SiC Power Fabs

Silicon carbide (SiC) power semiconductor fabrication is the archetype where semiconductor supply chain constraints intersect most directly with the broader electrification buildout. SiC MOSFETs and Schottky diodes enable the high-voltage, high-efficiency power conversion that makes 800V EV architectures, grid-scale battery energy storage, high-efficiency solar inverters, and next-generation industrial drives viable. The demand curve for SiC devices is being pulled simultaneously by nine distinct markets — EV traction, battery energy storage system (BESS) power conversion, solar PV inverters, industrial variable-frequency drives (VFDs), high-voltage DC transmission, datacenter power supplies, EV charging infrastructure, rail traction, and specialty aerospace — and all nine draw their device supply from one physics-limited substrate production pipeline. This is the defining structural story of SiC: nine markets converging on one wafer funnel.

SiC fabrication is also the archetype most in structural flux during 2025–2026. The Western vertical integration model pioneered by Wolfspeed (boule growth + wafer production + device fabrication under one operator) has come under severe financial pressure, with Wolfspeed filing Chapter 11 bankruptcy in mid-2025 and undergoing restructuring. Meanwhile Chinese producers (SICC, TanKeBlue, Sanan Optoelectronics, BYD internal) are scaling aggressively with state support and capturing increasing share of both substrate and device output. The Western operator landscape that looked relatively stable in 2022 has been reshuffled by 2025, and the industry structure that emerges from this period will define the global SiC supply chain for at least the next decade.

Why SiC Is Physics-Limited, Not Capital-Limited

The central structural constraint on SiC production is that boule growth is approximately two orders of magnitude slower than silicon boule growth. Silicon single crystals pull from the melt at 1–3 millimeters per minute via the Czochralski process, producing a 300mm silicon boule in roughly two days. SiC single crystals grow at 0.3–0.5 millimeters per hour via the Physical Vapor Transport (PVT) sublimation process, producing a 150mm or 200mm SiC boule in 1–2 weeks. This is not a process-optimization problem that capital will eliminate — it is a thermodynamic constraint rooted in SiC's extreme stability and high sublimation temperature (~2400°C). Growth rate improvements at PVT are incremental; fundamentally faster growth approaches (top-seeded solution growth, HTCVD, and others) have been explored but none has displaced PVT at commercial scale.

The consequence is that SiC substrate supply cannot be rapidly expanded in response to demand shocks. A demand curve that doubles in 18 months requires boule growth capacity that was committed three years earlier in tooling orders, facility construction, and crystal growth operator hiring and training. Every SiC fab expansion runs up against this upstream constraint — a device fab can be built in 2–3 years, but the substrate supply to feed it takes longer to scale. The substrate bottleneck has driven the device operator vertical integration trend (Wolfspeed's captive substrate + device model; STMicro-Soitec partnership; Infineon-Siltectra acquisition) and has structured the industry's competitive dynamics around substrate access more than device design.

The 150mm → 200mm → 300mm Wafer Transition

SiC's primary industry-level productivity lever is wafer diameter scaling. A 200mm SiC wafer has 1.78× the area of a 150mm wafer, which translates to approximately 2× die output per wafer (accounting for improved edge utilization at larger diameter). The industry is mid-transition from 150mm to 200mm as of 2026, with leading Western operators having ramped 200mm capacity substantially while Chinese operators are catching up. A future 300mm SiC transition would provide another ~2.25× area scaling but faces substantial technical challenges (defect control, tool adaptation, epitaxy scale-up) and commercial production remains years out.

The 200mm transition is the single most important near-term variable in SiC supply. Operators that have completed 200mm ramp operate with a structural cost and throughput advantage over operators still primarily at 150mm. Chinese producers' catch-up on 200mm will determine the competitive structure of the industry through the late 2020s. If 300mm SiC becomes commercially viable in the early 2030s, it would represent a second step-function increase in substrate availability — but the technical and economic case for 300mm SiC is not yet proven.

