By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel is divided on Micron's (MU) outlook, with some arguing that Burry's short is misguided due to the structural shift in memory demand driven by AI, while others warn of potential risks such as a squeeze or margin compression from increased supply.
Risk: A potential squeeze or liquidity-driven move in highly owned semiconductors, especially if rates ease or AI capex disappoints.
Opportunity: The shift in memory demand towards AI-driven margins, providing a revenue floor that didn't exist in previous cycles.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Michael Burry, the Scion Asset Management founder popularized in The Big Short, disclosed a new short position against Micron Technology (NASDAQ:MU) in a Substack post dated July 2. Burry stated that the puts "seemed expensive," so he shorted the stock directly and would add puts if volatility eases. He identified his disclosed entry at $1,051.87 per share.
The Micron short adds to a broader campaign. On June 30, Burry disclosed shorts against NVIDIA (NASDAQ:NVDA), Applied Materials (NASDAQ:AMAT), and the iShares Semiconductor ETF (NASDAQ:SOXX), arguing that AI-related chip stocks may be due for a 30% correction.
With the U.S. markets closed for the holiday, the disclosure lands into a market where every one of Burry's targets has ripped higher year to date. His thesis leans on cycle history and stretched technicals.
Micron shares are up 242% year to date (YTD) through July 2, with a market cap near $1.17 trillion. Burry argues that the rally has reached "historically extreme" levels, with Micron stock more extended above its 200-day moving average than at any point since 1984, "not even during the dot-com peak."
Don't wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.
Burry stated, "Micron defines cyclical like no other," citing 34 drawdowns of more than 30% over 42 years, a median return on invested capital (ROIC) of 4%, and return on equity (ROE) of 7%, which he called "frankly terrible." He added that "one quarter in every three, Micron is a destroyer of capital," with free cash flow negative 48% of the time.
His view: the move reflects fear of missing out (FOMO) and greater-fool dynamics around high-bandwidth memory (HBM) being "sold out through 2026." The timing carries some irony given President Trump praised Micron for a $250 million Trump Accounts commitment, and CEO Sanjay Mehrotra highlighted more than $200 billion in U.S. memory investment.
Four leading AI models discuss this article
"Micron's structural HBM demand through 2026 creates downside protection that prior 34 drawdowns of 30%+ never faced."
Burry's direct short in MU at $1,051.87, layered on NVDA, AMAT, and SOXX, correctly flags extreme technical extension—MU's distance above its 200-day MA is the widest since 1984. Yet the disclosure underplays how HBM supply remains sold out into 2026, a structural demand layer absent from prior memory cycles where MU posted negative FCF 48% of the time. With $50B Q4 revenue guidance and a $1,486 consensus target, the trade now requires not just mean reversion but an abrupt AI capex pause. SOXX's 86% YTD gain amplifies sector-wide beta, so any macro tightening could trigger the 30% correction Burry targets, but timing remains the unaddressed variable.
Even with the HBM backlog, a sudden slowdown in hyperscaler GPU orders or faster ramp from Samsung/SK Hynix could collapse pricing power faster than cycle history predicts, validating the short regardless of sold-out status.
"Micron’s transition to an HBM-focused oligopoly invalidates the historical cyclicality that Burry is using to justify his short position."
Burry is betting on mean reversion, but he is fundamentally misreading the structural shift in memory. Micron (MU) is no longer just a commodity DRAM player; it is an HBM (High Bandwidth Memory) oligopolist alongside SK Hynix and Samsung. The 'cyclical' argument ignores that HBM supply is effectively locked into long-term AI infrastructure contracts through 2026, providing a revenue floor that didn't exist in previous cycles. While technical extensions above the 200-day moving average are objectively stretched, the valuation re-rating is justified by the shift from cyclical commodity pricing to semi-contractual AI-driven margins. Burry is fighting a fundamental regime change with a historical playbook that no longer applies.
If enterprise AI spending hits a 'deployment wall' due to power constraints or insufficient software ROI, the HBM supply glut could arrive faster than anticipated, proving that Burry’s historical cycle analysis is actually the more prudent outlook.
"Burry correctly identifies technical excess but misses that Micron's current profitability metrics (50%+ gross margin, 25%+ ROIC) are structurally higher than his historical baseline, making the valuation less absurd than the 1984 comparison suggests."
