NEW YORK – Microsoft sits 31% below the peak it reached last October. Michael Burry, whose 2008 mortgage short made him the defining contrarian trade of that era, just bought call options on the stock that expire worthless unless it trades near $700 – 88% above where it closed on June 28 – by December 2028.
The bet, disclosed this week, consists of long-dated LEAP call options on Microsoft with a $700 strike price expiring December 15, 2028. Microsoft closed at $372.73 on June 28. The number of contracts held by Scion Asset Management has not been made public, because Scion deregistered with the Securities and Exchange Commission in November 2025 and no longer files the quarterly Form 13F reports that historically made Burry’s positions legible to markets. What is known is the strike, the expiry, and that Burry himself described the December 2028 LEAPs as “inexpensive” relative to his conviction on the stock – and that he views the $350 range as a rational entry point for an outright equity position.
That word – “inexpensive” – is doing a lot of work. At current prices, these options are roughly 88% out of the money. They expire worthless if Microsoft does not reach $700 by mid-December 2028. The implicit thesis is not simply that the stock will recover. It is that the market is fundamentally mispricing what Microsoft’s AI infrastructure investment will be worth in two and a half years, and that the cost of that thesis in option premium is asymmetrically low relative to the potential payoff.
The position sits in unusual company inside the same portfolio. Burry has not abandoned his bearishness on AI stocks. He still holds put options on Palantir Technologies with a $50 strike – a bet that one of the most richly valued AI-adjacent companies in the market is overpriced relative to its actual business. This week he covered approximately half of that short at $107.15 per share, reducing but not closing his bearish exposure. He also holds puts on Nvidia. The same disclosures that revealed the Microsoft calls showed him adding to positions in JD.com, acquiring shares at $24.79, and increasing his Adobe stake, where he has cited gross margins near all-time highs as evidence the market has mispriced a company caught in AI transition noise. The pattern is a valuation discipline applied selectively: not an AI bull, not a pure AI bear, but an investor sorting companies by what they cost relative to what they produce.
Microsoft’s fundamental case is built on numbers that have not moved the stock. Azure cloud revenue grew 40% in the first fiscal quarter of 2026 and 39% in the second, outpacing Amazon Web Services at 19% and Google Cloud at 30% by a margin that widened through the year. Commercial remaining performance obligations – a forward-looking bookings metric – reached $625 billion in Q2 fiscal 2026, up 110% year over year, with roughly 45% of that growth driven by OpenAI commitments. Revenue came in at $81.3 billion, above consensus, up 17%, according to Microsoft’s investor relations. The stock fell anyway.
What drove the selloff was not the revenue. It was the spending committed to sustain it. Microsoft has guided to approximately $190 billion in capital expenditure for fiscal 2026, driven by rising server memory costs and an accelerating data center build-out. The company plans to nearly double its data center footprint over two years and increase total AI capacity by more than 80% in a single fiscal year. For equity investors who need near-term earnings, that spending profile represents margin compression before it represents revenue growth. The Federal Reserve’s June pivot toward a projected rate hike compounded that pressure: higher rates raise the discount rate applied to long-duration earnings assumptions, and Microsoft’s AI return thesis is nothing if not long-duration.
A Microsoft trading at $700 would carry a market capitalization approaching $5.2 trillion – the highest valuation in the history of public markets. Getting there from today’s price would require roughly 25% annual compounding over two and a half years, a significant re-rating of the stock’s earnings multiple once capex spending moderates, or some combination of both. Burry has not said he expects $700. He has said the options are inexpensive, which is the vocabulary of someone who believes the bet is skewed favorably – that even partial upside represents a return the premium does not reflect.
The lack of a 13F filing means the position cannot be independently sized. Coverage of the bet as a “big” wager – the characterization that appeared across financial media this week – is accurate in directional terms but impossible to calibrate in dollars. The same deregistration that freed Scion from disclosure requirements makes it impossible to assess how large a conviction this represents relative to the rest of the portfolio. That ambiguity is part of the market’s response: an unknown bet from a known name moves stocks regardless of its actual size. Microsoft climbed after the disclosure was reported.
JD.com gained the same week. Adobe held. Palantir, still shorter by Burry’s remaining puts, has not yet told him he was wrong about overvaluation. The one position that requires a near-doubling of a $2.76 trillion company will take longer to resolve. December 15, 2028 is the date those calls either expire worthless or become the most discussed trade of this investing cycle.
