Michael Burry Sounds Alarm on Big Tech's AI Spending Binge
Michael Burry, the investor renowned for predicting the 2008 financial crisis and shorting the US housing market, has shifted his focus to Big Tech's massive artificial intelligence investments. In a series of posts on X, Burry, who manages Scion Asset Management under the handle Cassandra Unchained, posed a critical question largely ignored on Wall Street: when does the relentless AI spending actually stop?
Big Tech's $660 Billion AI Tab and Mounting Debt
Burry's argument is straightforward and stark. He points out that companies like Amazon, Google, Meta, Microsoft, Oracle, Nvidia, and Caterpillar are depleting their cash flows on AI data centers, resorting to unprecedented borrowing, and employing accounting maneuvers to obscure earnings impacts. "It is consuming all your cash flow, you are borrowing, you are financing in ways you never have," Burry wrote. "But if it scales, when does it end?"
The data supports his concerns. Amazon, Alphabet, Meta, and Microsoft are projected to spend approximately $660 billion on AI infrastructure in 2026, a staggering 165% increase from 2024. This expenditure surpasses, as a percentage of US GDP, historic projects like the Apollo space program, interstate highway construction, and the 1850s railroad expansion. Amazon alone has committed $200 billion, a sum expected to exceed its operational cash flow for the first time, prompting debt-raising plans.
Alphabet is issuing a 100-year century bond, its first since IBM in 1996, as part of a $20 billion-plus multi-currency debt raise, with long-term debt soaring to $46.5 billion. Meanwhile, the Stargate AI data center project, targeting $500 billion with partners like OpenAI and Oracle, is reportedly stalled with less than $10 billion secured and no major construction underway.
Market Reactions and Revenue Gaps
Wall Street has reacted negatively, with Amazon, Google, and Microsoft losing a combined $900 billion in market value post-earnings calls. Microsoft revealed that 45% of its $625 billion cloud backlog stems from a single customer, OpenAI, unsettling analysts. Burry emphasized that the combined revenues of Amazon, Apple, Alphabet, Microsoft, Meta, and Nvidia fall short of $2 trillion, highlighting a widening gap between income and outflows that normalizes leverage.
Historical Parallels: Electrification and Bubbles
When questioned if the AI buildout resembles a Cisco-style bubble, Burry offered a nuanced comparison to the late 19th and early 20th-century electrification boom. He noted that while the technology was real and demand high, with electric light as a killer app, companies like Westinghouse and Edison faced bankruptcies, price wars, and write-downs before stabilization. Many central power stations failed within decades, and financial panics in 1893 and 1907 wiped out numerous firms, underscoring that transformative tech doesn't guarantee investor safety.
Google CEO Sundar Pichai, at the India AI Impact Summit 2026, cited railroad and highway expansions to justify Alphabet's spending, yet just months earlier, he acknowledged "elements of irrationality" in AI investments, likening it to the dot-com crash, signaling a rapid shift from caution to conviction.
Depreciation Risks and Hardware Obsolescence
A significant concern is hardware depreciation. Nvidia's upcoming Vera Rubin chips are expected to offer five times the inference performance of current Blackwell GPUs in one generation, rendering six-year depreciation schedules unrealistic. Microsoft CEO Satya Nadella has expressed reluctance to depreciate hardware over four to five years.
Burry's analysis estimates that if chip useful life is 2.5 years, cumulative earnings overstatement across Big Tech could average 32% from 2026 to 2028, with Oracle potentially overstating by 62%. Even at three years, average overstatement is 18%. These discrepancies could lead to substantial write-downs, which, while minor against current multi-trillion market caps, could become critical if valuations decline.
Burry isn't dismissing AI's potential; his electrification analogy affirms its transformative nature. However, he warns that the financial engineering—debt, depreciation, and earnings adjustments—demands greater scrutiny. History shows that even groundbreaking technologies can lead to investor losses if not managed prudently.
