01. Why wide bandgap, why now
Silicon has governed power electronics for five decades because it is cheap, well-understood, and supported by a mature fabrication ecosystem. A silicon IGBT module driving an industrial motor at 650V works reliably and costs a fraction of what a SiC alternative costs. For applications where efficiency losses are acceptable, switching frequencies are low, and operating temperatures stay below 150°C, silicon remains the rational choice and will continue to be through 2030 and beyond. The problem is that EV powertrain requirements have moved outside the envelope where silicon performs well. A 400V or 800V traction inverter switches at 10 to 20 kHz, must operate at junction temperatures of 175°C or above, and needs to achieve system efficiency above 97% to meet regulatory and range targets that OEMs have committed to. At those parameters, silicon IGBTs dissipate too much heat and require thermal management systems heavy enough to cancel out part of the efficiency gain. Wide bandgap semiconductors — specifically silicon carbide and gallium nitride — have bandgap energies of 3.26 eV and 3.4 eV respectively, versus silicon's 1.12 eV, which is what allows them to operate at higher voltages, temperatures, and switching frequencies while producing less heat per watt of power handled. The commercial question is not whether wide bandgap wins — it will — but which material wins which application socket, on what timeline, and constrained by what supply chain realities. The performance benchmarks mostly favour SiC for high-voltage, high-current applications and GaN for lower-voltage, higher-frequency applications. But performance benchmarks are not what determines market outcomes. Supply chain economics, OEM qualification timelines, and substrate availability are.
"The SiC MOSFET will capture the 400V and 800V traction inverter socket. GaN will own onboard chargers and DC-DC converters. The fight is not SiC vs GaN — it is SiC vs silicon in traction, and GaN vs silicon in auxiliary power."
02. The EV traction inverter: SiC's home territory
The traction inverter converts DC battery power to the three-phase AC that drives the electric motor. In a 400V platform, the inverter handles 400 to 450V DC input at peak currents of 300 to 600 amps. In an 800V platform — adopted by Porsche Taycan, Hyundai IONIQ 6, Kia EV6, and increasingly by all premium OEMs — the input voltage is 700 to 900V DC. Silicon IGBTs at 800V require parallelling multiple devices to manage current density and thermal dissipation, which adds cost and system complexity. A SiC MOSFET rated at 1200V handles the same current in a smaller die with 65% lower on-resistance, producing less heat and enabling a smaller, lighter inverter module. Device value per vehicle is the metric that has made SiC commercially interesting to Infineon, STMicroelectronics, Wolfspeed, and ON Semiconductor. A conventional IGBT-based 400V inverter uses approximately USD 40 to USD 60 of silicon device content. A SiC MOSFET-based 800V inverter uses USD 200 to USD 400 of SiC device content — a four to six times increase in value per vehicle at a margin structure that automotive-grade SiC currently commands. By 2024, Tesla, BYD, Volkswagen Group, Hyundai-Kia, and General Motors had all transitioned primary traction inverter production to SiC MOSFETs. Toyota remains the largest holdout at scale, still qualifying SiC for its next-generation BEV platform after achieving volume production with silicon IGBTs in its bZ4X. The OEM qualification cycle for SiC is 18 to 36 months from device selection to production sign-off — which means every OEM now in qualification will be committed to SiC through at least 2028, and likely 2030.
03. Where GaN is actually winning
GaN does not compete with SiC in the traction inverter. The physics make it structurally unsuitable: lateral GaN HEMT devices have breakdown voltages below 700V in commercially viable die sizes, and vertical GaN devices capable of 1200V operation are still in development at Transphorm, GaN Systems (now acquired by Infineon), and Navitas. The 800V traction inverter socket requires 1200V rated devices with high pulse current capability, which GaN cannot consistently deliver at production yields today. What GaN is winning is the onboard charger and DC-DC converter market — the auxiliary power sockets in an EV that were previously contested between silicon and SiC but where GaN's superior switching frequency performance (1 to 10 MHz versus SiC's 100 to 500 kHz practical limit) creates a system size advantage. A 22kW onboard charger using GaN devices operates at higher switching frequency, which reduces the size of the passive components — inductors and transformers — needed in the power stage. At 6.6kW and 11kW charger power levels, GaN can achieve a charger module half the physical size of an equivalent SiC design at similar or lower system cost. The OBC transition to GaN is already underway. Navitas Semiconductor disclosed GaN IC design wins in onboard chargers at two undisclosed Tier 1 automotive suppliers in Q3 2024. STMicroelectronics launched its GaN HEMT series specifically targeted at 6.6kW and 22kW OBC applications. The commercial signal is clear: GaN is the preferred device for auxiliary power applications below 700V where high switching frequency reduces system cost, and SiC is the preferred device for traction where voltage and current requirements exceed GaN's current capability.
