By Arjun Mehta, Lead Analyst, Power Electronics and Semiconductors | Excellence Research | June 18, 2026
When Wolfspeed broke ground on its 200mm silicon carbide substrate facility in Siler City, North Carolina in September 2022, the project attracted USD 1.3 billion in announced investment and a visit from the U.S. Secretary of Commerce. Three and a half years later, that groundbreaking looks like the opening act of a supply chain arms race that has drawn in governments, sovereign wealth funds, and automotive giants on four continents.
The silicon carbide power semiconductor market reached USD 3.21 billion in 2026, according to new data from Excellence Research. By 2033, the firm projects that figure will climb to USD 10.47 billion, growing at 18.4% annually. For context, the broader semiconductor market is forecast to grow at roughly 7 to 8% over the same period. SiC is not tracking with its industry. It is lapping it.
The numbers make more sense when you follow the electric vehicles. A modern 800-volt EV traction inverter uses between 24 and 48 silicon carbide MOSFETs. At current pricing, that represents a silicon carbide content value of USD 91 to USD 298 per car, compared to roughly USD 27 to USD 85 for the silicon IGBT modules that SiC is replacing. The cost premium is real, but so is the payback: SiC inverters operate at higher switching frequencies with lower on-resistance, cutting switching losses by up to 60% versus silicon. That efficiency gain translates to either more range from the same battery pack or the ability to downsize the battery entirely.
BYD built 3.02 million battery-electric and plug-in hybrid vehicles in 2025. Tesla produced 1.79 million. Volkswagen Group's EV portfolio across Audi, Porsche, and VW brand accounted for another 890,000 units. All three have active SiC qualification programs; Porsche's Taycan and Audi e-tron GT already run SiC inverters sourced from Infineon Technologies. The procurement math from these volumes explains why Wolfspeed, STMicroelectronics, and onsemi are all running their SiC fabs at or above rated capacity through at least 2028.
The solar connection is less widely discussed but almost as important commercially. String inverter manufacturers including SMA Solar Technology, Sungrow, and Huawei Smart PV converted the majority of their product lines to SiC-based topologies between 2023 and 2025. The efficiency improvement at the inverter level translates directly to more kilowatt-hours harvested per installed panel, a metric that determines project IRR in competitive auction markets. With over 420 GW of new solar capacity installed globally in 2025, string inverter SiC consumption added approximately USD 556 million in demand that essentially did not exist in 2021.
The competition for supply has created a scramble at the substrate level that most industry observers have underestimated. Silicon carbide boule growth is a technically demanding process: a 150mm SiC crystal takes 5 to 7 days to grow in a specialized furnace running at temperatures above 2,000 degrees Celsius. The total installed global capacity for SiC substrate production is estimated at roughly 14 million 6-inch equivalent wafers per year as of mid-2026, against demand running closer to 17 million. That gap, approximately 3 million wafers, is why lead times for SiC MOSFETs from tier-1 suppliers extended to 52 weeks at peak in 2024 and have only partially recovered to 32 to 38 weeks as of June 2026.
China is moving aggressively to address its import dependency. The Chinese government committed over CNY 28 billion to third-generation semiconductor development through its 14th Five-Year Plan, and the results are beginning to show. Sanan IC, which produces both GaAs and SiC devices, expanded its SiC substrate line to 10,000 wafers per month in early 2026. SICC, a substrate specialist backed by provincial funds in Shandong, claims 150mm wafer defect density of below 1 micropipe per cm2, a benchmark that was considered a western-exclusive capability as recently as 2022. TanKeBlue is targeting automotive-grade qualification by 2028.
The Western response has been to lean into the automotive qualification moat. STMicroelectronics holds design-win approvals across 85-plus vehicle programs as of end-2025, a position that took roughly eight years of sustained qualification work to build. Infineon's automotive-grade SiC MOSFET family, the CoolSiC series, is qualified on over 60 vehicle platforms and has the broadest voltage range coverage of any single vendor's product lineup. These qualification relationships are not easily replicated by a competitor offering 15% lower die cost; the cost of a failed power module in an operational EV is measured in warranty claims and reputational damage, not just component value.
Beyond vehicles, the report identifies data center infrastructure as the sleeper application for SiC through 2030. A 10 MW AI training cluster uses approximately 1,200 SiC module positions in its power delivery network. With over 85 GW of hyperscale data center capacity in the global planning and construction pipeline, the SiC content opportunity from this segment alone could add USD 1.2 to USD 1.8 billion in annual demand by 2030 at current pricing, a figure not yet well-reflected in consensus market estimates.
For investors and strategic planners, the key variable to watch is the 200mm wafer transition. The first production-grade 200mm SiC substrates from Wolfspeed's Siler City facility are expected in late 2027. When 200mm becomes the mainstream format, estimated substrate cost per cm2 falls by 35 to 42% relative to today's 150mm standard. That cost deflation will drive average selling price erosion for SiC devices, but it will also accelerate adoption into mid-range vehicle segments and consumer-facing applications where the current price premium is still a barrier. The companies that navigate the 150mm to 200mm transition with the best yield will define the competitive hierarchy for the next decade.