The numbers from the latest round of tech earnings are hard to grasp. Meta, Microsoft, Alphabet and Amazon spent about $209 billion on capital expenditure in a single quarter, according to an analysis by the Thesis Rationale newsletter, most of it on data centers, chips and power for artificial intelligence. The question now is who ultimately pays for it.

Start with the companies themselves. Amazon's free cash flow over the past 12 months was negative $42.4 billion, compared with positive $14.7 billion a year earlier, as it pours money into AI. Alphabet reported negative free cash flow of $9.1 billion for the quarter, its second in a row, and Meta's turned negative at $2.1 billion. Amazon chief executive Andy Jassy said 2027 spending will be meaningfully higher than this year's roughly $220 billion.

The case for the spending

Supporters argue the demand is real. Google Cloud revenue more than doubled, Amazon Web Services grew 39% and Microsoft said customer demand still exceeds supply. If AI keeps finding paying customers, today's spending could look as smart as early investment in the internet itself.

Sceptics see warning signs. Investment is running well ahead of the cash it produces, and the bet depends on AI revenue growing for years. Markets are nervous: after the 3 November midterm elections, chip stocks fell on concerns about stricter data center rules, and the Nasdaq-100 closed about 10% below its early October high, according to Charles Schwab.

Apple offers a striking contrast. Its capital spending for fiscal 2026 was about $9 billion, roughly 1.9% of revenue, while it returned more than $100 billion to shareholders. Apple largely relies on others for heavy AI computing, but it is still feeling the effects. It expects memory costs, which have climbed amid the AI build-out, to keep rising into next year, it raised some iPhone and MacBook prices, and Reuters reported that Apple plans higher iCloud+ prices and usage limits on some AI features.

That is where the story reaches ordinary people. The AI boom is driving up the cost of memory chips and electricity, and those costs tend to flow through to device prices and subscriptions.

The fair conclusion is that both sides have a point. The spending is supported by real demand, but it carries real risks. For consumers, the practical takeaway is to expect AI features to keep arriving, and to expect more of them to come with a price tag.

Photo: ​中華民國總統府, CC BY 2.0, via Wikimedia Commons