Inside China’s Power Semiconductor Shakeout
China’s Power Semiconductor Industry Is Moving from Domestic-Substitution Dividends into a Brutal Period of Differentiation
Power semiconductors are the foundational devices of China’s new energy, AI data-center, solar-storage-charging, and industrial electrification systems. The most important signal in this industry now is that growth, overcapacity, price increases, price cuts, automotive qualification, and profit pressure are appearing at the same time.
This essay is part of China Industry Signals.
Executive Summary
China’s power semiconductor industry is entering a new stage: demand remains structurally strong, but competition, pricing pressure, depreciation, and technology gaps are forcing a major shakeout.
EVs, 800V platforms, charging infrastructure, solar-storage systems, AI server power supplies, industrial power, and low-altitude-economy applications are still expanding the addressable market for IGBT, MOSFET, SiC, and GaN devices.
Chinese companies are making real progress in vehicle-grade IGBT modules, SiC MOSFET validation, domestic wafer supply, and vertically integrated IDM capacity.
The best recent signals include Sanan’s SiC deliveries, Silan Micro’s vehicle-grade IGBT module recognition, domestic SiC MOSFET validation at automakers and overseas Tier 1 suppliers, and rising demand from AI power systems.
The bad news is equally important: low-end MOSFET and some SiC capacity face overexpansion, price competition, profitability pressure, heavy depreciation, and uneven product reliability.
China’s power semiconductor industry is moving from “substitution by price” toward “competition by reliability, vehicle qualification, module integration, customer stickiness, and system-level power electronics capability.”
Three Egrets (三鹭鸶), by Lin Fengmian (林风眠), 1950s
Its poised balance between separation and coordination mirrors the core argument of this essay: power semiconductors derive their strength not from a single device, but from an ecosystem of materials, manufacturing, applications, and industrial scale working together.
1. Power Semiconductors Are the Foundational Chips of China’s Electrified Economy
Power semiconductors do not attract public attention the way GPUs, AI models, and advanced logic chips do. But they are more like the foundational valves of the modern electrified economy. Every EV, every solar inverter, every energy-storage PCS, every fast-charging pile, every AI server power supply, and every industrial inverter needs power devices to convert, control, and protect electricity.
In the internal-combustion era, power semiconductors were not at the center of the automotive value chain. In the EV era, they enter the traction inverter, onboard charger, DC-DC converter, thermal-management system, and high-voltage platform directly. As automobiles shift from mechanical systems to electrical systems, power devices move from peripheral components to core devices. 800V platforms, ultra-fast charging, high-voltage electric drives, lightweighting, and higher-efficiency requirements are pushing SiC, IGBT, and high-end MOSFETs to the front line.
Over the past two months, Chinese domestic industry-media coverage of power semiconductors has shown a very interesting duality. On one side, EVs, solar-storage systems, AI power, the low-altitude economy, and industrial power continue to drive demand. Domestic companies are repeatedly releasing progress in automotive qualification, SiC mass production, IGBT modules, and domestic wafer supply. On the other side, the industry is also clearly feeling price wars, low-end overcapacity, depreciation pressure, falling profits, and insufficient high-end reliability.
This shows that China’s power semiconductor industry has passed the most comfortable phase of domestic-substitution dividends and is entering a more real and more brutal period of differentiation.
Power semiconductors are not simple chip replacements. To enter automotive, industrial, and energy systems, they must pass long-cycle validation. Customers care not only about price, but also efficiency, losses, temperature stability, short-circuit capability, lifetime, packaging reliability, batch consistency, and system compatibility. Chinese companies can use price and responsiveness to enter mid- and low-end markets, but to truly enter high-end automotive-grade, industrial, and energy systems, they must prove long-term reliability.
This is also the most valuable signal in recent domestic reporting: the industry is no longer only talking about “higher localization rates.” It is now facing a harder question — which companies can move from low-price substitution to high-reliability substitution?
