Microsoft Loses $360 billion as AI Spending Spooks Investors

For months, Microsoft has been treated as the safest way to own the AI boom, the company with the deepest cloud infrastructure. This week, one part of that story started to look less like dominance and more like a bill coming due.

On January 29, Microsoft shares slid sharply in the session after results, a move that quickly spread beyond one ticker and pulled the broader market lower.

The reality is that the panic was not triggered by weak revenue. It was triggered by what it costs tokeep the AI machine running.

The number that spooked Wall Street 

Microsoft’s fiscal second-quarter release showed real momentum. Revenue rose 17% to $81.3 billion, and net income jumped 60% to $38.5 billion for the quarter ending December 31, 2025.

The fear sat in the spending line. Capital spending hit $37.5 billion, up nearly 66% year over year, with much of it directed toward computing chips and the data-center buildout needed for AI.

That single figure reframed the quarter. Investors are no longer grading “AI leadership” as a vision story. They are grading the timing of returns. 

The Financial Times estimated the selloff wiped about $360 billion in market value.

Azure growth was solid, but the margin math shifted 

The cloud engine did not collapse. Azure and other cloud services revenue grew 39% in the quarter.

The issue was the pairing. When a company is spending at record levels, the market stops grading growth in isolation and starts grading the spread between growth and cost.

Bloomberg described the move as Microsoft’s biggest intraday slide since March 2020, a signal that investors are treating AI capex as a near-term earnings risk, not a distant strategic investment.

OpenAI goes from advantage to concentration risk 

The most revealing detail was not just that Microsoft’s backlog climbed. It was how concentrated the future commitments appear.

Microsoft said its long-term contract backlog rose 110% to $625 billion. On the earnings call, the company also disclosed that about 45% of that backlog is tied to OpenAI.

That does not mean demand is weak. It means the story now includes a new kind of risk: Microsoft is spending like the platform for everyone, but a meaningful slice of the narrative is still anchored to one AI heavyweight.

Copilot is the monetization signal, but it’s not the full offset 

Microsoft also highlighted a metric meant to show this is not only an infrastructure story. Management said Microsoft now has 15 million annual users for Microsoft 365 Copilot, the $30-per-user-per-month AI add-on for business customers.

That does not automatically translate into a clean revenue run-rate because “annual users” is not the same as “paid seats.” But it matters because it is one of the clearest indicators that Microsoft is trying to monetize AI inside its existing software base, not just by selling more compute.

Why this hit the whole AI sector 

This was treated as an AI market signal, not just a Microsoft signal. The spending surge revived investor anxiety about how quickly massive AI budgets translate into profit, even when demand remains strong.

In the same window, Meta’s results showed the market’s new rule: investors can tolerate heavy AI investment when the core business is accelerating and guidance feels clean, but they punish it when spending becomes the headline.

The real story is no longer whether Microsoft is all-in on AI. That part is settled. The story is that Wall Street has stopped paying for ambition on faith.