Big Tech's $724 Billion AI Bill Is Finally Making Investors Nervous

For most of the AI build-out, every extra dollar of Big Tech capex bought a higher stock price. Alphabet's Q2 print cracked that script: free cash flow turned negative for the first time since the company went public, capital expenditure guidance climbed to as much as $205 billion, and the stock dropped roughly 7% the week of the report even as the quarter itself beat consensus. The market is no longer rewarding the spend. It is starting to discount it.

The flip extends to the rest of the cohort. Google, Amazon, Meta, and Microsoft are projected to spend about $724 billion on capex in 2026 and nearly $950 billion in 2027, per analyst estimates compiled by Bloomberg. Through last year, each guidance raise got rewarded. This quarter, Alphabet's free cash flow went negative, the cash left over after the company paid for its AI build and its day-to-day operations. Investors took the print as a warning rather than a moonshot.

"People are really focused on capex, obsessed with it," one investor wrote. "It used to be the more the better, but now it is the less the better." The cohort is what turns Alphabet's print from a single-stock story into a four-day test. Microsoft and Meta report Wednesday. Apple and Amazon report Thursday. Each print is now a referendum on whether AI capex is producing returns fast enough to justify the cash burn, not just a quarterly growth check.

Alphabet's revenue climbed 24% year over year to $119.8 billion. Google Cloud grew 82% to $24.8 billion, and Search rose 17%. Operating margin held at 34%. On a P&L basis, the company is doing fine. The problem is below the line: capex of $195 billion to $205 billion in 2026 means the company is now spending more than it pulls in from operations, and it is funding the gap with cash, debt, and the mark-to-market value of its stakes in Anthropic and SpaceX, recorded as a roughly $98 billion gain in Other Bets this quarter. That line looks like strength on a headline. It is also the kind of line that does not repeat every quarter.

The other three hyperscalers face the same arithmetic, with different starting points. Microsoft is the closest parallel: its Azure capex has been the cleanest read on AI demand, and its free cash flow has compressed in step. Meta's capex narrative is tied to its Llama build-out and the cost of training compute for the next model generation. Amazon's is split between AWS server build and the Anthropic relationship, which is now partly Alphabet's exposure too via the mark-to-market line.

Apple is the visible outlier, and the most informative one. Its shares are up roughly 15% in July, on track for the best month in three years, by sitting out the arms race: Apple partners with model developers rather than building frontier models itself, and it sells the consumer hardware that runs on top of the other four's infrastructure. AI is still hitting Apple's costs. Memory-chip demand from the AI build has forced the company to lift MacBook and iPad prices. The exposure shows up on the cost line, not the capex line. That is the trade Wall Street is rewarding this month.

The harder question for the four-day stretch is whether the other hyperscalers' capex-to-free-cash-flow arithmetic looks like Alphabet's, a one-time negative print funded by balance sheet, or whether each company can show a path back to positive FCF before 2027. Investors are no longer paying for the spend; they are pricing the risk. Microsoft's Azure margin, Meta's next-model cost guidance, and Amazon's AWS growth re-acceleration will all be judged by that test.

Apple's print on Thursday is the cleanest counterpoint: the same AI demand, with none of the capex exposure. If the cohort's narrative holds, that contrast gets sharper with every quarter it lasts.