The Four-Tier SiC Supply Chain

SiC device production flows through four distinct supply tiers, each with its own operator concentration and strategic characteristics. Understanding the tier structure is essential to understanding where supply chain constraints bind.

The tier concentration varies substantially. Boule growth is concentrated at roughly ten operators globally with technical capability at commercial scale. Epitaxial growth splits between merchant suppliers (IQE, Resonac) and captive operations at vertically integrated device makers. Device fabrication is the broadest tier with roughly fifteen operators globally. The substrate tier (boule + wafer) is the structural bottleneck — downstream tiers can be expanded more readily than upstream boule growth capacity.

Western Restructuring and Chinese Scaling

The 2025 operator landscape has been reshuffled by simultaneous Western financial pressure and Chinese aggressive scaling. Wolfspeed filed Chapter 11 bankruptcy in mid-2025 following a period of substantial capital expenditure (Mohawk Valley 200mm fab; The Wolf 200mm facility at Siler City NC) combined with lower-than-projected near-term SiC device revenue. The bankruptcy restructuring is reshaping Wolfspeed's capital structure and operator footprint, with consequences still being worked through. The broader Western SiC industry has seen margin pressure as Chinese substrate supply has grown and device pricing has compressed.

Meanwhile Chinese SiC producers have scaled rapidly with state industrial policy support. SICC (Shandong Tianyue Advanced Technology) has grown to become a significant global SiC substrate supplier. TanKeBlue (Dongguan) and Sanan Optoelectronics (Chongqing facility) operate substantial Chinese device capacity. BYD operates captive SiC production for its own electric vehicles, reflecting a fully vertical integration play from wafer to vehicle. Chinese SiC device output has risen rapidly and has begun serving both domestic Chinese EV demand and select export markets.

The strategic implication is that the SiC industry is shifting from a Western-dominated structure that existed through 2022 toward a more bifurcated global structure with substantial Chinese production alongside Western operators. The EU Chips Act and US CHIPS Act include SiC-relevant provisions, but neither program fundamentally offsets the scale and pace of Chinese capacity expansion. The industry structure that emerges over the next 2–3 years will shape SiC availability and pricing through the late 2020s.

Operator Landscape

Product Types: Schottky, MOSFET, JFET, Modules

SiC power devices divide into several product categories, each with distinct volume trajectories and application profiles. Understanding the product mix is relevant because the demand curve for MOSFETs (EV traction-driven) has different scaling than for Schottky diodes (mature, power supply rectification).

SiC Schottky diodes are the most mature SiC product, widely deployed since the mid-2000s in power factor correction (PFC) circuits and secondary rectification. Volume is substantial but growth is moderate — Schottky diodes are commodity products with competitive pricing. Every SiC device operator produces Schottky diodes as baseline portfolio.

SiC MOSFETs are the growth product and the primary driver of SiC capacity expansion. Voltage ranges from 650V (solar, datacenter) through 1200V (primary EV traction at 800V architecture), 1700V (BESS, industrial), and 3.3kV (traction converters, industrial VFDs). The trench MOSFET structure pioneered by Rohm has become competitive with planar MOSFETs (Wolfspeed, STMicro historical approach); both structures are produced at different operators with different tradeoffs.

SiC JFETs (junction field-effect transistors) are specialty products with normally-on operation characteristics. Small market share compared to MOSFETs; United SiC (acquired by Qorvo, later divested) is a specialty operator.

SiC modules — pre-packaged assemblies containing multiple SiC MOSFETs and diodes in half-bridge, full-bridge, or three-phase configurations — serve EV traction inverters, BESS power conversion systems, and industrial drive applications where integrated modules are preferred over discrete components. Module suppliers include most device operators plus specialty module integrators.

Yield Gap vs Silicon

SiC device yield runs structurally lower than silicon device yield because SiC substrates carry defect populations that silicon substrates largely do not. Micropipes (hollow tube defects through the crystal), basal plane dislocations (BPDs that create stacking faults and degrade device reliability), threading screw dislocations, and surface defect densities all exceed equivalent metrics in modern silicon substrates. These defects translate directly into device-level killer faults that reduce the fraction of dies on a wafer that pass test.