2. SiC Is Still Growing, but Voices Saying “It Is Too Miserable” Have Already Appeared
SiC, or silicon carbide, has been one of the most sought-after areas in power semiconductors over the past few years. It is suitable for high-voltage, high-frequency, high-temperature, and high-efficiency scenarios. Compared with traditional silicon-based IGBTs, it has clear advantages in 800V EVs, fast charging, solar inverters, energy-storage PCS, and industrial power. Domestic reports generally believe that as 800V high-voltage platforms penetrate from high-end models into mid-range models, demand for SiC MOSFETs in traction inverters will continue to grow.
Industry data also supports the long-term growth logic. Domestic industry reports have noted that the value of SiC devices required by an 800V new-energy vehicle is about RMB 400 to 800, roughly three to five times that of a traditional 400V silicon-based solution. Solar and energy storage are the second-largest battleground, with SiC MOSFET penetration rising in string inverters and energy-storage PCS. Ultra-fast charging piles, especially those above 480kW, create rigid demand for high-voltage and high-current SiC devices. AI server power supplies are also beginning to be viewed as a new growth point.
But the other side is equally striking. Investment and industry media have recently used very direct headlines such as “Silicon carbide is too miserable.” These reports are not denying SiC’s long-term demand. They are pointing to stage-specific problems: expansion has been too fast, prices are falling, profitability pressure is rising, some companies are increasing revenue without increasing profits, and depreciation plus price competition are causing sharp profit declines.
StarPower is a typical example. As one of China’s leading domestic IGBT companies, its 2025 revenue reached a new high, but profitability was already under obvious pressure. Reports noted that in 2025 StarPower achieved operating revenue of RMB 4.012 billion, up 18.34% year-on-year; net profit attributable to shareholders was RMB 405 million, down 20.18%; and net profit excluding non-recurring items was RMB 377 million, down 22.61%. By the first quarter of 2026, pressure intensified further: main business revenue was RMB 864 million, down 6.0% year-on-year; net profit attributable to shareholders was RMB 26.6268 million, down 74.32%; and net profit excluding non-recurring items was RMB 16.7936 million, down 83.49%.
This shows that demand growth in power semiconductors does not automatically translate into corporate profit growth. Especially during SiC and IGBT expansion cycles, if industry supply is released too quickly, customer price-cutting pressure is too strong, production-line depreciation is too heavy, and product structures have not fully upgraded, companies can fall into the classic pattern of “revenue growth without profit growth,” or even profit collapse.
The real situation for SiC is therefore more complicated. It is not that there is no demand. Rather, demand growth, capacity expansion, price declines, and technological differentiation are happening at the same time. Low-end or homogeneous SiC products may face price wars, while high-reliability automotive-grade products and high-end industrial applications still have validation barriers. The industry will move from a stage where “everyone can tell a SiC story” to a stage where the key question is: who can stably deliver automotive-grade SiC, control costs, and pass long-term validation by automakers and Tier 1 suppliers?
3. Sanan: SiC Is Moving from a Capacity Story to a Customer-Validation Story
Hunan Sanan, under Sanan Optoelectronics, is one of the most important companies in recent domestic SiC reporting. In early May, Sanan disclosed the latest progress of its silicon-carbide business through investor interactions and industry-media reports, and the details were highly specific.
On capacity, Hunan Sanan already has supporting 8-inch SiC capacity of 1,000 wafers per month; Chongqing Sanan has current capacity of 3,000 wafers per month; and the Sanan-ST joint-venture production line currently has capacity of 2,000 wafers per week, with some products already completing validation and entering risk production.
In the automotive market, Hunan Sanan’s SiC MOSFETs for traction inverters have passed validation at leading domestic EV customers and are also being introduced for validation at several overseas Tier 1 customers. SiC MOSFETs for onboard chargers and air-conditioning compressors have also been sampled to or achieved small-batch shipments with several automaker customers.
This information is more valuable than ordinary “capacity expansion” news. The real key in power semiconductors is customer validation, especially automotive qualification. New-energy vehicle customers will not casually replace core traction-inverter devices just because of domestic substitution. Traction SiC MOSFETs must withstand high voltage, high temperature, high frequency, high current, and complex operating conditions. Their failure is not merely a chip problem; it is a vehicle reliability and safety problem.
