Michael Burry’s latest bet puts Nvidia stock and Micron’s AI boom on trial

Michael Burry’s latest bet puts Nvidia stock and Micron’s AI boom on trial
Devesh Kumar
04 Aug 2026, 10:45 AM

powered by

Invezz
NVDA puts (Dec 18, 2026)

Buy NVDA put options (strike in the low $100s, as in Burry’s trade). Thesis: hyperscaler capex may slow if AI returns disappoint, and Nvidia’s hardware cycle risk is real—buyers can delay upgrades if processors get obsolete before ROI is proven. Even without a crash, rising volatility and renewed “AI skepticism” can reprice puts upward.

Key Risk: Hyperscalers keep proving AI ROI (cloud growth + margins), forcing sustained Nvidia data-center spending and compressing volatility/sentiment back to bullish.

MU short (near $880)

Sell Micron (MU) short. Thesis: memory is becoming strategic, but the cycle still matters—capacity expansion (Samsung/SK Hynix/CXMT) can outpace demand, driving pricing down and hitting earnings leverage. Burry’s add near ~$880 signals belief the market is underpricing the oversupply risk.

Key Risk: Memory pricing stays firm because AI demand growth absorbs new capacity faster than expected, keeping margins resilient.

  • Michael Burry expands chip shorts as Wall Street defends durable AI demand.
  • Burry’s Nvidia puts challenge faith in sustained hyperscaler AI spending.
  • Micron’s memory boom faces Burry’s warning over future industry oversupply.

Michael Burry has increased bearish bets on Nvidia, Micron and the semiconductor sector, challenging Wall Street’s confidence in the artificial-intelligence boom.

In a July 30 update on his personal Substack, Burry said he added Nvidia put options expiring on December 18, 2026, increased a Micron short near $880 and expanded an iShares Semiconductor ETF short around $506.

Unlike a 13F filing, the update does not reveal position sizes, option premiums or hedges.

Burry is challenging Wall Street’s favourite AI assumptions

Burry’s Nvidia puts carry strike prices in the low $100s, far below Monday’s $206.64 close.

Nvidia does not need to fall below those strikes for them to gain as falling shares, higher volatility or weaker expectations could lift them before expiry.

His concern extends beyond valuation. Burry has questioned hyperscaler spending, circular financing between chipmakers and customers, and whether hardware improvements make costly processors obsolete before buyers recover their investment.

Micron represents a more cyclical wager as Burry added to his short near $880, while the shares closed Monday at about $829.11.

That addition is below its reported entry level, although his total exposure, average cost and borrowing expenses remain unknown.

Nvidia and Micron rose on Monday. Expanding bearish exposure into that rebound suggests Burry believes sentiment has not resolved the industry’s deeper risks.

The risk remains customer concentration as higher borrowing costs or disappointing AI returns could force cash-rich technology groups to become more selective.

Nvidia bulls see real returns from AI spending

The strongest challenge comes from Nvidia’s customers. Jefferies analysts said Microsoft and Amazon’s latest results offered evidence that AI investment can produce tangible returns.

The firm believes worsening sentiment towards semiconductor stocks may have bottomed.

If cloud growth and margins keep accelerating, hyperscalers will have less reason to reduce spending on Nvidia processors and data-centre infrastructure.

Morningstar equity analyst Brian Colello is also bullish. Business Insider reported that Morningstar values Nvidia at $280 and considers the shares undervalued, although it gives the stock a “very high” uncertainty rating.

Colello said Nvidia’s prospects remain underestimated because technology customers should sustain AI spending.

Morningstar also expects internally designed chips from Amazon, Google and others to complement, rather than replace, Nvidia’s broader hardware-and-software platform.

Micron is the harder cyclical test

Micron may provide the clearer verdict because memory shortages have historically encouraged investment that eventually creates oversupply.

Bank of America analyst Vivek Arya argues that memory is becoming a strategic AI resource rather than an ordinary commodity.

Barron’s reported that he expects memory to represent 35% to 40% of global cloud and AI infrastructure spending in 2027. He retained a Buy rating and $1,550 target.

TD Cowen analyst Krish Sankar has also forecast favourable pricing into 2027, according to MarketWatch, supported by higher memory content and high-bandwidth memory’s tighter supply structure.

Yet capacity is expanding. Micron has raised its investment plans, Samsung and SK Hynix are adding production, and China’s CXMT is pursuing further growth.