For a GaN power stage in Austria, the 630V C0G MLCC brief is wide in the catalogue and thin on the shelf: we counted 1,450 Walsin Technology (Walsin) NP0 part numbers rated 630 V or more, 60 of them in AEC-Q200 series, and 38 on our Hong Kong shelf as at 23 September 2026.
Venatech Limited, an authorised Walsin distributor in Hong Kong, wrote this from Walsin’s MLCC export (00_ALL_MLCC_parts_combined.xlsx, 24,026 rows, 23,601 distinct part numbers re-counted by us on 11 November 2026) and our stock list as at 23 September 2026. Walsin writes the C0G class as NPO; we use C0G and NP0 for the same class here. Every part number is real. The news statements name their source and the day we fetched them, 11 November 2026. This is engineering perspective, not legal or compliance advice.
What did Infineon and the European Commission announce for Villach?
Two linked announcements, one day apart. According to electrive on 28 October 2026, Infineon will invest about 1.1 billion euros by 2034 at Villach to set up gallium nitride power semiconductor production on 300 mm wafers, with production scheduled to start at the end of 2027.
Euronews reported on 28 October 2026 that the European Commission approved 319 million euros of Austrian state aid for the project under the European Chips Act. The electrive report lists AI data centres, electric vehicles, renewable energy, robotics and consumer electronics as target applications, and puts Infineon Austria’s workforce at about 5,800, roughly 4,800 of them in Villach. Those are the reports’ figures, not ours. We have no relationship with Infineon and say nothing about what it or its customers buy.
Why does a GaN stage push the capacitor brief towards 630V C0G MLCC?
Because a GaN switch on a 400 V bus switches fast enough that the snubber, resonant and bootstrap capacitors see sharp edges and overshoot. On a 400 V bus, a 630 V rating leaves about 1.6 times headroom over the nominal rail, which is our assumed design rule for this post and not a standard; the overshoot on your own layout decides the real margin.
Engineers usually reach for a class I dielectric in these positions because its capacitance barely moves with DC bias or temperature, while class II X7R parts lose capacitance under bias. Our post on MLCC DC bias and voltage derating covers that loss, and NP0 vs X7R vs X5R covers the dielectric codes. Check the dielectric behaviour in Walsin’s Middle and High Voltage datasheet, WTC_MLCC_Middle_and_High_Voltage.pdf, before you fix a value.
How many Walsin C0G MLCC are rated 630 V or more?
Plenty in the catalogue, few on the shelf. We parsed every row of the MLCC export and counted 1,450 distinct NP0 part numbers rated 630 Vdc or more, after removing the S2 safety rows, whose voltage field holds an impulse rating. 889 reach 1,000 V. 60 sit in the AEC-Q200 MT or ST series. 38 are on our shelf.
| Cut of the 630 V+ NP0 range | Distinct part numbers | On shelf 23 Sep 2026 |
|---|---|---|
| All NP0, 630 Vdc or more (S2 excluded) | 1,450 | 38 |
| Of which 1,000 Vdc or more | 889 | 31 |
| Of which AEC-Q200 MT or ST series | 60 | 0 |
| 100 pF to 1 nF window, 630 Vdc or more | 463 | 22 |
| 100 pF to 1 nF window, AEC-Q200 MT or ST | 21 | 0 |
By size the 1,450 split as 494 in 1206, 259 in 0805, 193 in 1812, 179 in 1210, 165 in 1808, 56 in 1825, 52 in 2220, 37 in 2225 and 15 in 1111. For comparison we counted 1,203 X7R parts at 630 V or more, so the C0G choice is not the narrower one in this range. Capacitance is the limit: the largest NP0 value at 630 V is 1.5 nF in 0805 and 10 nF in 1206, and at 1,000 V the 0805 ceiling drops to 390 pF. We have not seen these counts published anywhere else.
Which Walsin parts answer a 100 pF to 1 nF, 630 V C0G brief, and what differs?
