A bet made five years before it mattered
onsemi’s current silicon-carbide position traces back to a specific 2021 decision: acquiring GT Advanced Technologies (GTAT), giving the company in-house capability to grow its own SiC crystals rather than depending entirely on external substrate suppliers — the same substrate-supply bottleneck this cohort’s companion GaN-and-SiC briefing documents constraining Wolfspeed and other competitors in 2026 [1]. That single acquisition is the structural reason onsemi enters this briefing’s 2026 window with a materially different cost and supply position than peers still dependent on third-party boule and wafer supply.
The product-level payoff: EliteSiC and the M3e generation
onsemi’s EliteSiC M3e MOSFET generation, launched in late 2024, reduced conduction losses by 30%, and has become an industry-standard component for 800V EV traction inverters [1]. That efficiency improvement translates directly into real vehicle- level outcomes — less energy lost as heat during power conversion means more of a battery’s stored energy actually reaches the wheels, a concrete engineering payoff from onsemi’s underlying materials investment.
The 2026 manufacturing milestone: 150mm to 200mm wafers
In 2026, onsemi ramped 200mm SiC production at its Bucheon, South Korea facility, transitioning from the previous 150mm wafer standard — a change that increases the number of usable chips produced per wafer by roughly 80%, providing a substantial per-unit cost advantage [1]. This is the same kind of wafer-scale economics this cohort’s equipment-and-materials track documents driving cost curves across the broader semiconductor industry, applied specifically to the SiC materials system.
The customer-relationship evidence: NIO and 900V platforms
onsemi expanded its collaboration with Chinese EV maker NIO in 2026 specifically to accelerate the transition to next-generation 900V EV platforms [3] — a voltage architecture increase beyond the 800V standard EliteSiC M3e already serves, indicating onsemi’s roadmap is already being designed into automakers’ next platform generation rather than only serving current-generation vehicle programs.
The honest second half: a genuine 2026 slowdown
None of the above should be read as an unqualified success story. Reporting from mid-2026 describes onsemi facing a continued slowdown in SiC demand growth, alongside intensifying competition in a separate product line, its CMOS image sensor (CIS) business, from Sony [2]. This is directly consistent with this cohort’s companion GaN-and-SiC briefing, which documents EV adoption proceeding more slowly than earlier industry projections assumed — a demand-side headwind that no amount of onsemi’s own manufacturing-cost improvement can offset on its own.
Why “right bet, difficult cycle” is the fair characterization
onsemi’s 2021 GTAT acquisition and its 2026 wafer-transition execution both represent genuinely sound long-term positioning — the company controls more of its own supply chain and has real cost advantages over less vertically integrated SiC competitors. What it cannot control is the near-term pace of EV adoption, which this cohort’s power-and-automotive coverage documents running behind earlier industry expectations across multiple companies simultaneously, not onsemi specifically. The fair 2026 verdict on onsemi’s SiC strategy is neither “failed bet” nor “vindicated bet” but something more precise: a structurally sound position facing a genuinely difficult demand cycle it did not create and cannot single-handedly accelerate [4].
Why vertical integration matters more in a slow cycle, not less
A counter-intuitive point worth making explicitly: onsemi’s in-house crystal-growth capability arguably matters more during a demand slowdown than it would during a boom. In a supply-constrained upswing, nearly any SiC supplier can sell everything it manufactures regardless of cost structure. In a slower-growth environment, cost position becomes the decisive competitive variable, since customers have more leverage to choose the cheapest qualified supplier rather than accepting whatever capacity happens to be available. onsemi’s 80% chips-per-wafer improvement from the 200mm transition is precisely the kind of cost advantage that compounds in relative importance during exactly the demand environment the company is currently navigating.
The NIO relationship as a leading indicator worth tracking
Because the NIO 900V collaboration targets a next-generation platform rather than a current one, its commercial results will not show up in onsemi’s revenue for some time. Readers tracking this story going forward should treat design-win announcements like this one as leading indicators of onsemi’s medium-term trajectory, arriving well before they show up in quarterly financial results — the same lag this cohort’s other automotive-silicon briefings document between a platform design win and its eventual revenue recognition once vehicle production actually ramps.
