Hook: The Order Book Has a Shelf Life
A single SEC filing just dropped a signal on the semiconductor equipment sector. Third Point LLC, Dan Loeb's $18 billion hedge fund, disclosed it has unloaded a significant portion of its stake in Lam Research (LRCX). The market narrative is still digesting the news as a simple profit-taking move. But for anyone who has spent years auditing the capital expenditure (capex) cycles of wafer fabs, the filing reads like a technical alert. The market does not care about your narrative about AI; it cares about the next quarter's order book. This is not a sell-off driven by a broken product, but by a broken clock. The AI infrastructure build-out has a predictable cycle, and the smart money is already rotating out of the picks and shovels phase.
Context: The Gatekeeper of the 3D Stack
Lam Research is not a chip designer; it is a machine builder. It is the dominant supplier of plasma etch and thin-film deposition equipment, the critical processes that define the physical geometry of a modern semiconductor. When a wafer fab wants to build a 3D NAND chip with 300 layers, they buy Lam’s high-aspect-ratio etch tools. When they build HBM (High Bandwidth Memory) stacks for AI accelerators, they use Lam’s TSV (Through-Silicon Via) etch and deposition systems. The company holds a 40%+ market share in the etch-deposition segment for memory chips. Its revenue is a direct proxy for the intensity of global wafer fab equipment (WFE) spending. Third Point held a significant position in LRCX, riding the AI-driven rally from 2023. The filing reveals a structural shift in their conviction, not a tactical trim. The question is not whether Lam is a good company, but whether the current business cycle can support the stock's multiples.
Core: The Order Flow Decay Curve
To understand the timing of the sale, we must look at the hard data on WFE cycles. Lam’s revenue is a lagging indicator of wafer fab decisions. The leading indicator is the order book, which typically peaks 12-18 months before a capex cycle tops out. Let’s run the numbers. Global WFE capex is projected to hit $105 billion in 2025, driven by the AI/GPU/HBM expansion. This is an extreme peak relative to the historical average of 12% of semiconductor revenue. The current ratio sits at 15%, which is unsustainable. The device characteristic of the current cycle is that the incremental growth is heavily concentrated in HBM and advanced packaging (CoWoS). If you strip out the AI-related spending, the rest of the semiconductor market (PC, mobile, automotive) is in a mild recovery, not a boom. This creates a concentration risk. If the hyperscalers (Microsoft, Amazon, Google) even slightly trim their 2026 AI capex guidance—perhaps due to a return-on-investment reassessment—the equipment orders will collapse faster than the chip revenue. Third Point is not selling Lam because they think AI is a bubble. They are selling because they are modeling a capex deceleration in 2026. The logic is brute-force math: WFE grew ~30% in 2024, another 15% in 2025, and then the consensus expects a plateau. A plateau is a death sentence for a 30x PE stock. The order book data from the latest earnings calls confirms this: Lam’s management has been guiding for a “normalization of customer lead times,” which is a polite way of saying the surge is over. The sell-side is still modeling 12% EPS growth for 2026, but the buy-side is already discounting a 5% contraction. Trust is a variable; verification is a constant. The SEC filing is the verification.
Contrarian: The China Trap is the Real Story
The mainstream narrative is that Third Point is just banking profits. The contrarian view is that this is a structural repositioning based on the China tariff exposure. The US export controls on advanced semiconductor equipment are not a short-term hiccup; they are a permanent structural drag on Lam’s addressable market. Before the October 2022 export controls, China represented 29% of Lam’s revenue. By 2024, that number dropped to ~20%. The consensus expects it to fall to 15% by 2026. The bulls argue that the rest of the world (US, Korea, Japan, Europe) will fill the gap. But the math doesn't work. The US CHIPS Act is funding fabs, but the capacity is coming online slowly (2026-2027). The China market was the highest growth engine for mature node equipment (28nm+). Without that, Lam is left fighting for incremental HBM orders in Korea, where Tokyo Electron (TEL) is a fierce competitor. The danger is that the export controls not only cap Lam’s revenue but also increase its cost structure. The compliance and license application costs are rising. Furthermore, the Chinese are accelerating their domestic equipment production. Companies like Naura and AMEC are gaining traction in the 28nm etch/deposition segment. The retail investor sees a 20% revenue decline from China as a “one-time reset.” The institutional investor sees it as a permanent impairment of the growth trajectory. This is a classic blind spot. The stock is pricing in a V-shaped recovery, but the reality is an L-shaped landing for the China business. Arbitrage is the immune system of the protocol—in this case, the arbitrage is between the narrative of AI growth and the reality of deglobalization.
Takeaway: The Playing Field is Shrinking
Third Point’s exit is a tactical signal, not a terminal diagnosis. Lam Research remains a high-quality company with a strong moat in etch and deposition. The technology is not broken. The market is broken. The next 12 months will be a test of whether the stock can hold its 30x PE while the WFE cycle peaks. The 200-day moving average is the key support. If LRCX breaks below $850, the technical damage will confirm the structural rotation. The real question is not whether to buy Lam, but what to buy next. The smart money is moving from the equipment providers (the “picks and shovels”) to the end-use chip consumers (the “gold miners”). If you are long Lam, you are betting on a second wave of AI capex in 2027. That is a long time to wait. Yield farming is a strategy of time arbitrage—sometimes you have to harvest the yield and move to the next pool. Third Point just harvested. The market should take note.
