AIToday
AI Stocks & MarketsAI Business & IndustryYahoo Finance AIPublished: Sep 25, 2026, 22:00 JST

Michael Burry flags $3 trillion AI write-off risk at Microsoft, Amazon, Oracle

Michael Burry flags $3 trillion AI write-off risk at Microsoft, Amazon, Oracle

3 Key Points

  1. What happened

    Michael Burry said Microsoft, Amazon, Alphabet, Meta and Oracle carry about $3 trillion in AI-infrastructure commitments and exposures — leases, construction and purchase obligations — and flagged possible write-downs around 2028 or 2029.

  2. Why it matters

    Burry expects the investment to keep rising, so a mismatch between AI demand and capacity could force those companies to mark down the value of the infrastructure they have committed to, mirroring the excess capacity and weak returns of the late-1990s telecom buildout.

  3. What to watch

    Whether AI demand develops fast enough relative to new capacity is the test for whether those write-downs materialize. Burry also questioned Oracle's treatment of customer prepayments and cited rising financing concerns around its data-center projects.

WHO IT HITSThis lands hardest on the finance and investor-relations teams at Microsoft, Amazon, Alphabet, Meta and Oracle, who may face questions about how these lease, construction and purchase-obligation exposures are valued if AI capacity outruns demand.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

Burry's concern is not that the AI spending is happening, but that it may be getting ahead of the demand that is supposed to pay for it. He pointed to net capital investment by S&P 500 companies reaching 2.07% of GDP at June 30, a level he described as the highest in nearly four decades outside the period following the 2000 Nasdaq peak. That is the backdrop for the roughly $3 trillion in commitments and other exposures he attributes to Microsoft, Amazon, Alphabet, Meta and Oracle.

The comparison he draws is to the late-1990s telecom buildout, when heavy investment was followed by excess capacity, weaker returns and substantial depreciation. He expects the current investment to keep rising, and said write-downs could emerge around 2028 or 2029 if AI capacity exceeds demand. He also raised questions about Oracle's treatment of customer prepayments and cited rising financing concerns surrounding its large data-center projects.

The stakes appear to hinge on the pace of AI demand relative to new capacity, and on how these commitments are ultimately valued. If demand keeps up, the spending remains a growth story; if it lags, the same obligations could become the kind of asset write-down Burry is flagging, a risk that may weigh most on the finance and investor-relations teams at the five companies named.

FAQ
How large are the AI-infrastructure commitments Burry is pointing to?
He estimated Microsoft, Amazon, Alphabet, Meta and Oracle have about $3 trillion in commitments and other exposures linked to AI infrastructure, including leases, construction projects and purchase obligations.
When does Burry expect write-downs to show up?
He said write-downs could emerge around 2028 or 2029 if AI capacity exceeds demand.
What does Burry compare this AI spending cycle to?
He compared it with the late-1990s telecom buildout, when heavy investment was followed by excess capacity, weaker returns and substantial depreciation.
Yahoo Finance AIRead Original Article

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleFinancial Services Agency to scrutinize AI data center lending