Silicon carbide is already in your electric vehicle. The question nobody can answer yet is what happens when it fails.
A new paper published this month in IEEE Transactions on Electron Devices — authored by researchers from the University of Bremen, TU Chemnitz, BMW, Robert Bosch, Infineon, Semikron Danfoss, and FH Dortmund — documents what the automotive semiconductor supply chain has been working around: the IEC qualification standards that govern component reliability were written for silicon. Silicon carbide and gallium nitride have failure modes those standards do not cover.
SiC components already meet existing automotive qualification standards, the paper states. They are in commercial EVs. The problem is not current performance. The problem is what the standards do not test for.
Bipolar degradation and gate switching instability are WBG-specific effects that do not occur in silicon devices. The degradation mechanisms over a vehicle's service life can differ from silicon in ways the existing standards were not designed to detect. Production maturity of WBG semiconductors has not yet reached the level of silicon technology, the authors note. If standard screening tests cannot adequately cover WBG-specific defect modes, the fallback is burn-in testing — which is more expensive and slower.
The implications are not theoretical. Automotive qualification requires IEC compliance. No component ships to an OEM without it. But the IEC standards as currently written cannot fully characterize WBG behavior in the field. The paper's authors — which include engineers from every major automotive power semiconductor company — are not saying SiC is unsafe. They are saying the standards lag behind the technology, and the gap needs to close before WBG deployment in automotive reaches the maturity level silicon has.
The authors propose screening tests specific to WBG failure modes, and burn-in testing as a fallback where screening is not viable. Both approaches add cost and time to component qualification. For an industry already managing SiC supply constraints and pricing pressure, that is a non-trivial engineering and commercial problem.
There is also the data question. SiC has been in commercial EVs for several years, but the fleet is still young. Comprehensive field failure data — the kind that would validate whether current qualification standards are adequate or whether the WBG-specific failure modes are manifesting in ways the standards cannot detect — does not yet exist at scale. The authors note this explicitly: the service life degradation data needed to validate the standards' adequacy is still accumulating.
This is the aerospace analogy applied to automotive: you qualify to the standard, you fly the mission, you see what happens. SiC passed qualification. What happens in the field over 15 years of thermal cycling, vibration, and humidity is still being learned.
The commercial stakes are high. SiC inverters are central to EV efficiency — they convert battery DC to motor AC with lower losses than silicon IGBTs, which extends range. The major EV OEMs have built supply relationships with Infineon, onsemi, and others specifically to secure SiC capacity. If a field failure mode emerges that current standards do not catch, and if it manifests in a way that triggers recalls or redesigns, the supply chain implications are significant.
The paper is technical. The supply chain angle is the story: standards lag technology, and when the technology is already deployed in millions of vehicles, the catch-up has real consequences.
