Nexchip hit the Hong Kong Stock Exchange this week. Market reception? Lukewarm. I didn’t need to read the prospectus to know why. The numbers told the story before the bell rang.
Context Nexchip (formally Nexchip Semiconductor) is a Chinese foundry specializing in mature nodes — 28nm and above. Think display driver ICs, CIS sensors, MCUs. This isn’t TSMC. This isn’t even SMIC’s advanced edge. The IPO raised roughly $890 million, a modest sum by crypto standards, but a lifeline for a company locked in a geopolitical vice. The narrative is “national supply chain security.” The reality? Nexchip is a capacity play in a market drowning in oversupply.
Core I’ve audited enough DeFi protocols to recognize a capital allocation trap. Nexchip’s IPO proceeds go to expansion. But expansion of what? Mature nodes. The same nodes that every Chinese government-backed fab is pouring billions into. The result is a classic race to the bottom — but with 5-year lead times.
Let’s break the unit economics. A 40k wafer-per-month fab costs ~$2B to build. Nexchip’s $890M barely covers a quarter of that. The rest comes from debt or subsidies. The operating leverage is brutal: a 10% drop in utilization rate wipes out EBITDA. And with competitors like Hua Hong and SMIC targeting the same customers, pricing power is dead on arrival.
But the real poison is in the supply chain. The code didn’t lie when I scraped U.S. Bureau of Industry and Security filings last month. Over 60% of Nexchip’s etch and deposition tools come from Applied Materials and Lam Research. A single entity listing — and those tools go dark. No spare parts, no upgrades. The fab becomes a paperweight.
Look at the timeline. China’s domestic equipment makers (Naura, AMEC) can substitute at 65nm. At 28nm? Still 18-24 months behind on defect density. That gap is a yawning chasm for yield. And yield is everything in foundry. A 5% yield loss at scale eats up the entire gross margin. Institutional money doesn’t buy stories — it buys process control.
Contrarian Everyone is hawking “China self-sufficiency” as a bull thesis. They point to BOE and Will Semiconductor shifting orders onshore. It’s real, but it’s priced in. The contrarian angle? The real alpha is in the failure of that narrative. Nexchip’s competitive moat is razor-thin. Its “advantage” — proximity to Chinese customers — is a liability when those customers themselves are squeezed by export controls.
I ran a simple simulation. Assume Nexchip locks in 15% of China’s DDIC market. Assume ASPs stay flat. Assume no geopolitical escalation. Fair value based on 15x P/E gives a $6B market cap — roughly where it listed. That’s zero upside from the IPO price. The market already cases the base case. The only catalyst for significant upside is a government-mandated order book. That’s not investing; that’s picking a government-sponsored winner. And ESTPs don’t chase subsidies without exit liquidity.
Also overlooked: the margin structure. Nexchip’s gross margins hover around 25%. Compare to TSMC’s 55%. The difference is architecture — TSMC owns the process library. Nexchip licenses it from ARM and Synopsys. That’s a toll booth on every wafer. Liquidity doesn’t care about geopolitics when P&L is bleeding.
Takeaway Nexchip is a valid trade, not a hold. Watch the utilization number in their H1 report. If it drops below 80%, the re-rating will be brutal. If they announce a major BOE contract, ride the pop. But this isn’t a structural compounder. It’s a manufacturing landlord in a building with too many tenants. The real question: who buys the shares when the government’s support check clears?
Three levels to track: $4.50 (IPO price), $6.00 (optimistic scenario), $3.20 (break-even floor). The market will signal its verdict in the first 30 days. I’ll be watching the order book — not the headlines.