Most active commenters

    ←back to thread

    Apple reports fourth quarter results

    (www.apple.com)
    176 points mfiguiere | 26 comments | 30 Oct 25 20:34 UTC | HN request time: 0.422s | source | bottom
    1. oxqbldpxo ◴[30 Oct 25 21:15 UTC] No.45765471[source]▶
    All these companies depend on TSMC for their life.
    replies(4): >>45765560 #>>45765582 #>>45765677 #>>45777141 #
    2. sho_hn ◴[30 Oct 25 21:24 UTC] No.45765560[source]▶
    And TSMC depends on machines by ASML they can also sell to others.

    And ASML licensed the technology from EUV LLC.

    Which was a conglomerate of a bunch of state-funded US research labs.

    And the US cut its science funding.

    Misery all the way down!

    replies(2): >>45766464 #>>45767181 #
    3. seizethecheese ◴[30 Oct 25 21:25 UTC] No.45765582[source]▶
    If true, TSMC would command much higher margins. Their net revenue is a fraction of Nvidia or Apple
    replies(2): >>45765618 #>>45779238 #
    4. trenchpilgrim ◴[30 Oct 25 21:30 UTC] No.45765618[source]▶
    TSMC's business is much higher risk, each improvement to manufacturing process is a massive investment that's never a guaranteed success.
    5. nomilk ◴[30 Oct 25 21:38 UTC] No.45765677[source]▶
    Had to look up what TSMC meant (Taiwan Semiconductor Manufacturing Company).

    What would Apple's next best option be if a war rendered TSMC unavailable?

    replies(4): >>45765712 #>>45765755 #>>45766068 #>>45766664 #
    6. madeofpalk ◴[30 Oct 25 21:41 UTC] No.45765712[source]▶
    https://www.apple.com/newsroom/2025/02/apple-will-spend-more...

    > The fund’s expansion includes a multibillion-dollar commitment from Apple to produce advanced silicon in TSMC’s Fab 21 facility in Arizona. Apple is the largest customer at this state-of-the-art facility, which employs more than 2,000 workers to manufacture the chips in the United States. Mass production of Apple chips began last month.

    7. colechristensen ◴[30 Oct 25 21:45 UTC] No.45765755[source]▶
    >What would Apple's next best option be if a war rendered TSMC unavailable?

    Onshore TSMC fabs followed by Intel fabs.

    Properly motivated, I think Intel and Apple could do a lot relatively quickly.

    replies(2): >>45766075 #>>45779052 #
    8. 45764986 ◴[30 Oct 25 22:17 UTC] No.45766068[source]▶
    If a war rendered TSMC unavailable it would crash the global economy. There is no next best option.
    replies(1): >>45777073 #
    9. 45764986 ◴[30 Oct 25 22:17 UTC] No.45766075{3}[source]▶
    The supply chain for this simply does not exist.
    10. yieldcrv ◴[30 Oct 25 23:07 UTC] No.45766464[source]▶
    It’s a conglomerate of researchers that were employed by the feds and private institutions who met have received various forms of grants

    I think the science funding cuts will be inconsequential to that entity

    replies(1): >>45766688 #
    11. martinald ◴[30 Oct 25 23:30 UTC] No.45766664[source]▶
    There's an amazing book on Apple in China all about this issue (and more). It's a great read and I'd highly recommend if you're interested.

    Also Chip Wars is really good. I may be confusing which one is which because I read them back to back, but they overlap!

    replies(1): >>45766863 #
    12. lokar ◴[30 Oct 25 23:33 UTC] No.45766688{3}[source]▶
    But what about the next area where science can have a massive impact?
    replies(1): >>45766940 #
    13. nomilk ◴[30 Oct 25 23:54 UTC] No.45766863{3}[source]▶
    Thanks! I've added both to my reading list
    14. yieldcrv ◴[31 Oct 25 00:08 UTC] No.45766940{4}[source]▶
    sounds like a totally different thread to me
    15. throw0101a ◴[31 Oct 25 00:47 UTC] No.45767181[source]▶
    > And ASML licensed the technology from EUV LLC.

    And glass/mirrors from Zeiss, amongst a whole bunch others:

    > ASML employs more than 42,000 people[1] from 143 nationalities and relies on a network of nearly 5,000 tier 1 suppliers.[6]

    * https://en.wikipedia.org/wiki/ASML_Holding

    * https://www.robotsops.com/complete-list-of-all-suppliers-and...

    Let's also not forget the the two most prominent chip design software companies, Cadence and Synopsys, are American:

    * https://en.wikipedia.org/wiki/List_of_EDA_companies

    There are all sorts of inter-dependencies between companies and countries: welcome to globalization.

    16. astrange ◴[31 Oct 25 21:47 UTC] No.45777073{3}[source]▶
    Samsung, Intel, SMIC are not incredibly far behind. TSMC is the best because we (the US and its customers) trust them more than its competitors and so fund its R&D and license them more exclusive technologies.
    replies(1): >>45777513 #
    17. jayd16 ◴[31 Oct 25 21:56 UTC] No.45777141[source]▶
    I mean... do they? TSMC is the best but in a world where they had to use Samsung or Intel is it really a death sentence?
    replies(3): >>45777377 #>>45778308 #>>45781054 #
    18. ◴[31 Oct 25 22:27 UTC] No.45777377[source]▶
    19. dontlaugh ◴[31 Oct 25 22:42 UTC] No.45777513{4}[source]▶
    SMIC in particular have made very quick progress. They’d probably match TSMC first in such a scenario.
    replies(2): >>45777779 #>>45783911 #
    20. ta9000 ◴[31 Oct 25 23:18 UTC] No.45777779{5}[source]▶
    In case of a war, SMIC would likely also be unavailable.
    replies(1): >>45783933 #
    21. tomcam ◴[01 Nov 25 00:50 UTC] No.45778308[source]▶
    My intuition is that we would adapt just fine. Maybe we'd have to drop to assembly more often, read the chip docs closely, etc. Unheard-of performance benefits are still being found from Commodore 64s and first-gen IBM PCs, for crying out loud. What if we wrung every last cycle out of the Samsung or Intel chips?
    22. throwaway31131 ◴[01 Nov 25 03:34 UTC] No.45779052{3}[source]▶
    TSMC in Taiwan has a significant share of the wafers produced by the world every month. If those wafers were not produced the global economy would suffer badly.

    It takes years to bring a fab online. Fab 21 in Arizona took 5 years to enter mass production from ground breaking. Some believe it could be done in two but that’s yet to be demonstrated. Then there are the wafers themselves. The total time it takes to process one wafer at the single nm scale is around 100 days.

    So realistically, even if one makes up their mind to make a fab fast, you’re looking at 3 years before you have your first sellable wafer.

    23. tyre ◴[01 Nov 25 04:23 UTC] No.45779238[source]▶
    TSMC is famous for not taking margin when they could. It’s part of their strategy.
    24. ◴[01 Nov 25 12:15 UTC] No.45781054[source]▶
    25. astrange ◴[01 Nov 25 18:14 UTC] No.45783911{5}[source]▶
    No they haven't, they're just trying to show off by running older processes with high failure rates very hard to make it look like they can keep up.

    We're not giving them EUV and they can't reinvent it, so they're stuck.

    26. dontlaugh ◴[01 Nov 25 18:16 UTC] No.45783933{6}[source]▶
    Unavailable to whom exactly?