Microsoft to Borrow $25 Billion as AI Spending Spree Shows No Signs of Slowing

The software giant is tapping its top-tier credit rating to help fund a $170 billion capital spending plan driven by demand from OpenAI, Anthropic, and other AI customers.

Microsoft is going to the bond market for $25 billion, the company announced Thursday, as it looks for more ways to pay for the data centers and chips powering its AI ambitions.

The timing isn't subtle. A day earlier, Microsoft told investors it's targeting $170 billion in capital expenditures for 2026, a figure driven largely by demand from AI labs Anthropic and OpenAI, along with other cloud customers. Building and equipping the data centers needed to serve that demand costs an enormous amount of money, and borrowing is one of the levers Microsoft has decided to pull.

It can pull that lever cheaply. Microsoft carries a AAA rating from S&P, the highest grade the credit agency hands out, which means lenders see it as about as safe a bet as a corporate borrower gets. That rating translates into lower interest costs on whatever debt Microsoft issues, which matters a lot when you're raising $25 billion.

According to a company statement, the proceeds will go toward "general corporate purposes," a catch-all that Microsoft said could include working capital, capital expenditures, stock buybacks, acquisitions, and paying down existing debt. The offering is expected to close November 4.

Why borrow when you're already profitable

Microsoft isn't short on cash. The company reported positive free cash flow on Wednesday, a notable distinction right now in big tech. Alphabet posted negative free cash flow for the second straight quarter. Meta said last week its free cash flow had gone negative. Amazon reported the same on Thursday. Against that backdrop, Microsoft generating more cash than it spends is a meaningful data point, one that suggests its AI investments are, for now, not outrunning the business.

So why borrow at all? Finance chief Amy Hood addressed that directly on the company's earnings call Wednesday. Microsoft has been leaning on leasing as a financing tool, she said, but hasn't yet turned to the debt markets, despite having the credit standing to do so cheaply.

"We're fortunate to have access in the debt markets, given where our ratings are, and we haven't yet tapped that," Hood said.

That changed within 24 hours. The logic is straightforward: when a AAA-rated company can lock in debt at favorable rates, it often makes more financial sense to borrow for long-term infrastructure than to drain cash reserves or slow down building. Capital expenditures for AI infrastructure, the land, power, servers, and chips that make up a data center, are the kind of long-lived assets that companies commonly finance with debt rather than cash on hand.

Microsoft isn't the only one borrowing

The bond market has become a popular stop for companies racing to build AI infrastructure. Amazon and Alphabet have both issued debt this year for their own infrastructure projects. CoreWeave, the AI-focused cloud provider, has also relied heavily on borrowed money to expand its data center footprint.

Oracle joined the list this week too. On Wednesday, the same day as Microsoft's earnings call, Oracle said it plans to raise $10 billion in new debt. Oracle also said it expects to return to positive free cash flow in the second half of its 2028 fiscal year, a timeline that underscores just how long some of these AI infrastructure bets take to pay off.

There's a backdrop worth keeping in mind here: central bankers have been raising interest rates to fight inflation, and investors have grown warier of companies carrying large debt loads as borrowing gets more expensive. That makes a AAA rating like Microsoft's more valuable than ever, it's one of the few that still unlocks relatively cheap capital in a tighter-rate environment.

The bigger picture

What's happening across Microsoft, Amazon, Alphabet, Meta, Oracle, and CoreWeave amounts to a real-time stress test of how the biggest players in AI are choosing to fund the buildout. Some, like Microsoft, are generating enough cash to self-fund while still reaching for debt opportunistically. Others, including Alphabet, Meta, and Amazon, are already running cash flow negative, meaning they're spending more than they bring in and need external financing just to keep pace.

Microsoft's position looks the strongest of the group right now: profitable, cash-generating, and borrowing from a position of strength rather than necessity. But the fact that even a company with Microsoft's balance sheet is turning to a $25 billion bond offering says something about the scale of money AI infrastructure demands. The capex figures keep climbing, and so does the appetite for debt to match it. Whether that pace holds, and whether free cash flow across the industry stays this divided, is something investors will be watching closely into next year.