Qualcomm Cloud AI 100 Ultra – 150W AI Accelerator Card with 128GB at 548GB/s(www.qualcomm.com)
70 pointsby peter_d_sherman3 hours ago7 comments
  • thway152690372 hours ago
    What's the price?
    • simlevesque2 hours ago
      According to their own website, they cost 1 cent.
      • wewewedxfgdfan hour ago
        Less than I expected for this level of GPU power.

        It's a trend I like.

      • bigwheelsan hour ago
        8,750 in stock at 1 cent, too. The more expensive 47 and 63 cent variants are all sold out.
      • thway152690372 hours ago
        Yeah, I clicked that too. But knowing qualcomm they obviously cost an arm and a leg. Still interesting how much.
        • bigwheels36 minutes ago
          In real life, my research yielded the price to be $10-17K USD.
          • IgorPartola20 minutes ago
            I mean at ten bucks it’s not a bad deal but that’s quite a range.
            • 6 minutes ago
              undefined
        • an hour ago
          undefined
    • 2 hours ago
      undefined
    • martin197542 minutes ago
      hopefully less than 5090?
    • dingdingdang2 hours ago
      Honestly, let me buy one, it's many times faster than the minipc I run local ai on so if the price is right I'm in.
  • tokepi2 hours ago
    If the specs seem outdated, it’s because this chip launched in 2024. It’s available at cloud providers including Cirrascale
    • jauntywundrkindan hour ago
      Thats relieving!

      I'm starting to believe in Qualcomm! The Snapdragon X2 Elite Extreme is alas only available in a single system, with it's tri-channel on-package (!) memory good for 228 GB/s. But it's a beast. The core is just good, even in the mainstream form in other systems: some of the best single core performance you can get, shy of Apple.

      And they are upstreaming their stuff much better! They hired a bunch of really good known FOSS developers and it seems to be working!

      Now they are talking up High Bandwidth Compute as their next big thing. And I want to believe, even if it makes me nervous that they too intend to leave consumer computing completely in the dust. https://www.qualcomm.com/news/onq/2026/11/qualcomm-hbc-valid...

      This is, as commented, at M4 Max levels of GB/s. It is not impressive. It has some uses! But if this was their modern opening salvo I would not be impressed.

      • redanddead24 minutes ago
        > I'm starting to believe in Qualcomm

        Someone drop the YouTube link of their robot presentation

  • pearl_panel52 hours ago
    Curious how the memory bandwidth actually holds up under sustained inference loads compared to the spec sheet numbers.
  • swiftcoder2 hours ago
    So it's an M4 Max Mac Studio, RAM-wise, with a GPU's worth of compute?
    • trvz24 minutes ago
      (I know it's you comparing it to the M4 Max, not them, but:) it's Qualcomm, so it'll be compared to the M4 Max, while being slower than a M2 Max.
  • zitterbewegung2 hours ago
    Sort of interesting design It has half a terabyte of SRAM on the accelerator with a 128 GB of RAM replacing HBM with SRAM but keeping regular RAM.
    • 2001zhaozhao2 hours ago
      half a gigabyte not terabyte lol
      • zrm2 hours ago
        Yeah, this thing is unusual. It resembles a consumer GPU but using normal DRAM with more channels, so then you get more RAM but less memory bandwidth than GDDR/HBM, which is the same thing Apple does. And the TDP is pretty attractive for what it is.

        But what is that amount of SRAM for? It simultaneously seems like a lot, because SRAM is so expensive, but still not very much, because what can you fit in 0.5GB? Is there a specific workload that needs something that looks like this?

        • zamadatixan hour ago
          The Intel D270, Nvidia DGX Spark, and (somewhat) AMD Strix Halo also use the same approach, so it seems to be getting pretty common.

          SRAM wise it doesn't seem too insane. About the same ratio of memory:SRAM as the 5090 (IIRC). Still more than I would have expected. I assume they found it helps process the current layer faster or something.

  • aetherspawn2 hours ago
    Ok so it’s significantly worse than Apple M6 and we don’t know if it’s cheaper or not

    Edit: Ah sorry you’re right, the bandwidth is lower but the TOPS is a lot higher. I’m not sure how much effect you get from higher TOPS with such low bandwidth though.

    • zozbot2345 minutes ago
      > I’m not sure how much effect you get from higher TOPS with such low bandwidth though.

      Raised arithmetic intensity. It would be quite good for LLM prefill, and decode would want to adapt by pushing parallelism higher (saving on memory bandwidth by reusing the same layer weights for multiple streams) or, if unable to do so (e.g. due to KV cache constraints) by resorting to speculative decoding (where the expensive 'validate' step handles tokens in parallel within a single stream). It does not make the platform inherently unviable. It would also do well at other tasks such as image/video generation, which is already quite compute bound.

    • 2 hours ago
      undefined
    • dist-epoch2 hours ago
      Apple M6 has 80-100 TOPS.

      Qualcomm Cloud AI 100 Ultra has 800 TOPS, almost 10 times more.

      • grosswaitan hour ago
        This could be very fast at running embedding models (if it supports the right modes)
      • ipsum22 hours ago
        Considering that memory bandwidth is the most important for inference for most neural networks, that matters more.
        • 35 minutes ago
          undefined
    • dist-epochan hour ago
      Modern inference stacks use different GPUs for decode versus prefill. This is a GPU suitable for prefill.
  • varispeed2 hours ago
    Aren't these specs already outdated? Seems meh. I would think players would like to bring something new to the table instead of bringing what was okay two years ago or so.

    Give me 1TB RAM and 5TB/s or something that exciting.

    • entrope2 hours ago
      For most purposes, this seems reasonable? This looks like a single-height card, so fitting four of these in a workstation or server would give 512 GB RAM and 2 TB/s total bandwidth. That's pretty attractive for a lot of current LLMs, even if one wishes for 1 TB capacity at 5 TB/s bandwidth.
      • varispeed34 minutes ago
        But there are already plenty of products for most purposes.
    • tempest_2 hours ago
      5 year old GPUs are still at crazy prices.

      Plenty of businesses want in local inference and are priced out.

      • varispeed33 minutes ago
        You can't do quality inference on such a toy. It's a myth.
        • redanddead20 minutes ago
          A great opportunity for huawei

          Nobody helps China like American corps helps China. Domestic anticompetitive behavior spurs demand for international competition

    • mciancia2 hours ago
      8 of those cards give you 1TB of ram and 4.4TB/s
      • PhunkyPhil2 hours ago
        It may be a semantic thing but I don't think bandwidth pooling works like that? Unless your blocks are distributed perfectly between cards there's going to be data passed between each card and from VRAM to compute. Correct me if I'm wrong though
        • msandfordan hour ago
          I think the way it works is that the weights have to get streamed through the compute and the weights can be split very easily. With multiple cards you do have a coordination/communication phase after each layer of weights but that's usually much less significant than the compute phase.