Europe's quiet dependence on Chinese legacy chips

Legacy chips — made on established process nodes at 28nm and above — attract little attention next to the advanced-node story. They are also where Europe's actual semiconductor dependency sits, and the 2025 disruption at a single packaging plant demonstrated it more clearly than any policy paper.

China's share of global mature-node capacity has passed 30%, with a stated target of 50% by 2030 and projections of nearly 47% of global mature process capacity by 2027.

Where China already leads

Not everywhere, and the specifics matter.

Specialty process. Hua Hong Semiconductor is a global leader in BCD process technology, the core specialty process for analogue and power chips. SMIC holds a top-three global position in foundry services at 28nm and above.

Analogue and programmable logic. Chengdu Huahong's 8-bit 64GSPS ultra-high-speed ADCs perform comparably to international products, in applications spanning aerospace, radar and 6G. In FPGAs, Chinese firms have mass-produced 70 million-gate parts and are developing hundreds-of-millions-of-gates devices.

Packaging and test. China holds 46% of the global market, with JCET among those deploying chiplet techniques.

Power devices. BYD Semiconductor ranks second globally in automotive IGBT market share, behind only Infineon, with fully domestic mass production. GigaDevice and SG Micro lead in cost-effective 32-bit general-purpose MCUs and power management ICs across appliances, industrial control and fast charging.

The next push is explicitly up-market: high-end analogue (ADC/DAC) where European and US firms still dominate, full-series automotive-grade chips including ASIL-D MCUs and domain control SoCs, wide bandgap SiC and GaN — which run on mature nodes and existing legacy fabs — long-lifecycle industrial MCUs and industrial Ethernet chips for EtherCAT and PROFINET, and high-reliability automotive memory.

The price war is real, and it is segment-specific

A useful corrective to the assumption that Chinese capacity is uniformly subsidised into dumping.

Support comes mainly as equity investment, R&D subsidy, tax incentive and land policy — not direct product price subsidy. But in low-barrier segments the competition is genuinely destructive. In consumer-grade 8-bit MCUs, entry-level PMICs, low-density consumer NOR Flash and basic discrete devices, prices fell 30–50% between 2023 and 2025, with some firms selling at or below marginal cost and widespread losses among smaller players. Subsidy has let loss-making firms survive longer than they otherwise would, prolonging the fight.

In automotive-grade, industrial-grade and high-end analogue, there is no such competition. Technical barriers are high, automotive qualification runs two to three years, functional safety requirements are strict, and only a handful of Chinese firms can mass-produce. BYD Semiconductor and CXMT have explicitly abandoned low-price strategies for high-margin positions.

M&A guidelines introduced in 2024 encourage consolidation to remove inefficient capacity. The shakeout is under way.

Worth noting that Chinese manufacturers describe the pressure as running the other way too — aggressive price reduction by American analogue chip majors squeezing domestic firms, which is part of the argument for protective tariffs on imported chips to restore a profit-to-R&D cycle.

What Europe actually depends on

This is the section European industrial planners should read twice.

In products:

• European automotive, industrial and consumer manufacturers depend on Chinese volume production of 8-bit and 32-bit MCUs, PMICs, discrete power devices and NOR Flash, where European capacity is insufficient
• Volkswagen, BMW, Mercedes-Benz and Stellantis depend on Chinese production of automotive-grade IGBTs, MOSFETs and discretes. Nexperia's Dongguan facility handles over 70% of the firm's global automotive chip packaging capacity, and European automakers faced production halts in 2025 when wafer supply to that plant was disrupted
• Europe's solar and energy sector relies on Chinese IGBTs, MOSFETs and SiC devices, where Chinese suppliers hold over 60% global share
• Even European equipment leadership has dependencies underneath it: subsystems including vacuum pumps and high-purity gas filtration draw partly on Chinese suppliers, and 67% of processor cooling fans are imported from China
• Industrial IoT, smart home and medical device makers depend on Chinese low-power MCUs, MEMS sensors and connectivity chips at cost and scale European suppliers cannot match

In value-added steps:

• European fabless and fab-lite firms depend on SMIC and Hua Hong for 28nm-and-above manufacturing, because Europe has insufficient 8-inch and 12-inch legacy fab capacity
• Almost all major European chip firms rely on Chinese OSAT providers — JCET, TF Micro, Huatian — for back-end packaging and test of legacy parts
• Upstream equipment and materials firms depend on China for quartz, ceramic components, special gases, sputtering targets and packaging substrates
• 94% of Europe's gallium and 87% of its germanium are imported from China

That last pair is the one that turns a commercial dependency into a strategic one, because both are subject to Chinese export controls.

