Infineon Makes Chips on 300 mm Silicon Carbide as Rivals Pull Back

Infineon said on November 9 that it has turned 300 millimeter silicon carbide wafers into working power devices. The work was done at its fab in Villach, Austria. Each wafer yields 2.3 times as many chips as a 200 millimeter wafer. That could bring down the cost of power chips for electric vehicles and AI data centers.

• Infineon announced working power devices on 300 millimeter silicon carbide wafers on November 9, 2026.

• The wafers were processed in Villach, Austria, using existing silicon manufacturing equipment.

• A 300 millimeter wafer yields 2.3 times as many chips as a 200 millimeter wafer.

• Infineon gave no production timeline and will keep its 200 millimeter roadmap.

• onsemi cancelled a Czech silicon carbide expansion worth up to $2 billion days earlier.

What Infineon actually did in Villach

Infineon has processed 300 millimeter silicon carbide wafers into working power devices inside a full scale fab. The company laid out the result in a November 9 announcement. It claims to be the first chipmaker to do so in a full production environment.

What is significant is the tooling, because Infineon processed the wafers on tools built for 300 millimeter silicon, which already fill its lines at Villach. A bigger wafer might therefore not need a new fab built around it. Building that fab is usually the biggest cost of moving to a new wafer size.

Infineon also reports progress further upstream. At its Dresden site, the company has shown that its Cold Split technology works on 300 millimeter material. Cold Split separates wafers from a silicon carbide crystal without wasting as much material or taking as long as wire sawing.

Chief executive Jochen Hanebeck framed the result as a matter of scale. In the company's announcement, he said:

Infineon is showing the first processed wafer this week at electronica in Munich, which runs November 10 to 13. The company also plans a 300 millimeter gallium nitride facility at the Villach site. That project is worth about 1.1 billion euros.

Why does wafer size matter for silicon carbide?

Wafer size is one of the most important factors in the cost of a chip. A fab processes one wafer at a time, so the more chips it can get out of a wafer, the lower the cost of each chip. Infineon puts the gain from 200 to 300 millimeters at 2.3 times.

Silicon carbide has lagged in adopting larger wafers. The material is nearly as hard as diamond, grows slowly at extreme temperatures and is wasteful when sliced. Most of the industry is still moving from 150 to 200 millimeter wafers, more than two decades after silicon reached 300.

The reward is efficiency, because silicon carbide switches high voltages with less energy lost as heat than silicon. That makes it the choice for traction inverters, on board chargers, fast chargers and the power supplies of AI data centers.

A breakthrough in a year of retreat

The announcement came in the same week that a competitor backed away. onsemi said it is not moving forward with a silicon carbide expansion in Roznov pod Radhostem, in the Czech Republic. The project was announced in June 2024, with a value of up to $2 billion.

onsemi said it had reassessed global market conditions for silicon carbide as well as the economics of the project. Its existing Czech production, including silicon carbide, will continue without interruption. Reports on the decision point to pressure from lower priced wafers from China.

Infineon is not the only company at 300 millimeters. On January 13, Wolfspeed said it had produced a single crystal 300 millimeter silicon carbide wafer. Its chief technology officer, Elif Balkas, said at the time:

A wafer is not the same as a finished device, however. Wolfspeed has shown the material, and Infineon now says it has built working devices on it. Neither company has shown that it can produce them in volume.

The questions Infineon left open

Infineon did not give any detail on costs, nor a date for production. The company said it will work with customers and partners to plan the technology and the timing of the ramp. Meanwhile, it will continue with its 200 millimeter roadmap.

Supply is the first open question, because a 300 millimeter line needs a steady flow of large crystals with few defects. Those are still hard to obtain and costly to produce. Yield is the second, because a single crystal defect can scrap a large power chip.

Demand is the third, since sales of electric vehicles have grown more slowly than chipmakers expected when they announced their new fabs. Hanebeck's own summary of the news was confident:

The market will judge that claim on price and delivery, not on the size of the wafer. Competitors have not stood still, and on November 10 onsemi launched a new trench based silicon carbide platform for traction inverters.

