Michael Burry buys beaten-down mega tech stock

Michael Burry initiated a new long position in Microsoft and added to MSCI, PayPal, and Adobe on April 23, while also buying puts on QQQ, Nvidia, and the iShares Semiconductor ETF. He is betting that software names like Microsoft have been oversold on AI fears, with Microsoft down roughly 25% from its July 2025 high and trading around 26x forward earnings versus a five-year median of 34x. The trade is a targeted relative-value call rather than a broad bullish tech view, and Microsoft’s April 29 earnings are the next key catalyst.

Analysis

The signal here is not “buy software”; it is that the market is collapsing two very different AI exposures into one trade. Enterprise software with sticky installed bases, recurring revenue, and embedded distribution should be able to pass through AI capex and monetization better than headline sentiment implies, while the crowded compute stack remains vulnerable to any slowdown in hyperscale spend or multiple compression. That creates a cleaner relative-value setup in the software cohort than in semis, especially after a selloff driven more by narrative contagion than company-specific deterioration.

Microsoft is the highest-conviction expression because it can absorb AI investment longer than peers and still compound FCF, which matters when rates remain non-trivial and investors are increasingly punishing long-duration growth without visible monetization. If the market is discounting AI cannibalization of software seats, the second-order effect is that it may be underestimating pricing power for bundled platforms; Copilot-style attach rates can defend ARPU even if unit growth slows. The more important catalyst is not the broader macro tape, but whether upcoming guidance confirms that demand elasticity is intact after the IBM/ServiceNow scare.

The hedge matters more than the long: short exposure to NVDA/QQQ/SMH implies Burry sees asymmetry in the most consensus-owned part of the AI trade. That positioning suggests the next leg down in risk could come from a de-rating of infrastructure beneficiaries rather than a collapse in software fundamentals. In other words, even if AI remains strategically important, the trade may be overcrowded on the supply side and underappreciated on the application side.

The contrarian miss is that consensus may be extrapolating a sector-wide demand slowdown from a few weak guides while ignoring balance-sheet strength and installed-base resilience at the best franchises. But the risk is timing: this can stay mispriced for months if capex remains elevated and earnings beat the worst-case setup. The cleanest way to express the view is relative value, not outright beta, because the market can still punish all tech in a risk-off tape before fundamentals separate.

Key Decisions for Investors

• Long MSFT vs short NVDA as a 3-6 month pair trade: own the cash-flow compounder with pricing power and hedge the crowded AI infrastructure leg; target 10-15% relative outperformance if software sentiment normalizes.
• Long MSFT / short QQQ into earnings over the next 1-3 weeks: expresses selective quality within tech and reduces index beta; risk is a broad tape drawdown that overwhelms company-specific fundamentals.
• Buy ADBE on weakness for a 2-4 month rebound trade: the stock is priced for a prolonged demand air pocket; upside/reversal is highest if guidance implies enterprise spend is stabilizing rather than rolling over.
• Use put spreads on SMH or NVDA for event-risk protection over the next 1-2 months: semis are most exposed to multiple compression if AI capex enthusiasm decelerates, while defined risk limits bleed if momentum continues.
• Avoid chasing PYPL/MSCI outright; prefer small tactical longs only on confirmation of post-selloff stabilization, because these are lower-quality sentiment rebounds and less likely to rerate without a macro/multiple tailwind.