Burry's short is tactically sound on technicals but strategically risky on fundamentals. Yes, Micron is cyclical and 242% YTD is extreme—but HBM demand is real, not FOMO. The article omits that Micron's gross margins expanded to 50%+ (vs. historical 30-40%), and Q4 guidance of $50B revenue implies ~$15B EBITDA, a 12-year high. Burry's 4% median ROIC is backward-looking; current-cycle ROIC is 25%+. A 30% correction is plausible if AI capex disappoints, but timing matters enormously. Shorting at $1,051 risks a 6-12 month squeeze if earnings confirm the cycle shift is structural, not speculative.
If Micron's HBM backlog extends to 2026 and AI data-center capex sustains 30%+ YoY growth through 2025, Burry is early and wrong—and early shorts on cyclicals with genuine demand acceleration get destroyed by short squeezes and margin calls before the correction arrives.
"Durable AI-driven demand and ongoing data-center capex could keep the semiconductor cycle healthier than a 30% correction implies."
Burry’s bets lean on a cyclical downturn and stretched charts, but the missing context matters. First, AI compute demand and data-center capex have stayed resilient; memory pricing can stabilize as supply tightness persists into 2025-26. Second, AMAT and SOXX are proxies for capex cycles, not just speculative hype; a re-acceleration in orders could drive multiple expansions. Third, the risk of a squeeze/liquidity-driven move in highly owned semis exists, especially if rates ease or monetization of AI accelerates. Finally, the article's '30% drop' target feels arbitrary without a probabilistic scenario and time frame.
AI-driven capex tailwinds could keep the sector bid, and a longer-than-expected horizon before a material correction occurs. Also, short-squeeze risk in a crowded semi space cannot be ignored.
"Competitor capacity ramps scheduled for 2025 could erode HBM pricing power before AI demand absorbs the extra supply."
Claude correctly flags squeeze risk if HBM backlog holds, yet both he and Gemini understate how SK Hynix and Samsung capacity ramps already scheduled for 2025 could flood supply even while Micron's own lines stay booked. This creates a classic oligopoly breakdown: contracted volumes get honored but at lower spot pricing once incremental output hits, compressing the 50%+ margins Claude cites before any AI capex pause arrives.
"HBM manufacturing complexity and yield constraints act as a structural barrier to supply-driven price erosion, regardless of capacity expansion."
Grok, your oligopoly breakdown assumes pricing power is purely elastic, but you're ignoring the 'yield' problem. HBM3E is notoriously difficult to manufacture; even if SK Hynix and Samsung ramp capacity, their effective output is constrained by low binning yields. Micron’s superior yield curve is the real moat here. You are treating HBM like standard DRAM, but the complexity of stacking and thermal management creates a supply bottleneck that prevents a price-collapsing glut in 2025.
"Yield complexity delays but doesn't prevent supply-driven margin compression once competitors hit production scale."
Gemini's yield-moat argument is credible but incomplete. Yes, HBM3E binning is brutal—but yields improve exponentially once volume ramps past initial learning curves. Samsung and SK Hynix have shipped HBM2 at scale; HBM3E yield curves compress faster than historical DRAM. The real question: does Micron's current 50%+ margin assume they're the only viable supplier, or does it hold at 60-70% of current volume? If the latter, Grok's oligopoly breakdown still triggers even with yield constraints.
"HBM yields matter, but demand timing and contract pricing risk erodes any yield moat."
Gemini overstates the moat: yields help, but the real driver is demand mix and contract pricing. Even with HBM3E binning improvements, capacity ramps by Samsung/SK are not a pure moat if AI capex slows or shifts to non-HBM architectures. The risk is demand-led margin compression that a yield advantage cannot shield; MU could still see multiple compression if bookings weaken or contract prices renegotiate lower.
The panel is divided on Micron's (MU) outlook, with some arguing that Burry's short is misguided due to the structural shift in memory demand driven by AI, while others warn of potential risks such as a squeeze or margin compression from increased supply.
The shift in memory demand towards AI-driven margins, providing a revenue floor that didn't exist in previous cycles.
A potential squeeze or liquidity-driven move in highly owned semiconductors, especially if rates ease or AI capex disappoints.