The bottleneck moves to the package

A less expensive silicon carbide chip does not necessarily make for a less expensive inverter. The dies are small, run hot and switch fast, so heat and stray inductance in the package limit what they can do. A big part of a power module's cost and reliability now lies outside the die, in the layers that hold it, cool it and connect it.

This is where materials that most drivers have never heard of come into play. Ceramic substrates carry heat away from the die while insulating it from thousands of volts. Silver sintering and copper clips replace the solder and bond wires that fail under repeated heating and cooling.

Sentec E&E, a Taiwanese manufacturer that started in 1968, works on that side of the problem. The company makes silicon nitride and aluminum nitride ceramic substrates, and it assembles power modules with silicon, gallium nitride and silicon carbide devices. Its processes include silver sintering, copper clip bonding and molding, and it builds integrated motor and inverter units too.

Sentec is one of eight companies in the Taiwan Mobility Pavilion, which Sparknify previewed last week. The company will be at Booth #4577 in LVCC West Hall during CES 2027, January 6 to 9. Power electronics teams can book a meeting with Sentec before the show.

Will bigger wafers make silicon carbide cheap?

The near term is likely to bring a slow change rather than an abrupt one. Infineon has shown that 300 millimeter silicon carbide can be made on standard tools. Scaling up will require a good crystal supply, solid yields and customers willing to requalify parts.

Automotive customers tend to be slow and careful about any change, and they have reasons. A traction inverter chip must be qualified for years of heat, vibration and voltage stress. Carmakers will not rush to switch wafer sizes on a supplier's schedule, so 200 millimeter production will likely handle most of the volume for some years.

There is a real debate in the industry over whether the capacity is even needed. Some believe demand from electric vehicles, energy storage and AI data centers will outstrip what existing fabs can supply. Others see cancelled projects and dropping wafer prices as a sign of too much capacity. On that reading, a wafer that yields 2.3 times as many chips arrives just when there are too many chips.

Both views can hold at different times. Cheaper chips could expand the market, as was the case for silicon, and get silicon carbide into cheaper cars and more power supplies. They could also squeeze every company that invested in capacity on the assumption that old prices would last. Governments that subsidized fab building in Europe and the United States have an interest in which outcome comes first.

These concerns are not just for chipmakers. If the price of the die goes down, the substrate, sintering and assembly will make up a higher share of what a power module costs. That will put pressure on those suppliers to scale up and lower prices as well.

The opportunity for founders and investors is around the chip rather than in it. Bigger wafers favor the biggest chip manufacturers, but they also drive demand for crystal growth tools, defect inspection, substrates and packaging. A startup that can cut one cost out of this chain will find customers on three continents.

Four things will show which way this goes.

1. Infineon naming a start date or a first customer for 300 millimeter silicon carbide.

2. Prices for 200 millimeter silicon carbide wafers through 2027.

3. Further cancellations or delays of silicon carbide factories in Europe and the United States.

4. Carmakers qualifying 300 millimeter silicon carbide devices for traction inverters.

The wafer got bigger this week, and the harder work of making it pay has only started.

Subscribe to Sparknify for tech news and the people behind it.

Questions & Answers

Q1: I design traction inverters at an EV startup. Should I wait for 300 millimeter silicon carbide chips, or put my effort into module packaging now?

Don't wait. Infineon did not provide a production date, and 200 millimeter chips will carry most of the volume for years to come. Decisions about substrates and silver sintering define heat and reliability today. Engineers can book a meeting with Sentec at sparknify.com/sentec-en to talk about substrates and module assembly at CES 2027, Booth #4577, January 6 to 9.

Q2: Will AI data centers end up competing with electric vehicles for silicon carbide chips the way they already compete for memory?

They already draw on the same supply. Infineon lists AI data center power supplies along with traction inverters as uses for its silicon carbide. Memory shows how that contest can play out, with the squeeze on cockpit memory raising costs for carmakers this year. Larger wafers could ease the pressure if they reach volume production.

Q3: Which other Taiwanese suppliers will be at CES 2027 with Sentec, and what do they make for carmakers?

Sentec is sharing Booth #4577 in the West Hall of the LVCC with seven other companies. Between them, they span vision AI, memory, over the air software, edge AI, 3D sensing and power electronics. The Taiwan Mobility Pavilion page shows all eight exhibitors and links to each one's meeting page.