Sanan also has one detail with strong field texture. On May 7, the company announced cooperation with Megmeet. Sanan SiC chips had been mass-produced and introduced into Megmeet’s customized power solution for DJI drones, passing multiple rounds of flight testing. Data showed that in the first quarter of 2026, Hunan Sanan delivered more than 600,000 silicon-carbide chips to Megmeet, helping the customer shorten the development and introduction cycle for new products by 30%.
This shows that SiC applications are expanding from traditional EVs and solar-storage systems into drones, the low-altitude economy, and high-power-density power-supply scenarios. Drones have different requirements from automobiles, but they also emphasize efficiency, weight, temperature, reliability, and dynamic operating conditions. Whether the low-altitude economy can form a large market still needs observation, but the entry of SiC into such scenarios shows that domestic SiC devices are seeking diversified customers, rather than relying only on a few automaker traction-inverter projects.
Sanan’s advantage lies in vertical integration. It covers SiC substrates, epitaxy, chips, and packaging-related capabilities, and has formed a joint-venture production line with ST. Domestic reports emphasize its integrated push from substrate to device and from China to overseas markets. This model is especially important in the SiC industry, because substrate quality, epitaxial defects, chip design, manufacturing processes, and packaging reliability all influence one another. Strength in a single link does not necessarily guarantee final device reliability.
But this also means heavy capital expenditure and depreciation pressure. SiC vertical integration is not an asset-light story. Production lines must ramp, customers must validate, yields must improve, and prices may still fall. Sanan’s progress shows that domestic SiC is entering the stage of customer validation and batch delivery, but the final outcome will still depend on reliability, cost, and long-term customer relationships.
4. Silan Micro: The IDM Model and Automotive-Grade Power Modules
Silan Micro is another important sample in China’s power semiconductor industry. It is not only a single-device company, but an IDM covering design, manufacturing, packaging, and testing. Its products include power semiconductors, IPM modules, MEMS, and analog circuits. Domestic reports note that Silan Micro covers 5-inch to 12-inch silicon-based production lines and is expanding into compound-semiconductor production lines, making it one of the few domestic semiconductor companies with relatively complete IDM capability.
Over the past two months, Silan Micro’s progress in automotive-grade power modules has been worth recording. On May 22, Silan Microelectronics disclosed that its B7 power module SGM270SS8B7TFM for new-energy vehicles had been selected as one of the “Top Ten New Domestic Automotive-Grade Chips of 2026” and received strong evaluation. The product uses Silan Micro’s self-developed fifth-generation Trench-Field-Stop IGBT process and features low stray inductance, high blocking voltage, high power density, and high reliability.
This product signal has two layers of meaning.
First, IGBT has not been completely replaced by SiC. Although 800V platforms and high-end EVs are driving SiC growth, IGBT will still maintain long-term demand in many 400V platforms, hybrid vehicles, industrial control, solar power, rail transit, white goods, and mid- to low-voltage high-power scenarios. SiC is not replacing everything; it is expanding its share in high-voltage and high-efficiency scenarios. Chinese companies’ accumulated capability in IGBT remains very important.
Second, automotive-grade power modules are a higher-value form of competition than single chips. Automotive customers need modules, system compatibility, and reliability, not isolated chips. Power modules must solve chip paralleling, heat dissipation, package parasitics, insulation, soldering, bonding, thermal cycling, and mechanical reliability. Module capability determines whether domestic devices can truly enter high-value vehicle-level links.
Silan Micro’s IDM model has advantages at this stage. Unlike logic chips, many high-reliability power products depend heavily on process stability, device design, packaging, and application coordination. IDM companies can better control manufacturing processes, package structures, and quality systems. For domestic power semiconductor companies to enter automotive, industrial, and energy customers, IDM capability will become increasingly important.
But Silan Micro cannot escape industry pressure either. Price increases and shortages may appear in some power-device products, but price wars still exist in low-end devices. The IDM model has heavy capital expenditure and depreciation, and if product-structure upgrading cannot keep pace, profitability will also be pressured. The key question for Silan Micro is not only capacity, but whether it can continuously scale automotive-grade IGBT, SiC, IPM, and module products, and build stable share among high-end customers.
5. Jiejie Microelectronics’ Price Increases Show That the Industry Is Not Simply in a Price War
Power semiconductors have another interesting phenomenon: some products face price wars and profit pressure, while some companies are announcing price increases.