European firms have responded by moving in, not out

The revealing thing is what European manufacturers have actually done, as distinct from what European policy says.

Infineon opened its largest overseas IGBT wafer fab in Wuxi. STMicroelectronics established a SiC wafer fab joint venture with Sanan Optoelectronics. NXP expanded manufacturing and R&D in Tianjin and Chongqing. European firms have moved from back-end packaging in China to full front-end legacy-node fabs, and are increasing their procurement of Chinese wafers, materials and equipment for those fabs — both for cost and to meet local content expectations from Chinese customers.

Many European fabless firms have shifted mature-node foundry orders to SMIC and Hua Hong. The structure of choice has moved toward joint ventures, technology licensing and co-development rather than wholly owned facilities, as a hedge against geopolitical risk.

Meanwhile the market position has changed underneath them. China is 41.5% of global chip consumption. European firms have lost mid-to-low-end share and retreated to high-margin niches — ASIL-D automotive, high-reliability industrial MCUs, premium power modules. Demand has shifted from general-purpose commodity parts to customised solutions for NEVs, PV inverters and industrial automation, requiring faster development cycles than standardised global products allow. And Chinese OEMs increasingly mandate dual-supply, mixing European and domestic chips, which puts a hard ceiling on European share growth even as the market expands.

The export strategy is deliberately below the control threshold

China's approach to Europe in legacy chips is coherent, and its first principle is avoiding the fight it would lose.

By concentrating on 28nm and above — largely exempt from US export controls — Chinese firms access the European market without sanctions exposure, positioning legacy chips as a stable alternative amid supply chain volatility.

The sequence has been: enter with high-volume commodity parts and take share in price-sensitive segments; then move upward into automotive and industrial legacy chips certified to EU standards — AEC-Q100, ISO 26262, CE marking — with BYD Semiconductor, StarPower and GigaDevice running dedicated European product lines; then compete on customisation, offering IGBTs and SiC optimised for European PV inverter makers and chips designed for European EV platforms, with faster cycles than incumbents.

Localisation supports it: regional headquarters, R&D centres, sales offices and logistics hubs in Germany, Italy, France and the Netherlands, providing on-site technical support and after-sales service. Joint ventures with European OEMs, Tier 1s and automation firms give European partners stable supply and co-development.

To pre-empt trade restrictions, firms are adopting a China design, global manufacturing, European local service model, with packaging and test capacity in Southeast Asia and Eastern Europe to spread tariff risk.

And on the state side, the recent posture favours exemptions and licences over restriction — when the Dutch government moved on Nexperia, China granted an exemption permitting exports to resume for civilian use, keeping the supply chain intact.

What this means if you are on the European side

Map your legacy-node exposure before your advanced-node exposure. The advanced-node debate is strategically important and commercially distant. The dependency that halted European car production in 2025 was packaging capacity in Dongguan.

Gallium and germanium are the chokepoints with no substitute. At 94% and 87% import dependence, and with export controls already demonstrated as an instrument, these deserve treatment as a sovereign supply question rather than a procurement one.

Trade restrictions on Chinese legacy chips would hit European buyers first. Europe lacks the mature-node fab capacity and the OSAT capacity to absorb the volume. Restricting supply does not create European alternatives; it raises costs while the dependency persists.

The relationship has moved from China as a consumer market and production base for European chipmakers to China as a supply chain competitor and rule-making participant. European firms have adapted commercially — building fabs, forming joint ventures, buying Chinese inputs — considerably faster than European policy has recognised.