The SEC filing hit the tape on a quiet Tuesday: Third Point LLC, Dan Loeb’s $12 billion hedge fund, had offloaded a significant stake in Lam Research. The semiconductor equipment giant, a linchpin in the production of chips for AI training and, by extension, crypto mining ASICs, saw its shares dip 2% on the news. For most market participants, this was a routine portfolio adjustment. For those of us who track the physical infrastructure underpinning digital assets, it was a flashing red indicator.
Lam Research is not a household name in crypto circles. Its equipment—etching and deposition tools for wafer fabrication—does not appear in whitepapers or on-chain explorers. Yet every NVIDIA H100 GPU, every Bitmain S21 miner, and every high-bandwidth memory (HBM) stack in an AI server passes through tools manufactured by Lam, Applied Materials, or Tokyo Electron. When a sophisticated actor like Third Point reduces exposure to the “pick-and-shovel” supplier of the AI era, the implications ripple into the crypto mining supply chain, GPU availability, and the cost of network security.
Context: The Lam Research Conundrum
Lam Research is a dominant player in the wafer fabrication equipment (WFE) market, with roughly 20% share. Its core competency lies in high-aspect-ratio etching for 3D NAND and HBM, the memory technology that fuels AI accelerators. In 2023, the company generated $17.4 billion in revenue, with about 25% coming from China. The U.S. export controls imposed in 2022 and 2023 directly targeted the advanced equipment Lam sells to Chinese customers, effectively capping a major growth engine. The stock surged over 80% in 2023 on AI hype, but by mid-2024, its forward P/E sat at 35x, a premium to historical averages.
Third Point’s sale is not an isolated event. Other institutional investors have trimmed positions in semiconductor equipment names over the past quarter. The consensus narrative: AI capital expenditure growth is peaking, and the equipment cycle, which typically leads wafer fab buildouts by 12-18 months, is turning down. As I’ve written before, “Follow the liquidity, find the leak.” Here, the liquidity is leaving Lam, and the leak is in the China revenue stream and the valuation bubble.
Core: Systematic Teardown of the Sale Signal
To understand why Third Point sold, we must dissect the three layers of the semiconductor equipment thesis: geopolitics, cycle timing, and valuation.
First, geopolitics. Lam’s China revenue has dropped from 29% of total in FY2021 to an estimated 20% in FY2024. The U.S. Bureau of Industry and Security maintains a “presumption of denial” for advanced equipment licenses to China. This is not a temporary blip; it is a structural shift. The Chinese government’s Phase III Big Fund, valued at $47 billion, is pouring money into domestic equipment makers like AMEC and Naura Technology. In the medium term, Lam will lose market share in mature-node etching and deposition. The only question is how fast. “Trust the code, not the press release” applies here: the code is the SEC filing, the press release is the geopolitical noise.
Second, the capital expenditure cycle. Global WFE spending is projected to reach $100 billion in 2025, but the growth rate is decelerating. Cloud hyperscalers—Amazon, Microsoft, Google—are expected to spend over $200 billion on AI infrastructure in 2024, with a 30% increase in 2025. However, equipment orders are a leading indicator. If the marginal dollar of AI capex shifts from GPUs to networking or software, Lam’s order book will feel the pain first. The HBM equipment market, which grew 50% in 2024, will likely see growth slow to 20% in 2025 as the technology matures. Third Point’s sale is a bet that the “easy” growth in AI hardware is behind us.
Third, valuation. At 35x trailing earnings, Lam trades at a premium to peers (Applied Materials at 28x, KLA at 30x). The historical average for the semiconductor equipment sector is 25x. A mean reversion to 25x would imply a 30% downside from current levels. Hedge funds are not in the business of holding through multiple compression. As one analyst quipped, “Run the numbers, ignore the hype.” The numbers here show a stock priced for perfection in a world where export controls, a capex slowdown, and Chinese competition are all risks.
Contrarian Angle: What the Bulls Got Right
I am not one to dismiss a counter-narrative out of hand. The bulls on Lam Research point to the company’s technological moat in high-aspect-ratio etching, a capability that is difficult to replicate. The HBM market, driven by AI memory demand, will require Lam’s tools for TSV etching and deposition for years. Furthermore, the company’s services and spare parts revenue—which carries higher margins—is less exposed to export controls. In FY2023, services accounted for roughly 30% of Lam’s revenue, and that proportion is growing. “Transparency is a feature, not a promise,” and Lam’s financial disclosures are transparent enough to see that cash flow remains robust.
Third Point’s sale could simply be a tactical rotation: selling Lam to buy NVIDIA or another pure-play AI beneficiary. The fund’s overall tech exposure may not have changed; it just shifted from the “pick-and-shovel” to the “miner.” In crypto terms, this is akin to selling a mining rig manufacturer to buy the mining pool operator. The underlying asset—AI compute—remains in demand.
However, the contrarian view must contend with the fact that Lam’s stock has already begun to underperform the broader semiconductor index since the filing. The market is pricing in a higher probability of a cycle downturn. “One exploit, one lesson, zero excuses” — the exploit here is the valuation bubble, the lesson is that institutional flows are a leading indicator, and there are no excuses for ignoring the signal.
Takeaway: The Crypto Infrastructure Connection
For crypto miners and investors in GPU-as-a-service platforms, the Third Point sale is a warning shot. If Lam’s equipment orders slow, the supply of advanced chips (HBM, GPUs) could tighten in the short term but then flood as overcapacity builds. The cost of mining equipment and the profitability of ASIC farms are directly tied to the semiconductor capital expenditure cycle. Hedge funds are rotating out of the equipment layer, signaling that the peak of the AI hardware cycle is near. The question for the crypto ecosystem is whether the next wave of capex will go into decentralized compute networks or centralized cloud providers. Based on this signal, I would bet on the latter.
Silence from the team speaks volumes — Third Point has not issued a statement, but the filing is the message. Follow the liquidity, find the leak. The leak is in the semiconductor equipment supply chain, and it will eventually reach the crypto mining rigs you rely on.