Hook
Third Point LLC filed a 13F with the SEC yesterday, revealing a complete liquidation of its stake in Lam Research. The sale, executed in Q4 2024, removed approximately $1.2 billion in Lam shares from the fund’s portfolio. Data doesn’t. This is not a casual portfolio rebalance. It is a calculated signal from a hedge fund that historically front-runs macro shifts in hardware cycles. For crypto participants, this filing is more than a semiconductor headline. It is a leading indicator for the next leg of the AI infrastructure buildout—and the tokens tied to it.
Context
Lam Research is a top-three wafer fab equipment (WFE) supplier. Its core business—etch and deposition tools—is essential for manufacturing HBM (high-bandwidth memory), advanced logic nodes (3nm/2nm), and 3D NAND. That makes Lam a direct beneficiary of the AI capex super-cycle driving cloud providers, GPU manufacturers, and memory makers. Since 2023, Lam’s stock has surged 80%+ as the market priced in a decade of HBM growth. But the Third Point exit raises a cold question: Is the AI capex expansion already priced in, and is the marginal growth rate peaking?
Crypto’s intersection with this is twofold. First, AI tokens (e.g., Render, Akash, Bittensor) rely on the same GPU and HBM supply chains. Any slowdown in wafer-start commitments translates to tighter GPU availability for decentralized compute networks. Second, the Bitcoin mining industry has been shifting toward ASICs, but newer miners (e.g., Canaan, Bitmain) also compete for advanced packaging capacity. A capital expenditure deceleration in the broader semiconductor ecosystem squeezes supply for all non-AI endpoints.
Core
From a technical perspective, the Third Point signal is backed by three quantifiable factors. First, Lam’s valuation: trailing P/E of 32x, above the 5-year average of 26x, and well above the 25x sector median. The forward P/E on FY2025 consensus earnings is 28x—still rich given that revenue growth is expected to decelerate from 15% in FY2024 to 10% in FY2025. Second, export controls: Lam’s China revenue share dropped from 29% (FY2021) to ~20% (FY2024), and the U.S. Commerce Department’s “presumption of denial” policy for advanced equipment is unlikely to reverse. Third, the storage cycle: HBM equipment demand surged 50% in 2024, but the replacement cycle for HBM3e → HBM4 is longer than the market assumes. Leading-edge equipment orders for 3D NAND are already softening as inventory levels normalize.
Based on my audit experience of supply chain data during the 2021 DeFi Summer liquidity stress tests, I learned that front-running hardware cycles requires tracking the slope of capital expenditure growth, not the level. The slope is flattening. Cloud hyperscalers (AWS, Azure, GCP) collectively guided $220B in AI capex for 2025, up 30% YoY. But the incremental YoY growth was 45% in 2024. The deceleration, at the margin, is what equipment suppliers like Lam feel first. The Third Point sale is a bet that the WFE market will plateau or decline in 2026, compressing Lam’s P/E back to 22x.
On-chain metrics > Twitter polls. The on-chain equivalent is tracking the number of active GPU compute nodes on Akash or the hash rate expansion of Bitcoin miners. Both are currently showing signs of deceleration. Akash’s provider count grew 12% in Q4 2024, down from 22% in Q3. Bitcoin’s hash rate growth slowed to 8% annualized, the lowest since the 2022 bear market. This is not a coincidence. Hardware supply constraints are already biting.
Contrarian
The contrarian angle that the broader market misses is that Third Point’s exit is not a vote against Lam’s technology—it is a vote against the market’s pricing of certainty. Lam’s moat in high-aspect-ratio etch and TSV deposition remains unchallenged by Applied Materials or Tokyo Electron in the near term. The real risk is not that Lam loses share, but that the entire sector is over-earning relative to the cycle. Historical data shows that WFE revenues peak at 14% of global semiconductor sales, then revert to 11%. In 2024, WFE hit 13.5%. We are near the peak.
For crypto, the contrarian take is that a slowdown in AI capex could actually benefit decentralized compute networks. If hyperscalers pull back on new data center builds, the secondary market for GPUs (e.g., NVIDIA H100s) becomes more liquid, lowering prices for smaller AI startups and decentralized GPU rental platforms. This is a rotation away from centralized infrastructure providers toward distributed ones. Third Point’s move may inadvertently accelerate that shift.
Another blind spot: the Chips Act subsidies. The U.S. government is injecting $39B into domestic fabrication, which will sustain Lam’s revenue for the next 2-3 years regardless of the commercial cycle. The order book for Intel’s Ohio fab and TSMC’s Arizona fab is already locked in. The Third Point sale may be ignoring this government-backed floor.
Takeaway
Verify the hash, ignore the hype. The Third Point filing is a flashing red light for the AI hardware bull case. Crypto participants need to watch Lam’s next earnings call (February 2025) for order guidance, and cross-reference with on-chain GPU utilization metrics. If the WFE peak is confirmed, expect a rotation from AI infrastructure tokens to value-oriented DeFi protocols. The next six months will separate the narratives from the fundamentals.