Intel revised its 2026 capital expenditure upward by $3 billion. The market immediately priced in gains for Japanese equipment makers—Lasertec, Tokyo Electron, Disco. But the crypto crowd? Dead silent. That silence is a mistake.
Volatility isn’t just price. It’s also the hidden rhythm of supply chains that shape the cost of every hash. When a chip foundry spends billions on advanced nodes, it doesn’t just produce more CPUs. It produces the machinery that builds the ASICs miners depend on. The connection is opaque, but the data is clear.
Context: Why Intel Matters to Crypto
Intel’s IDM 2.0 strategy aims to turn the company into a major foundry for external clients—including makers of Bitcoin mining ASICs. While Bitmain and MicroBT currently rely on TSMC and Samsung, Intel’s 18A node (targeting 2025-2026) offers a potential alternative: lower latency chips with better power efficiency. The $3 billion capex increase signals a bet on that future. But the real winners, according to Goldman Sachs, are not Intel itself. They are the Japanese equipment vendors who provide the tools Intel needs to reach 18A and beyond.
Core: Three Stocks, One Thesis
Goldman Sachs identified three key beneficiaries: Lasertec (EUV mask inspection, ~85% market share), Tokyo Electron (coating, etching, deposition, ~25-30% in etching), and Disco (precision dicing and grinding for advanced packaging). The logic: Intel’s ramp toward 18A and 14A nodes, plus its EMIB-T advanced packaging for AI chips, will require massive orders of these tools.
Let’s break down the technical data.
- Lasertec: Its ACTIS series is the only production-ready tool for High-NA EUV mask defect inspection. Without it, a single pinhole defect can render an entire wafer useless. Intel’s transition to 18A requires High-NA EUV, and ASML has only shipped one such system so far—to Intel. Every additional Intel wafer means one more mask inspected by Lasertec. Revenue per wafer grows linearly with complexity.
- Tokyo Electron: TEL’s etch and deposition tools are used in gate-all-around (GAA) fabrication and PowerVia backside power delivery. Intel’s 18A relies on both. However, TEL faces fierce competition from Applied Materials and Lam Research. Its market share in etching is ~25%, not dominance. The $3 billion capex uplift will be split among all three U.S. and Japanese players.
- Disco: This is the clearest winner. Disco’s dicing saws and grinders are essential for chiplet-based designs, like Intel’s EMIB-T. The AI chip trend—where Nvidia, AMD, and cloud giants like Google and Amazon design chiplets—is structurally bullish for Disco. Unlike TEL, Disco holds a near-monopoly (~50-80% share) in precision dicing for advanced packaging. Every new AI ASIC that uses chiplet architecture is a direct order for Disco.
Market forecasts: If Intel’s 18A captures 10% of the external foundry market by 2027, the cumulative equipment revenue for these three could reach $8-10B over three years. The $3B capex increase is just the tip.
Contrarian: The Hidden Risks Goldman Missed
Goldman’s thesis is correct on the macro trend, but it underestimates two critical risks—both directly relevant to crypto miners who might think this is a “buy and forget” play.
Risk 1: Intel execution failure. Intel has a history of delays. Its 18A node is scheduled for 2025, but if it slips to 2026 or 2027, the equipment orders get pushed back. Worse, if Intel’s foundry fails to win any major external clients (e.g., Nvidia, AMD, or a Bitcoin ASIC designer), the entire capex plan may be cut. In that scenario, the $3B uplift evaporates. The three Japanese stocks would then revert to their TSMC/Samsung baseline, which is still strong but without the Intel premium. The crypto mining community should watch Intel’s foundry announcements closely: if a major miner like Bitmain announces an Intel 18A-based chip, that’s the signal.
Risk 2: Geopolitical squeeze. The U.S. CHIPS Act provides $52 billion in subsidies but mandates that recipients “not expand advanced semiconductor manufacturing in China.” More importantly, the Commerce Department could issue guidance forcing Intel to prioritize American equipment vendors (Applied Materials, KLA, Lam Research) for national security reasons. That would shrink the Japanese vendors’ share of Intel’s capex. Tokyo Electron is most vulnerable—its etching tools compete directly with Lam’s. Lasertec and Disco have stronger moats, but investors should not assume the status quo.
My own experience: During the 2020 DeFi summer, I audited a flash loan attack on Uniswap V2. I tracked wallet clusters moving liquidity out before the exploits were public. That taught me to look beyond the surface narrative. Similarly, today’s narrative around Intel and Japanese equipment makers hides a deeper structural shift: the real growth is in advanced packaging for AI, not in Intel’s foundry success per se. Crypto miners and investors should focus on Disco’s role in chiplet packaging, because that demand is secular, not dependent on Intel’s execution.
Takeaway: What to Watch Next
Security is a promise; liquidity is the proof. In this case, the liquidity is in the order books of these Japanese firms. Over the next 6-12 months, track three signals: 1. Intel’s quarterly earning calls: any mention of 18A yield and external customers. 2. Disco’s backlog growth: If it accelerates beyond 20% YoY, the chiplet trend is real. 3. The U.S. CHIPS Act Office’s guidelines on equipment procurement.
If Intel delivers, Lasertec and Disco are multi-baggers. If not, they still hold value from TSMC and Samsung orders. The downside is limited; the upside is asymmetric. But do not buy the hype without the data. Fast money leaves fast scars.
Chaos is just data waiting to be organized. This Intel capex surge is data. Organize it, and you’ll see the next move before the market does.