Watch AMD's Computex 2024 keynote live stream here at 9:30 pm ET / 6:30 pm PT / 1:30 am UTC

Lisa Su and AMD
(Image credit: Computex Taipei)

AMD CEO Lisa Su is set to take the stage for a Computex 2024 keynote address here in Taipei at 9:30 a.m. local time on Monday, June 3rd. If you aren't here in Taiwan or you are and you haven't pre-registered, you can't get a seat in the audience. But you can turn watch Dr. Su's presentation as a YouTube Live stream, which we've embedded below.

The event starts at 9:30 a.m. Monday Taiwan time, which is 9:30 pm Sunday (June 2nd) if you're in the U.S. Eastern time zone. It's 6:30 pm Sunday in the Pacific time  zone, 2:30 a.m. BST and 1:30 a.m. UTC.

What can you expect from AMD's Computex keynote presentation? Computex's event description reads simply: "Dr. Lisa Su will explore how AMD, together with its partners, is pushing the limits in AI and high-performance computing from the data center, to the edge and end user devices."

So we can see that she's going to be talking about both data center and end user devices, which means PCs. Will she announce new chips for desktops and laptops? It seems quite possible.

In recent weeks, we've seen leaked benchmarks for the upcoming Zen 5 processor line and motherboard vendors have also added support for Zen 5 (aka Ryzen 9000 series) in their firmware updates.  We've also seen leaked benchmarks from AMD's rumored Strix Point mobile CPUs.

Su has to focus on AI performance in her keynote, not just because AI is the tech industry's hottest buzzword, but also because Microsoft recently made a point of releasing its first-gen Copilot+, AI-enabled PCs with Qualcomm Snapdragon chips only, leaving AMD and Intel behind. To meet the Copilot+ requirements, a computer must have an NPU that delivers at least 40 TOPs (trillion operations per second) and neither AMD nor Intel, at this moment, have a processor in the market that's capable of that. 

Qualcomm can't have a monopoly on Copilot+ PCs for long so expect AMD and Intel to respond with some more details about their plans for AI on the desktop. This keynote would be a great place for Su to showcase AMD's solutions.

Avram Piltch
Avram Piltch is Tom's Hardware's editor-in-chief. When he's not playing with the latest gadgets at work or putting on VR helmets at trade shows, you'll find him rooting his phone, taking apart his PC or coding plugins. With his technical knowledge and passion for testing, Avram developed many real-world benchmarks, including our laptop battery test.
  • usertests
    Tune in to watch CEO Lisa Su discuss her company's AI plans.

    But really, we can expect to hear about Granite Ridge, Strix Point, or both.
  • Metal Messiah.
    Same as previous leaks pointed out. More to come.....

    I'm more interested in the Strix Point Ryzen AI lineup though. They haven't mentioned the MSRP of these SKUs yet....
  • Metal Messiah.
    Here it comes. :grinning:
  • usertests
    +16% IPC, better than Zen 4 if accurate. Going from Zen 2 to Zen 5 is +56% IPC. Sonoma Valley looking good.

    Thanks for the cringe branding. Maybe it's only for the Ryzen 9.

    The 10-core cuts two of the Zen 5c cores, not the higher clocked Zen 5. Nice.

    Let's see 890M graphics compared to 780M plz.

    VideoCardz lists 15-45W configurable TDP for these two Strix Point SKUs, instead of 35-54W.

    9600 non-X theory busted.
  • Metal Messiah.
    Strix Point branding seems confusing as before. Product differentiation for an average consumer is not gonna be easy, unless they first dig into the specs of each SKU, especially the core count.

    Too bad only 2 Mobile processors were announced.

    Both are Ryzen AI 9 SKUs, and the only naming differentiator seems to be the HX moniker for now. These product codenames also don't tell anything about the iGPU specs, be it 890M, or 880M, or how many CUs they sport.

    Ryzen AI 9 HX 370
    Ryzen AI 9 365.
  • usertests
    AM5 socket commitment extended to "2027+".

    RDNA 3+ became RDNA 3.5 again lol.

    XDNA2 gains "Block FP16" capability, with the accuracy of FP16 and the performance of INT8 (they claim).