MetaX’s Sales More Than Double on Chinese AI Chip Demand

Revenue rose to 1.6 billion yuan ($232M) in 2025, up ~115% from 743 million yuan a year earlier, driven by strong domestic AI chip demand as Nvidia is squeezed out of the Chinese market. Losses narrowed to 779 million yuan and this is the company's first earnings release since its December IPO, indicating rapid top-line expansion but continued unprofitability.

Analysis

China’s nascent domestic AI-chip demand will create a multi-layered uplift across the semiconductor value chain rather than a single-company win: expect durable utilization gains at mature-node foundries, a wave of orders for back-end packaging & test facilities, and incremental demand for local materials and mature-node equipment that global suppliers don’t prioritize. The mechanics matter — 16/28nm processes and advanced packaging capacity can scale faster than bleeding-edge EUV fabs, so expect revenue and margin expansion concentrated in incumbents that already have high-utilization lines; a 10–20% jump in utilization typically translates to ~200–400bps of incremental gross margin for packaging/test operators within 6–12 months.

Second-order pricing effects will show up in customer negotiations: as domestic AI model builders switch to locally sourced silicon, they will trade higher unit cost for supply security, allowing domestic suppliers to sustain 5–15% price premiums in the near term versus pre-shift levels. That in turn accelerates CAPEX at local fabs and backend players — watch quarterly capex guidance and equipment import approvals as leading indicators; a 3–6 month uptick in equipment order flow precedes visible revenue growth at testers/packagers.

Key risks are policy and technology: tighter export controls or a sudden re-opening to Western incumbents can both reverse momentum quickly. Catalysts that would validate continuation include multi-quarter share gains by domestic foundries in key cloud AI buyers, a sustained >80% utilization across packaging plants, or announced multi-year supply deals with national champions; negative catalysts include a pause in government subsidies, three-plus quarters of underutilization, or a major defect discovery in new chips.

Key Decisions for Investors

• Long SMIC (0981.HK) — 12–18 month horizon. Size as a thematic overweight (3–5% NAV). Rationale: direct exposure to mature-node fab demand and likely subsidy support; target +40–70% upside if utilization and ASPs sustain, stop-loss at -30% from entry or on signs of subsidy pullback.
• Long JCET / major Chinese packaging & test leader (600584.SH) — 6–12 months. Trade: buy-to-open a 6–9 month call spread to limit capital outlay. Rationale: near-term margin leverage from higher utilization; upside 25–50% vs downside limited to premium paid. Roll on stronger orderbook prints.
• Pair trade: Long SMIC (0981.HK) / Small hedge via NVDA (NVDA) put options — 3–6 months. Size hedge as 10–15% of the long position. Rationale: protect against geopolitical repricing that would shrink China addressable market; if NVDA puts spike, reduce long exposure.
• Long ASE (3711.TW) or Taiwan packaging specialist — 9–12 months. Trade: purchase shares or 12-month calls sized modestly. Rationale: benefits from outsourced packaging demand; target 20–40% realized gain if cross-strait shipping remains stable, cut at -25% or on orderbook deterioration.