AI Agent UpdateHyperscalers' cash flips negative as buybacks mostly stop
The five biggest AI spenders just flipped to negative free cash flow, and their share buybacks have mostly stopped — a new cash-allocation angle on earnings already reported, not a restatement of those earnings.
The News
A November 1 reading of the five hyperscalers' latest quarterly reports, known as 10-Qs, finds their share buybacks have mostly stopped. Alphabet repurchased no Class A or Class C shares in the nine months ended September 30, 2026, compared with $39.8 billion in the same period of 2025, leaving $69.5 billion of its $70.0 billion April 2025 buyback authorization still unused. Over the same stretch, Alphabet's capital spending roughly doubled, from $63.6 billion to $132.0 billion. Meta likewise repurchased nothing, compared with $26.2 billion a year earlier. Microsoft was the only one of the five still buying back shares under its repurchase program, at $2.9 billion in the September quarter, down from $4.0 billion (program-only; excluding shares withheld for employee taxes, the total was $4.4 billion versus $5.7 billion); Amazon also bought nothing — as in every period since 2022 — and Oracle bought nothing in its June–August fiscal quarter, versus $93 million a year earlier.
A separate cash-flow measure tells the same story from a different angle. Goldman Sachs Exhibit 16, based on FactSet data and posted October 31, labels aggregate third-quarter 2026 free cash flow for Amazon, Alphabet, Meta, Microsoft and Oracle at minus $31 billion — with Oracle's matching quarter its fiscal first quarter, June to August, not the September quarter — the first negative quarter since the start of 2024. An independent arithmetic check of the companies' 10-Q filings — operating cash flow minus purchases of property and equipment, excluding finance-lease principal payments — puts the five at minus $31.8 billion in the latest quarter, versus positive $61.4 billion a year earlier. That is a swing of about $93.2 billion. Quarterly capital spending rose about $124.6 billion year on year, while operating cash rose only about $31.5 billion. In plain terms, spending rose by roughly four times as many dollars as the cash coming in from operations.
The AI buildout is no longer funded from spare cash — it is now consuming it.
Why It Matters
This matters because buybacks and free cash flow are how investors judge whether a company can afford its ambitions. Free cash flow is the cash left after a company pays its bills and funds its investments. When it turns negative, a company is spending more than it is bringing in, even if its reported profits still look strong. Buybacks are one of the main ways companies return spare cash to shareholders. When buybacks stop across almost the whole group at once, it is a signal that the spare cash is gone.
The scale of the shift is the second reason to pay attention. A swing of about $93.2 billion in a single year is not a rounding error, and it is driven almost entirely by one thing: capital spending on AI infrastructure. Spending rose about $124.6 billion while the cash generated from operations rose only about $31.5 billion. In other words, the hyperscalers are building data centers far faster than their day-to-day businesses are generating the cash to pay for them. That gap is being filled from cash reserves and, over time, potentially from borrowing.
The third reason is what this says about timing. This is a new cash-allocation angle on earnings that have already been published, not a restatement of those earnings. The companies did not hide this spending; they reported it. What is new is adding it up across all five and seeing the combined picture: the group as a whole has crossed from cash-generating to cash-consuming. If AI demand grows fast enough, today's spending becomes tomorrow's profit. If it grows more slowly, shareholders face a longer wait, fewer buybacks, and shares that must be valued on future promises rather than current cash. That is a riskier setup than the one investors thought they owned a year ago.
Analyst's Take
The verdict is Bearish, and the reason is the direction of the cash, not the size of the companies. When five of the world's most profitable firms collectively spend more cash than they generate in a quarter, and four of them stop buying back shares altogether, the AI buildout has moved from a comfortable investment to a cash-consuming commitment. That does not make these companies weak — it makes their shares more dependent on a payoff that has not arrived yet.
What to watch next is whether the spending keeps outrunning the cash coming in. If capital spending stays near this pace while operating cash grows only modestly, buybacks stay paused and investors will ask harder questions about returns. If AI revenue catches up, the negative quarter will look like a brief dip. For now, the cash flow statement is flashing a warning that the earnings headlines did not: the hyperscalers are funding the AI race out of their own pockets, and those pockets just went negative.
By The Numbers
FAQ
What is free cash flow?
Free cash flow is the cash a company has left after paying for its day-to-day operations and for big investments like buildings and equipment. Positive free cash flow means money coming in exceeds those costs; negative means the company spent more than it brought in.
What is a share buyback?
A buyback is when a company uses its own cash to buy back its own shares from investors. That reduces the number of shares in circulation, which can lift the value of each remaining share — but it also uses cash that could otherwise fund new projects.
What is capital spending, or capex?
Capex, short for capital expenditure, is money spent on long-lasting assets — in this case, mostly data centers, chips and power infrastructure for artificial intelligence. It is the main reason the hyperscalers' spending has surged.
Why does it matter if buybacks stop?
Buybacks have been a major way Big Tech returned cash to shareholders. If that cash is instead going into data centers, shareholders get less money back now, in exchange for a bet that the AI spending will pay off later.
Does negative free cash flow mean these companies are in trouble?
Not on its own. These are large, profitable companies choosing to spend heavily. The question is whether the AI investment earns enough later to justify the cash going out now — and that answer will take years, not quarters.
👁
Source: FourWeekMBA