| Application | Voltage Range | 2024 Leader | 2024 Share | 2030E Leader | 2030E Share |
|---|---|---|---|---|---|
| Traction Inverter | 400–900V | SiC | SiC 82% | SiC | SiC 91% |
| Onboard Charger (OBC) | 200–500V | SiC | SiC 54% | GaN | GaN 68% |
| DC-DC Converter | 48–400V | Silicon | Si 71% | GaN | GaN 58% |
| EV Fast Charger PSU | 200–1000V | SiC | SiC 61% | Split | SiC 52%, GaN 38% |
04. The substrate supply chain is the real constraint
Every discussion of SiC market growth eventually arrives at the same bottleneck: SiC boule growth is slow, yield-limited, and capital-intensive in a way that silicon wafer production is not. A 150mm silicon wafer takes hours to grow by Czochralski process. A 150mm SiC boule takes three to five days by physical vapour transport, and producing a defect density low enough for power device qualification requires additional process control that most producers have not yet achieved at scale on 200mm boules. The 150mm to 200mm transition is the critical inflection point. 200mm SiC wafers offer approximately 77% more die area per wafer than 150mm — which translates directly to lower die cost per device at the same process yield. Wolfspeed's Mohawk Valley facility is the first fab to produce power devices on 200mm SiC wafers at commercial scale, and it commenced volume shipments in March 2025. Infineon's Villach 200mm SiC line and Onsemi's Hudson, Ohio facility are both targeting 2026 production readiness. ROHM and STMicroelectronics are on 2027 timelines for 200mm production at volume. GaN substrate supply is a separate and equally significant constraint. GaN-on-SiC, used in RF and microwave applications, relies on SiC substrate supply — which means RF GaN demand competes with power GaN and power SiC for the same upstream substrate. GaN-on-silicon is a viable alternative for power electronics applications below 650V, enabling GaN device fabrication on standard 200mm silicon wafers at silicon wafer costs. This is why companies like Navitas, GaN Systems, and EPC have built their power GaN product lines on GaN-on-silicon rather than GaN-on-GaN: it decouples their supply chain from the SiC substrate bottleneck.
"The constraint on SiC market growth through 2028 is not device design or fab capacity — it is boule growth yield and 200mm substrate availability. Capital expenditure cannot fix a three-to-five-day crystal growth cycle."
05. OEM qualification pipelines — who is where
OEM qualification status determines demand visibility for SiC suppliers and is the most commercially relevant signal for procurement and strategy teams. A completed qualification means committed volume; an in-progress qualification means a forecast demand signal with 18 to 36 months to production conversion.
06. 2030 forecast: a divided market
The Nodvolt base case for the wide bandgap power semiconductor market by 2030 reflects a structurally divided outcome that neither the SiC-only nor the GaN-everywhere narrative captures correctly. SiC will hold 85 to 90% of traction inverter device revenue through 2030, with no credible GaN challenger at the 800V traction level in commercial production within that timeframe. Traction inverter SiC revenue reaches USD 8.4 Billion by 2030 from USD 2.8 Billion in 2024, driven by 800V platform proliferation, rising device value per vehicle, and expanding OEM adoption beyond the current lead adopters. GaN will hold 60 to 70% of onboard charger device revenue by 2030, transitioning from SiC's current position as the dominant OBC device. GaN-on-silicon economics enable below-USD-5 device cost at 650V class ratings, which makes GaN the rational choice for every new OBC design at 6.6kW and above that begins qualification in 2025 or later. GaN OBC revenue reaches USD 2.1 Billion by 2030 from USD 380 Million in 2024. The substrate constraint is the downside scenario risk. If Wolfspeed's 200mm SiC ramp underdelivers — a real risk given their 2024 financial restructuring and customer allocation commitments — SiC device pricing will remain elevated, making some 400V OEM applications defer to silicon IGBTs for an additional design cycle. The upside scenario is 200mm SiC yield improvement accelerating faster than Nodvolt's base case, driven by improved boule growth process control, which would pull forward the die cost crossover with 150mm and accelerate SiC adoption in mid-tier EV platforms currently priced out of SiC economics.
07. What this means for procurement and strategy
For semiconductor company strategy teams, the application segmentation of wide bandgap is the primary planning variable. A company pursuing traction inverter sockets needs SiC device capability, 1200V MOSFET ratings, automotive-grade qualification, and a credible 200mm substrate sourcing plan. A company pursuing auxiliary power sockets should be evaluating GaN-on-silicon as the cost-competitive path, accepting that GaN wafer sourcing on silicon substrates is structurally lower risk than SiC substrate dependency. For procurement teams at automotive Tier 1 suppliers, the substrate supply chain is where strategic sourcing attention should be focused, not device procurement. The companies that secure long-term 200mm SiC substrate agreements with Wolfspeed, Coherent, and SiCrystal before 2026 will have a cost structure advantage in traction inverter production that latecomers cannot close through device procurement alone. The qualification lead time means the sourcing decision being made today determines 2028 and 2029 production economics. For sales heads at wide bandgap device companies, the TAM conversation needs to shift from market-level revenue to application-level device content. The total wide bandgap power semiconductor market growing at 22% CAGR masks the structural divergence between a traction SiC segment growing at 18% and an auxiliary GaN segment growing at 38%. The accounts worth pursuing, and the product specifications worth qualifying, differ entirely between those two growth trajectories.