The biggest, most valuable company in tech is having a month it would rather forget — and the reason is the very AI build-out that was supposed to power its next decade.
The move
Microsoft stock has fallen more than 20% in June, putting it on course for its worst month since December 2000, MarketWatch and Benzinga reported. The slide stands out: the broad market has been roughly flat to slightly higher over the same stretch, so this is a Microsoft-specific repricing, not a sector-wide rout.
Why investors are selling
In one word: capex — capital expenditure, the money a company spends on long-lived assets like data centers and chips. Bank of America estimates Microsoft's 2026 capital spending will approach $190 billion, as reported by MarketWatch, with spending up sharply year over year. The problem for shareholders is what that does to free cash flow — the cash left after capital spending, and the pool from which buybacks and dividends are paid. As capex surges, free cash flow falls, leaving less to return to investors.
AI infrastructure is also front-loaded: the data centers, GPUs and networking gear must be built and paid for years before the revenue that justifies them shows up at scale. That timing gap is the heart of investor anxiety — the spending is certain and immediate, the payoff uncertain and deferred.
The bull and bear cases
Wall Street is split. Bears argue the Street is over-estimating near-term AI revenue and that enterprise adoption remains a bottleneck; one analyst downgrade this month cited exactly that. Bulls counter that Microsoft is the best-positioned large-cap in AI, that its Azure cloud growth should reaccelerate later this year, and that the sell-off has left the stock cheap relative to its history. Most analysts tracking Microsoft still rate it a buy — a reminder that a sharp drawdown is not the same as a verdict on the business.
We don't make price predictions; the point is the debate itself. The fall reflects a question now hanging over the whole AI trade, not just Microsoft.
The bigger picture
Microsoft is the most visible test of a question dogging every hyperscaler: is the AI build-out arriving faster — and costing more — than the revenue it is meant to unlock? JPMorgan recently estimated global AI infrastructure spending on track for around $5.5 trillion through 2030, with the biggest cloud players spending hundreds of billions in 2026 alone, and judged the economics "holding, for now." Microsoft's June slide shows that "for now" is doing a lot of work: investors are willing to fund the build-out, but they are demanding evidence the returns are coming. The crucial near-term signal will be whether Azure growth reaccelerates in the second half, as Microsoft has promised — enough to close the credibility gap the stock has opened this month.
