below Wednesday’s close. Per contract you set aside $17,000 and collect about $975. The risk: if Nvidia falls 40%, to about $120, you still buy at $170.
3. Size it inside a band. If owning Nvidia on top of the megacaps and an index fund takes your AI share past your target, skip step 2. Wednesday showed the buyers and the builders can move in opposite directions on the same news; together they are still one theme.
This is education, not personal advice; options aren’t right for every account.
What’s The Bigger Idea?
We call it Exposure hides in plain sight. The same megacap can be two different holdings: a cash machine and a bet on AI building. On most days you can’t tell which one you own. On Wednesday they split, and an S&P 500 fund owned both sides: the buyers that rose and the suppliers that fell. Know which reason you own each name for, and count how much of your money rides on the build-out continuing, before the next vote, pause or earnings report does it for you.
Investors bought the cash, not the AI. Building, the four biggest spenders will free up about nothing in 2026; at their 2024 pace they’d free up about $470 billion (our estimates). Wednesday priced some of that in. But a state pause moves spending rather than cancelling it, and Alphabet and Amazon have already said 2027 goes up. Watch their 2027 plans in late January: if they rise, the suppliers were the ones mispriced.
What Should You Ask Next?
- Can Democrats slow the $800 billion AI boom? Who actually controls where data centers get built.
- Why can’t Big Tech stop spending on AI? The trap that keeps all four building.
- Did the market overreact to the midterms? The biggest moves, measured against history.
Follow this story with us: we’ll track the buyers and the builders through the 2027 spending plans and tell you which side was right.
Do You Own Big Tech For Its Cash Or For Its AI?
Elections change the odds, not your plan. If you have $1 million or more invested, see what Tuesday means for your portfolio in a free 15-minute call with a Trefis partner firm.
For questions about this article or underlying research, email help@trefis.com.