The Shovel Sells First: Third Point's Lam Research Exit Signals a Cycle Shift for AI and Crypto

To sell a shovel in a gold rush is to bet on the depth of the mine. When Third Point filed its SEC disclosure this week, showing a full exit from Lam Research (LRCX), the move was not just a portfolio rebalance. It was a quiet, deliberate signal—one that speaks directly to the heartbeat of the AI infrastructure machine that crypto, too, depends on. Lam is the backbone of the etching and deposition equipment that carves the HBM stacks and advanced nodes powering every NVIDIA GPU and every ASIC miner. If the shovel maker is being sold, who is left to dig? Lam Research is a semiconductor equipment oligarch. Its etching tools are the invisible hands that sculpt the 3D NAND layers and TSV interconnects inside every HBM3e chip. Without Lam, there is no AI training cluster. Without Lam, there is no Bitcoin ASIC upgrade. The company's 40%+ market share in high-aspect-ratio etching makes it a de facto toll booth on the road to silicon. Third Point's exit, therefore, is not a critique of Lam's technology—it is a critique of the cycle. Over the past 18 months, the global Wafer Fab Equipment (WFE) market surged past $100 billion, driven by a frenzy of AI capital expenditure. Lam's stock tripled from its 2022 lows. Yet the math of a capital expenditure cycle is unforgiving: the peak of orders always precedes the peak of spending by 12 to 18 months. Third Point, with its event-driven DNA, is reading the tea leaves. The US Dollar's strength, the tightening of export controls to China, and the reality that hyperscalers (Microsoft, Amazon, Google) are already rethinking the ROI of their AI buildouts—all point to a deceleration in equipment orders by late 2025. Based on my own audit of tokenized hardware supply chains for a decentralised AI protocol, I've seen the same pattern play out in crypto mining. When the price of ASICs surges, the market assumes infinite demand. But the moment the next-generation chip hits production, the previous generation's value collapses. Lam is the ASIC of the AI world. Third Point is simply selling before the next-gen overhang arrives. The contrarian truth, however, is far more interesting for the crypto ecosystem. A slowdown in semiconductor equipment spending does not mean a collapse in AI advancement. It means a shift from building the most expensive, bespoke chips to scaling more efficient, commoditized silicon. This is exactly the window where decentralized AI networks—like those running on Bittensor or Akash—can benefit. Cheaper hardware lowers the barrier to entry for community-run nodes. The same overcapacity that will hurt Lam's earnings will eventually flood the market with affordable HBM and GPUs, making it economically viable for a thousand small miners to participate in inference tasks. To own nothing is to feel everything, deeply. The crypto ethos is about resilience through redundancy. While Wall Street sees a cycle peak, I see the beginning of hardware abundance. The soul does not mint; it manifests. Third Point is betting on the contraction. I am betting on the distribution. Here is the raw technical signal: the industry's R&D intensity is shifting from achieving the smallest node to optimizing the most interoperable node. Lam's expertise in extreme ultraviolet (EUV) post-etch processes is critical for 2nm, but the next wave of demand will come from chiplet integration and hybrid bonding—technologies that reduce the need for high-end etching. The equipment that built the HBM boom may not be the equipment that services the HBM plateau. Third Point is not wrong; they are just early. For the crypto builder reading this, the takeaway is not to panic. It is to prepare. The signal from Third Point is a reminder that the hardware cycle is a pendulum. When the pendulum swings back, the cost of compute will drop. And when it drops, the promise of a truly decentralized, permissionless AI infrastructure becomes not just a dream, but a practical reality. Trust is not a transaction; it is a resonance. The market is speaking. Listen to the frequency, not the noise.