On April 17, Jiejie Microelectronics disclosed an investor-relations activity record, noting that due to persistently high raw-material prices and tight capacity, it was adjusting prices for main products including thyristors, protection devices, MOSFETs, and IGBTs. Among them, finished MOSFET prices were raised by 10% to 20% from February 1, 2026; finished IGBT prices were expected to rise by 10% to 20% from May 1, 2026. The company said the price adjustment was intended to respond to rising production costs and ensure supply-chain stability and long-term operating capability.
This case shows that the power semiconductor industry is not in “comprehensive overcapacity.” Differentiation across products, applications, and customer tiers is very large. Low-end consumer-grade MOSFETs may suffer price wars, while automotive, industrial, specific protection devices, some IGBTs, and high-reliability devices may still have room for price increases due to raw materials, capacity structure, and customer demand.
This differentiation is critical. Power semiconductors are not one single market, but a collection of markets. Consumer electronics, white goods, industrial control, automotive traction inverters, OBCs, solar inverters, energy-storage PCS, server power supplies, the low-altitude economy, power grids, and rail transit all have completely different device requirements. Low-end products compete on price; high-end products compete on reliability, qualification, and customer binding.
Jiejie Microelectronics’ price increase also shows that raw-material and supply-chain costs still affect the industry. Power semiconductors are not like pure design companies. They are deeply related to wafer manufacturing, packaging, metal materials, ceramic substrates, copper materials, molding compounds, and testing capacity. Upstream cost increases can pass through to device prices, but whether price increases can be accepted depends on the product competitive landscape and customer tolerance.
So the current industry is not simply “good” or “bad.” It is structurally differentiated: low-end is oversupplied, high-end remains tight; some companies’ profits are falling, some products are rising in price; SiC has long-term growth but short-term internal competition; domestic substitution is progressing, but high-end customers remain cautious.
6. The Nexperia Incident Exposes the Geopolitical Risk of Globalized Power Semiconductors
There is another special case in power semiconductors: Wingtech Technology and Nexperia.
Nexperia was originally NXP’s standard-products business and was later acquired by Wingtech Technology. It was one of the largest cross-border acquisitions in Chinese semiconductor history. Nexperia has global standing in diodes, transistors, MOSFETs, protection devices, and automotive-grade standard products, and it is also an important asset for Chinese companies participating in the global power and standard-device markets.
But over the past year, the Nexperia incident has become a typical case of geopolitical and corporate-governance risk. Domestic media recently reported that due to U.S. sanctions and Dutch intervention, Wingtech Technology fell into a crisis in which Nexperia’s overseas entities became “out of control.” The company’s stock has been subject to delisting-risk warning since May 6, 2026, and its stock abbreviation was changed to “*ST Wingtech.” Financial reports show that *ST Wingtech recorded a net loss of RMB 8.748 billion in 2025, and in the first quarter of 2026 its net profit attributable to shareholders turned from profit to loss year-on-year, falling more than 172% to negative RMB 189 million.
This case has enormous significance for China’s power semiconductor industry. It shows that global acquisitions do not automatically bring stable control. Chinese companies previously entered global semiconductor supply chains through overseas acquisitions, but as geopolitical pressure rises, overseas assets, governance structures, employee permissions, customer relationships, and supply chains can all be subject to intervention.
Nexperia China has also begun shifting wafer supply to local sources. Earlier domestic reports noted that Nexperia China had locked in domestic wafer fabs to ensure that from 2026 onward, wafer supply for core products such as IGBTs would shift toward domestic sources. This move is defensive, but it also has long-term industrial significance. It shows that China’s power semiconductor supply chain is being forced to localize not only because of cost and substitution logic, but also because globalized assets themselves have become unstable.
The Nexperia incident is bad news and also a warning. The bad news is that globalized assets acquired by Chinese companies through overseas M&A may be torn apart by geopolitics, and Wingtech has also suffered huge financial and capital-market pressure. The warning is that China’s power semiconductor industry cannot rely only on overseas assets and brand channels. It must build real control in domestic wafers, packaging, customer validation, and system capability.
