Rattner warns AI boom faces $205B debt test as rates rise

Big Tech's $205 billion AI infrastructure bet faces a reckoning as investors demand wider margins of safety from companies borrowing to fund capital spending.

Steven Rattner said artificial intelligence will be a major contributor to the economy, but the market faces unresolved questions about debt levels, China exposure and higher interest rates that could test Big Tech's record capital spending. The financier and former auto industry advisor pointed to growing skepticism about whether the unprecedented investment in AI infrastructure will generate proportional returns.

"Investors are starting to demand a wider margin of safety from companies borrowing to fund massive capital spending," Rattner said.

Google raised its 2026 capital expenditure forecast to as much as $205 billion, up $15 billion from its April guidance, as it ramps up investment in AI data centers and cloud infrastructure. The search giant's stock fell 3.9% after the announcement, reflecting growing investor unease about whether hyperscalers can generate sufficient returns on unprecedented infrastructure spending. Alphabet Chief Financial Officer Anat Ashkenazi said expenditures will be "significantly" higher next year as the company continues to acquire more compute capacity.

The stakes are enormous: Google's cloud division reported a $514 billion backlog in contracted work and revenue of $24.7 billion, up 82% from a year earlier. But Google has also fallen behind Anthropic and OpenAI in releasing frontier AI models, and delayed its Gemini 3.5 Pro model after it failed to meet internal benchmarks. The company struck a $920 million-a-month agreement with SpaceX for access to its data centers, underscoring the extent to which even the largest cloud provider lacks sufficient capacity.

Google's $205B CapEx Pushes the Limits of Investor Patience

Bond market anxiety over AI capital spending has been building for months as companies across the technology sector borrow to fund data center construction, GPU purchases and cloud infrastructure. The concern centers on whether the massive upfront investment will translate into sustainable revenue growth before debt servicing costs erode returns, particularly in a higher interest rate environment.

Google's situation illustrates the tension. Its Tensor Processing Units have become increasingly attractive for inference workloads, offering strong performance and lower operating costs for certain AI deployments. But the company's consumer Gemini app, while reaching 950 million monthly active users, faces questions about engagement depth compared with ChatGPT, which third-party data suggests is used considerably more actively.

The competitive landscape adds another layer of risk. Anthropic's Fable and Mythos models have dominated the AI conversation for more than a month, while OpenAI's Codex is gaining enterprise customers. Google's delayed Gemini 3.5 Pro was meant to be its major push into coding and enterprise workflows, but internal benchmarks were not met, according to Bloomberg.

For investors, the math is straightforward but the outcome uncertain. Google trades at roughly 22 times forward earnings, and its cloud backlog of $514 billion provides visibility into future revenue. But with capital spending set to rise again next year and competitors advancing faster on frontier models, the margin of safety that Rattner described is narrowing. If the AI infrastructure buildout fails to deliver the expected returns, the companies that borrowed most aggressively to fund it will face the steepest reckoning.

This article is for informational purposes only and does not constitute investment advice.