Intel surpasses one million High-NA EUV wafers processed, outpaces the rest of the industry combined — company also trailblazing giant 6×12 photomasks to speed production and lower costs

ASML EUV machine
(Image credit: ASML)

Intel announced on Monday that it had processed more than one million 300-mm wafers using its High-NA EUV scanners, less than two and a half years after its first tool was assembled. For now, the company intends to use industry-standard 6-inch photomasks, which can expose 26×16.5 mm half-fields and therefore require field stitching for larger chips. However, Intel is also working on larger 6×12-inch photomasks that would enable High-NA EUV scanners to expose full 26×33 mm fields without stitching.

One million High-NA wafers

Latest Videos FromTom's Hardware
Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

  • mac2213
    The reporting on the technical details in this article is outstanding, but it fails to recognize the confusing information that both Intel and ASML had pushed on the market regarding HNA tool deliveries starting in the spring of 2024. Even the recent comments by ASML at Semicon Taiwan 26 were intentionally misleading, but a bit closer to the truth where they talked about 4 HNA customers and 13 total HNA tools delivered, but failed to include IMEC and NY Crest Dev efforts in either count. It would be nice if Tom's could lead the reporting on the tool count and timing on delivery in their reporting. Restating the belief that Intel only has 3 or 4 tools out of 13 total (or 16 including IMEC/NYDEV) doing this kind of volume is confusing everyone further.
    Reply
  • moreslawplease
    mac2213 said:
    The reporting on the technical details in this article is outstanding, but it fails to recognize the confusing information that both Intel and ASML had pushed on the market regarding HNA tool deliveries starting in the spring of 2024. Even the recent comments by ASML at Semicon Taiwan 26 were intentionally misleading, but a bit closer to the truth where they talked about 4 HNA customers and 13 total HNA tools delivered, but failed to include IMEC and NY Crest Dev efforts in either count. It would be nice if Tom's could lead the reporting on the tool count and timing on delivery in their reporting. Restating the belief that Intel only has 3 or 4 tools out of 13 total (or 16 including IMEC/NYDEV) doing this kind of volume is confusing everyone further.
    Intel could have easily made a million wafers in this time with just one machine. Check ASML page for the 5000. Output is still 100+ per hour. It's not like they are claiming 1M fully finished wafers using every feature of high na... I'm not saying it's one but the million doesn't mean it needs to be more than one either.
    Reply
  • Stomx
    1 Million, 1 Billion...Where is something I'd like to buy ?
    Reply
  • adamXpeter
    So TMSC multi-pass to avoid buying these machines and Intel multi-pass because they bought these machines?
    Reply
  • mac2213
    moreslawplease said:
    Intel could have easily made a million wafers in this time with just one machine. Check ASML page for the 5000. Output is still 100+ per hour. It's not like they are claiming 1M fully finished wafers using every feature of high na... I'm not saying it's one but the million doesn't mean it needs to be more than one either.
    Without sharing all that I know here, I'm pretty sure my data would convince you that you are wrong. I would share privately with you if that is possible and you are interested.
    Reply
  • EzzyB
    If the effort comes to fruition, then Intel will likely have a considerable first-mover advantage over its industry peers because it will define and set the standard for the projection lithography industry for decades to come, an advantage that is hard to overestimate.

    Intel payed a premium to get the first of these machines some time ago when they were mired in their 12 and 10nm (class) then used TSMC to bridge the gap. And, they did it at a time when many were declaring Intel's imminent demise. It's looking like it may very well be paying off for them.

    I remember saying at the time that Intel hadn't forgotten how to make computer chips. I know very few of the commentariat here like to hear anything at all positive about Intel, but, good things like this simply drive on the fierce competition that keeps our processors getting better and better. Will AMD/TSMC respond? Of course they will in time and that's good for all of us.

    I'd love to see Intel use this technology to build a 5070/5070ti class Arc GPU priced around $450. It would set the GPU market on it's head and force more competition there. Almost certainly a pipe dream, but a good one.
    Reply
  • Eximo
    EzzyB said:
    I'd love to see Intel use this technology to build a 5070/5070ti class Arc GPU priced around $450. It would set the GPU market on it's head and force more competition there. Almost certainly a pipe dream, but a good one.

    My understanding is they shuttered every project below a certain profitability. Bringing their GPU production under their roof would only help a little, the rest of the build costs are still there, and still inflated. Memory, PCB fab, assembly, etc all outsourced.

    A more realistic hope is that Nvidia (possibly AMD as well) moves their mid-range GPUs to Intel fabs. Intel could use a steady customer or two, and it would save more of the top-spec TSMC fab time for the stuff they clearly deem more important.
    Reply
  • usertests
    Stomx said:
    1 Million, 1 Billion...Where is something I'd like to buy ?
    Based on the article, they are using it for Panther Lake, but not from the beginning. Panther Lake is generally recognized as a good APU, albeit expensive in these trying times.
    Reply
  • EzzyB
    usertests said:
    Based on the article, they are using it for Panther Lake, but not from the beginning. Panther Lake is generally recognized as a good APU, albeit expensive in these trying times.
    Eximo said:
    My understanding is they shuttered every project below a certain profitability. Bringing their GPU production under their roof would only help a little, the rest of the build costs are still there, and still inflated. Memory, PCB fab, assembly, etc all outsourced.

    A more realistic hope is that Nvidia (possibly AMD as well) moves their mid-range GPUs to Intel fabs. Intel could use a steady customer or two, and it would save more of the top-spec TSMC fab time for the stuff they clearly deem more important.
    My thoughts were a bit darker. Many, many companies in AI are now designing their own chips for fabrication somewhere. Microsoft, Meta, Google, Open AI, and Anthropic I believe are all doing it. So any open capacity is going to get snapped up for things like that.

    There was some talk only a few months ago that Intel was working with Apple to fab it's M processors as well.
    Reply