Several do, and the stocked ones sit in the Middle and High Voltage series. For a brief of C0G, 100 pF to 1 nF, at least 630 V, the export returns 463 parts; 22 are stocked, including 1812N681J631 at 680 pF and 630 V and 1210N471J102 at 470 pF and 1,000 V, while all 21 AEC-Q200 parts in the window are to-order.
| Walsin part | Series / size | Value / voltage | On shelf 23 Sep 2026 | Match / mismatch with the C0G, 100 pF to 1 nF, 630 V+ brief | Your incumbent |
|---|---|---|---|---|---|
1812N681J631 |
Middle and High Voltage, 1812 | 680 pF, 630 Vdc | Yes | Match on dielectric, value window and voltage; mismatch if your footprint is 1206 | |
1210N471J102 |
Middle and High Voltage, 1210 | 470 pF, 1,000 Vdc | Yes | Match with extra voltage headroom | |
0805N181J631 |
Middle and High Voltage, 0805 | 180 pF, 630 Vdc | Yes | Match on all three; smallest stocked footprint here | |
1808N471J102 |
Middle and High Voltage, 1808 | 470 pF, 1,000 Vdc | Yes | Match; 1808 is a less common land pattern | |
MT31N471J631 |
MT, AEC-Q200, 1206 | 470 pF, 630 Vdc | No | Match on all three and on AEC-Q200 series; to-order only | |
SH31N471J631 |
SH soft termination, 1206 | 470 pF, 630 Vdc | No | Match; soft termination adds board-flex margin |
A match here is a match on export fields only. Confirm against both datasheets and qualify in your own circuit, including ripple heating at your switching frequency, which the export does not carry.
Which voltage fields in the export should an Austrian engineer double-check?
Three high-voltage C0G rows whose part-number voltage code disagrees with the Rated Voltage column. We decoded the last three characters of all 23,599 MLCC part numbers that carry a voltage code and found 17 that disagree with the column, including 2225N103J101 (code 100 V, column 1,000 V), and 1206N182G631 and 1206N182K631 (code 630 V, column 1,000 V).
The other 14 sit in RF, High Q, MT and ST series and lower voltages. We cannot tell from the export which field is right, so we treat the lower figure as the working assumption until the datasheet settles it. For a GaN snubber, reading 1,000 V where the part is 630 V, or 100 V where it is 1,000 V, changes the design outright.
How can Venatech help an Austrian GaN design team?
How Venatech helps starts with the free paste box, and the arithmetic above took one export and 23,601 part numbers. The cross-reference tool, VenaMatch, decodes 18 manufacturers’ part numbers across 800,000+ crossable parts, accepts up to 500 part numbers or an Excel or CSV file, filters by commercial or automotive grade, and states every difference from your original.
Upload a whole list at BOM upload for review and a request for quote. The ready-stock list is a Hong Kong snapshot dated 23 September 2026, never live-synced; the MLCC range lists the series. Walsin is the only line we are authorised for; Stackpole and Viking are lines we supply. We serve from Hong Kong in small lots, give no price, lead time or stock guarantee, hold no office or stock in Austria, and give no compliance advice: confirm with your compliance team.
The short version
GaN on a 400 V bus pulls the snubber and resonant positions towards C0G at 630 V or more. Walsin lists 1,450 such NP0 parts, 463 in the 100 pF to 1 nF window, but only 38 are on our shelf and none of the 60 AEC-Q200 ones. Check three rows whose voltage code and voltage column disagree.
Paste your snubber and resonant capacitor lines into the free cross reference, or send them to us, and we will return the Walsin candidates with every difference stated. Which matters more on your board: the 1206 footprint or the AEC-Q200 series?
Sources
Fetched 11 November 2026.
- electrive, “Infineon to invest 1.1 billion euros in Villach”, 28 October 2026: https://electrive.com/2026/10/28/infineon-to-invest-e1-1-billion-in-villach
- Euronews, “EU approves 319 million in state aid to Austrian first-of-a-kind next-generation semiconductor”, 28 October 2026: https://www.euronews.com/2026/10/28/eu-approves-319-million-in-state-aid-to-austrian-first-of-a-kind-next-generation-semicondu
- Walsin, Middle and High Voltage MLCC datasheet: https://www.passivecomponent.com/wp-content/uploads/datasheet/WTC_MLCC_Middle_and_High_Voltage.pdf
Related reading
- MLCC DC bias and voltage derating
- NP0 vs X7R vs X5R MLCC dielectric
- Soft termination MLCC: SH, SG, ST and VT
- Free tools: VenaMatch cross reference and the ready stock list
Last reviewed 11 Nov 2026.
