TSMC: Our 2nm Node Will Beat Intel's 1.8nm Tech

TSMC
(Image credit: Shutterstock)

Although Intel's 20A (2nm-class) and 18A (1.8nm-class) fabrication technologies are set to be available earlier than TSMC's comparable manufacturing processes, the world's largest contract maker of chips believes that its N3P (3nm-class) technology will offer comparable characteristics with Intel's 18A, whereas its N2 (2nm-class) will beat it across the board in terms of power, performance, area (PPA) advantages.

"Our internal assessment shows that our N3P […] demonstrated comparable [power performance area] to [Intel] 18A, my competitor's technology, but with an earlier time to market, better technology maturity, and much better cost," said C.C. Wei, chief executive of TSMC, at the company's earnings call (via The Motley Fool). "In fact, let me repeat again, our 2nm technology without backside power (N2) is more advanced than both N3P and 18A, and will be the semiconductor industry's most advanced technology when it is introduced in 2025."

Intel's 20A fabrication technology, set to arrive in 2024, promises to be a breakthrough in terms of innovations as it will introduce RibbonFET gate-all-around transistors as well as backside power delivery network (BSPDN), two technologies that are designed to enable higher performance, lower power consumption, and increased transistor density. Meanwhile, Intel's 18A production node is designed to refine 20A's innovations further and offer further PPA improvements already in late 2024 – early 2025.

By contrast, TSMC's 3nm-class N3, N3E, N3P, and N3X fabrication processes all rely on proven FinFET transistors and traditional power delivery network. The world's largest foundry does not seem to be in a hurry with its nanosheet GAA transistors and BSPDN, so the former will be introduced by TSMC's N2 node, set to enter high-volume production in 2H 2025, whereas the latter will be added to N2P that will start mass production in late 2026.

One of Intel's main goals in the coming years is to beat TSMC is terms of technology leadership and land foundry orders from companies that need leading-edge nodes. To do so, Intel is set to introduce three advanced manufacturing processes over the next five quarters and initiate volume production on its 2nm and 1.8nm-class manufacturing technologies in 2H 2024 – 1H 2025. But TSMC believes that even its N3P node that will be used in 2025 will offer comparable PPA to Intel's 18A at a lower price, whereas its N2 will beat it, albeit a year after it will enter the market.

Given that TSMC hasn't disclosed much in terms regarding numbers about its N3P and N2, it is hard to make any conclusions about their competitive capabilities against Intel's 18A. Meanwhile, it is evident that TSMC is very confident in its upcoming process nodes.

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.

  • brandonjclark
    I believe this is officially called a "pissing contest".
    Reply
  • thisisaname
    brandonjclark said:
    I believe this is officially called a "pissing contest".
    Or my Dad is better than your Dad.
    Reply
  • JamesJones44
    Usually when companies start telling you their not worried, they really are. Setting up to be a fun fight. I hope Samsung can get its hat in this ring too. The more competition the better.
    Reply
  • Co BIY
    They've both had "comparable" abilities for a while now.

    What will be new is Intel offering theirs as a foundry.
    Reply
  • jkflipflop98
    Sweet summer children and their short memories. You'll soon be reminded who the leader really is.
    Reply
  • Toadster88
    so is it their 3nm or 2nm that is comparable?

    "
    TSMC: Our 3nm Node Comparable to Intel's 1.8nm TechBy Anton Shilov published about 8 hours ago
    TSMC believes its 2nm technology will beat Intel's 1.8nm-class process."

    or is this the whole node vs nanometer naming circus?
    Reply
  • dehjomz
    The proverbial appendage measuring contest. My pp(a) is better than your pp(a). But for TSMC to do this that means they’re feeling the heat because they haven’t mentioned intel’s nodes in years. Perhaps they’re trying to head off losing some customers to Intel. But if TSMC is feeling the heat then that might mean Intel has something viable.
    Reply
  • watzupken
    Toadster88 said:
    so is it their 3nm or 2nm that is comparable?

    "
    TSMC: Our 3nm Node Comparable to Intel's 1.8nm TechBy Anton Shilov published about 8 hours ago
    TSMC believes its 2nm technology will beat Intel's 1.8nm-class process."

    or is this the whole node vs nanometer naming circus?
    Perhaps TSMC 2nm is just a optimized 3nm, just like their 6nm that's pretty much derived from their 7nm?

    Anyway, its funny that TSMC can compare with a node that is not ready till 2024. I mean TSMC can claim all they want, or Intel can give their node some impressive name. But at the end of the day, it all boils down to the performance of the actual product. In my opinion, TSMC's 3nm seems like a flop for now, given that Apple's A17 SOC don't appear to perform much better when it comes to power vs performance. It performs better in GPU because of the extra GPU core, but at the expense of higher power draw it seems.
    Reply
  • Lucky_SLS
    i just want the good old days when the max power draw for a gpu was around 250W. A 650W gold PSU to power everything!!!
    Reply
  • usertests
    Honestly, the original headline: "TSMC: Our 3nm Node Comparable to Intel's 1.8nm Tech" is more impressive if true.

    Intel's renaming scheme was intended to bring their nodes in line with TSMC's nodes, i.e. Intel 4 is the equal of TSMC N4, Intel 20A is the equal of TSMC N2. If the reality is that TSMC N3 is as good as Intel 18A, then Intel's node-a-year strategy is the least they can do to stay competitive.

    Then on top of that, they claim N3 costs less than 18A. Probably true since it's a FinFET node, nothing fancy. They move to GAAFETs at N2.
    Reply