The Slurry Problem

A process engineer at Intel’s Hillsboro, Oregon fab watches the cost of abrasive slurry tick upward for the third consecutive quarter. CMP slurry is a liquid mixture of nanoscale particles that grinds silicon wafers to atomic smoothness. Without it, no advanced chip gets made. The slurry’s key ingredients—antimony compounds and tungsten particles—now flow through a gate that opens and closes from Beijing.

The Mine That Sits Idle

China controls more than 80 percent of global primary antimony production and comparable tungsten supply. What changed is that Beijing began using them as active leverage—imposing export restrictions that function less like permanent blockades than like valves, tightened during confrontation and loosened during negotiation.

The price signal was immediate. Tungsten rallied 557 percent from pre-restriction levels, driven partly by intensive US military consumption that drained domestic stocks—a spike severe enough to briefly exceed gold’s per-kilogram cost. Washington responded with a $500 million critical minerals initiative aimed at building non-Chinese supply chains through companies like United States Antimony and Perpetua Resources. But those projects stretch years to completion.

Beaver Brook, a Canadian antimony mine capable of producing roughly 6,000 tonnes of concentrate annually—about 5 percent of global supply—sits idle. Its owner is China Minmetals. Analysts flagged rumors that China Minmetals could selectively restart the mine to flood the market and depress prices, undercutting US-funded competitors before they reach scale. The mechanism is straightforward: restrict supply to raise prices, then selectively release supply to destroy rival project economics.

A valve you can close at any time disciplines behavior even when it’s open. Export controls are not trade policy. They are infrastructure.

288 Cores and a Yield Curve

Those mineral restrictions land directly on the fab floor where Intel is trying to build America’s semiconductor future. Intel’s 18A process node requires tungsten for interconnects and antimony-bearing compounds in CMP slurries. The global CMP slurry market, currently valued at $3.5 billion, is projected to reach $6.1 billion as chipmakers push to smaller geometries. A significant share depends on materials that currently flow predominantly through China.

Intel CEO Lip-Bu Tan has staked the company’s foundry future on 18A, announcing a 288-core Xeon processor as the node’s flagship product. But analyst Azam Barkatullah of Foretag Consulting estimates that 18A won’t reach industry-standard yield levels until 2027. That timeline creates a window during which Intel’s mineral consumption ramps—ordering more tungsten, more slurry—while yields remain too low to generate revenue justifying the investment.

Molybdenum and cobalt are being explored as tungsten alternatives for interconnects at sub-2nm nodes. But “being explored” and “qualified for volume production on a node already behind schedule” are separated by years of process development. You cannot pause an 18A ramp to requalify alternative materials without falling further behind TSMC.

Yield curves don’t negotiate with export ministries.

The Substrate Trap

Intel’s escape route from semiconductor dependence on Asia runs directly through a mineral supply chain that Asia’s dominant power can throttle at will. The $500 million critical minerals initiative and CHIPS Act foundry subsidies aim at domestic chip production but operate on incompatible timelines. Intel needs tungsten and antimony-grade slurry compounds now, in increasing volumes, as it ramps 18A through 2026 and 2027. Mining projects meant to provide non-Chinese supply are likely years from meaningful production—industry timelines typically stretch five to seven years from funding to output. China Minmetals’ Beaver Brook option means even those projects face price-suppression risk upon approaching viability.

When ASML export controls restricted China’s access to advanced lithography equipment, Beijing accelerated domestic alternatives—but the timeline mismatch between restriction and substitution created years of vulnerability. Washington is watching the mirror image: restricting its own dependence on Chinese minerals while the substitution timeline lags the fabrication timeline by half a decade.

By investing billions in domestic fabs without parallel mineral sovereignty measures, Washington may have shunted pressure from a visible fabrication chokepoint to a less visible materials chokepoint—one that procurement teams see daily but policymakers have been slower to address. That Hillsboro process engineer absorbs the consequence because Intel cannot refuse to buy slurry, cannot substitute it on the 18A timeline, and cannot source it outside China’s reach at required volumes.

Every quarter Intel spends ramping 18A at sub-standard yields, it consumes more material per good die produced. Waste rates at low yields mean more slurry, more tungsten, more exposure to the export gate. Every signal that 18A is struggling gives Beijing less reason to ease restrictions, because a delayed American fab is a weaker negotiating counterparty. China’s 92 percent control of rare earth processing has remained “structurally stable for nearly two decades” despite Western policy efforts. Someone decided to use the mineral gate precisely when the customer on the other side had no alternative.

What to Watch

Intel’s 18A yield data over the next two quarters will tell you whether mineral exposure stays manageable or becomes acute. If yields remain below 50 percent through Q3 2026, material consumption per good die stays elevated, and Intel’s procurement team faces a choice between building strategic slurry inventory at inflated prices or risking production gaps. Watch for any MOFCOM announcement adjusting tungsten or antimony export quotas during the next round of US-China trade talks, likely before September 2026.

A key signal is whether any US-allied government—Japan and South Korea being most likely given their semiconductor exposure—moves toward bilateral mineral stockpiling agreements for semiconductor-grade materials. If no such agreements emerge by late 2026, policymakers still believe fabrication investment alone is sufficient. I predict that announcement comes before year-end. If you see it, the thesis is tracking. If you don’t, the gap between fabrication ambition and material reality is wider than anyone in Washington wants to admit.

If This Thesis Is Wrong

The strongest competing explanation is simpler: Intel’s 18A challenges are primarily lithographic and architectural, not material-constrained, and mineral price spikes are commodity-market noise that procurement teams hedge routinely. CMP slurry costs are a rounding error in a fab’s operating budget, and the real bottleneck is EUV tool availability and process integration.

The 557 percent tungsten rally was driven substantially by military demand, not export restrictions alone—meaning the price signal overstates China’s leverage. If US military consumption normalizes, tungsten prices could retreat without any change in Beijing’s posture. The thesis weakens if foundry-grade antimony and tungsten compounds can be sourced from non-Chinese refiners already operating at scale, or if Intel’s 18A process uses alternative planarization chemistry that reduces antimony dependence.

The constraint isn’t in the chip. It’s in the polish.

Word count: 1,247

Market Implication

If this thesis holds, JKM LNG futures express the immediate risk—any Hormuz tension or cargo delay spikes spot prices as importers scramble for replacement cargoes within Taiwan’s 11-day window. The prediction market question is whether a major TSMC customer (Apple, NVIDIA, AMD) discloses LNG-related supply chain risk in SEC filings by December 2026, as the article explicitly forecasts. The kill signal is Taiwan’s Guantang terminal decision by September 2026: environmental clearance means structural buffer expansion, invalidating the 11-day constraint. The sleeper trade is long Cheniere Energy—Taiwan’s 25-year supply deal starting June 2025 positions Cheniere as the primary beneficiary if Taiwan accelerates contracting to address this vulnerability, a second-order effect not yet priced into LNG equity.

Analytical implication, not financial advice.

Sources