Frequently asked questions
<!– wp:rank-math/faq-block {"questions":[{"id":"faq-question-01","title":"What did Infineon announce for Villach, Austria?","content":"According to electrive on 28 October 2026, about 1.1 billion euros by 2034 to set up GaN power semiconductor production on 300 mm wafers in Villach, with production planned from the end of 2027. Euronews reported 319 million euros of approved state aid the same day.","visible":true},{"id":"faq-question-02","title":"How many Walsin 630V C0G MLCC part numbers exist?","content":"We counted 1,450 distinct NP0 part numbers rated 630 Vdc or more in Walsin's MLCC export, excluding the S2 safety rows. 889 reach 1,000 V, 60 are in AEC-Q200 MT or ST series, and 38 are on our Hong Kong shelf as at 23 September 2026.","visible":true},{"id":"faq-question-03","title":"Are any AEC-Q200 C0G MLCC at 630 V on the shelf?","content":"No. All 60 AEC-Q200 MT and ST NP0 parts at 630 V or more are to-order, including the 21 in the 100 pF to 1 nF window such asMT31N471J631. Walsin’s datasheets say compliant; request the qualification report for the exact part number.”,”visible”:true},{“id”:”faq-question-04″,”title”:”Why use C0G rather than X7R in a GaN snubber?”,”content”:”Class I C0G capacitance barely moves with DC bias or temperature, while X7R loses capacitance under bias. In a snubber or resonant position that loss shifts the tuned value. Check the behaviour in Walsin’s datasheet and qualify the part in your own circuit.”,”visible”:true},{“id”:”faq-question-05″,”title”:”Which Walsin MLCC rows have a voltage code that disagrees with the voltage column?”,”content”:”We found 17 of 23,599. The high-voltage C0G ones are 2225N103J101, 1206N182G631 and 1206N182K631. We treat the lower figure as the working assumption until the datasheet settles it.”,”visible”:true},{“id”:”faq-question-06″,”title”:”Do you hold stock or an office in Austria, or give compliance advice?”,”content”:”No. Our locations are Hong Kong, Shenzhen, Singapore and Jiangmen, with Chengdu in preparation. We give no legal, customs or compliance advice and no price, lead time or stock guarantee. Confirm with your compliance team.”,”visible”:true}],”listStyle”:””,”titleWrapper”:”h3″,”titleCssClasses”:””,”contentCssClasses”:””,”sizeSlug”:””} –>
What did Infineon announce for Villach, Austria?
According to electrive on 28 October 2026, about 1.1 billion euros by 2034 to set up GaN power semiconductor production on 300 mm wafers in Villach, with production planned from the end of 2027. Euronews reported 319 million euros of approved state aid the same day.
How many Walsin 630V C0G MLCC part numbers exist?
We counted 1,450 distinct NP0 part numbers rated 630 Vdc or more in Walsin’s MLCC export, excluding the S2 safety rows. 889 reach 1,000 V, 60 are in AEC-Q200 MT or ST series, and 38 are on our Hong Kong shelf as at 23 September 2026.
Are any AEC-Q200 C0G MLCC at 630 V on the shelf?
No. All 60 AEC-Q200 MT and ST NP0 parts at 630 V or more are to-order, including the 21 in the 100 pF to 1 nF window such as MT31N471J631. Walsin’s datasheets say compliant; request the qualification report for the exact part number.
Why use C0G rather than X7R in a GaN snubber?
Class I C0G capacitance barely moves with DC bias or temperature, while X7R loses capacitance under bias. In a snubber or resonant position that loss shifts the tuned value. Check the behaviour in Walsin’s datasheet and qualify the part in your own circuit.
Which Walsin MLCC rows have a voltage code that disagrees with the voltage column?
We found 17 of 23,599. The high-voltage C0G ones are 2225N103J101, 1206N182G631 and 1206N182K631. We treat the lower figure as the working assumption until the datasheet settles it.
Do you hold stock or an office in Austria, or give compliance advice?
No. Our locations are Hong Kong, Shenzhen, Singapore and Jiangmen, with Chengdu in preparation. We give no legal, customs or compliance advice and no price, lead time or stock guarantee. Confirm with your compliance team.