Mature SiC operators achieve device yields in the 70–85% range on 150mm wafers, with improvements expected (and being realized slowly) at 200mm. Silicon yields at comparable node maturity run 95%+ or higher. The yield gap means that SiC die cost per functional unit is disproportionately higher than the wafer cost difference alone would suggest, and yield improvement is a primary industry R&D focus. The substrate defect density is the primary lever — device process improvements have approached the yield ceiling set by substrate quality, and further yield gains require cleaner substrates.

The EV Traction Demand Driver

EV traction inverter adoption of SiC has been the dominant demand driver since Tesla integrated SiC MOSFETs in the Model 3 main traction inverter (original supply from STMicro). The structural case for SiC in EV traction is clear: efficiency gains (2–4 percentage points over silicon IGBT inverters) translate directly to range improvements or battery capacity reductions for equivalent range, both of which reduce total vehicle cost despite higher power semiconductor cost. 800V battery architectures (Porsche Taycan, Hyundai E-GMP, Kia EV6, Chinese BEV platforms, next-generation Tesla) specifically require SiC because silicon IGBTs cannot efficiently handle 800V operation. 400V architectures increasingly adopt SiC for efficiency gains where cost permits.

Tesla's next-generation "downsized SiC" approach — reducing total SiC semiconductor area per vehicle while maintaining performance — has been watched closely by the industry. If Tesla's approach demonstrates that vehicles can use less SiC per unit while maintaining performance benefits, it creates downward pressure on per-vehicle SiC content even as unit volumes scale. The industry-level SiC demand forecast depends on both vehicle production volumes and per-vehicle content; both variables are in flux.

See ElectronsX for the broader electric vehicle traction inverter coverage, BESS architectures for battery energy storage SiC applications, and solar inverter coverage for renewable energy SiC demand.

Strategic Framing: Nine Markets, One Wafer Funnel

The most important strategic observation about SiC is that the device demand from nine distinct end markets (EV traction, BESS, solar inverters, industrial drives, HVDC, datacenter PSUs, charging infrastructure, rail traction, aerospace specialty) converges on a single physics-limited substrate supply. Expansion of any one market increases pressure on substrate availability for all nine. EV growth tightens SiC supply for datacenter PSU manufacturers. BESS scale-up tightens SiC availability for industrial drive customers. The markets do not have independent substrate supply chains; they share one.

This convergence is different from the silicon semiconductor market, where leading-edge logic, memory, analog, and specialty categories operate on essentially independent wafer supply chains. SiC's one-funnel structure means that cross-market demand surges produce compounding tightness that silicon markets do not experience. The multi-year substrate capacity investment lead time combined with the one-funnel demand structure makes SiC capacity planning one of the most challenging in semiconductors — forecast accuracy across nine demand segments is required to avoid either oversupply or acute shortage.

Fabs in This Archetype

The specific SiC fab inventory is maintained in the Fab Facilities dataset. Notable SiC fabs include Wolfspeed Durham NC, Wolfspeed Mohawk Valley NY, Wolfspeed Siler City NC, STMicro Catania, Infineon Villach, Infineon Kulim, onsemi Bucheon, onsemi Hudson NH, Rohm Chikugo, Bosch Reutlingen, Mitsubishi Electric Fukuoka, Coherent Easton PA, SICC Jinan, and the Chinese scaling operators. See Fab Facilities for the full inventory.

Related Coverage

Parent: Wafer Fabs

Peer archetype pages: Leading-Edge Logic · Mature Logic · DRAM · 3D NAND · GaN Power & RF · Analog & Mixed-Signal · CMOS Image Sensor · MEMS · III-V Compound Semiconductor · Silicon Photonics · Rad-Hard & Rad-Tolerant

Cross-pillar dependencies: Power Semiconductors · SiC & GaN (EX–SX Interface)

Cross-network EV supply chain: ElectronsX · EV Traction Inverters · BESS · Solar Inverters

Supply chain framing: Bottleneck Atlas · U.S. Reshoring