21 comments

  • gpugreg4 hours ago
    Groq stopped serving Kimi K2 (1T params) when they got aquihired by NVIDIA, so I guess NVIDIA took most of the hardware in addition to the employees. The largest model they serve now is the relatively minuscule gpt-oss-120b.<p>The community support forum is also getting retired and there haven&#x27;t been any posts by support employees in forever anyway, so they are probably gone, too. Also, the number of issues have been piling up, suggesting that the developers are gone as well. <a href="https:&#x2F;&#x2F;community.groq.com&#x2F;c&#x2F;forum&#x2F;4" rel="nofollow">https:&#x2F;&#x2F;community.groq.com&#x2F;c&#x2F;forum&#x2F;4</a> (archive link for when it goes down <a href="https:&#x2F;&#x2F;web.archive.org&#x2F;web&#x2F;20260602064050&#x2F;https:&#x2F;&#x2F;community.groq.com&#x2F;c&#x2F;forum&#x2F;4" rel="nofollow">https:&#x2F;&#x2F;web.archive.org&#x2F;web&#x2F;20260602064050&#x2F;https:&#x2F;&#x2F;community...</a>)<p>To me, it looks they are trying to raise 650M with a few remaining (ancient) LPUs and no employees.
    • batperson1 hour ago
      Before they removed it, I was using groq Kimi K2 model for a chat bot in small community site&#x2F;chat. It was really good, seemed to have incredibly vast general world knowledge and the fast speed (400tok&#x2F;s if I remember right) meant that chat users got a response instantly which was a much better experience compared to other SOTA models at the time.<p>On the bright side it looks like Cerebras might be serving Kimi K2.6 at 1000tok&#x2F;s soon <a href="https:&#x2F;&#x2F;www.cerebras.ai&#x2F;blog&#x2F;cerebras-kimi-k2-Enterprise" rel="nofollow">https:&#x2F;&#x2F;www.cerebras.ai&#x2F;blog&#x2F;cerebras-kimi-k2-Enterprise</a>
  • conshama1 hour ago
    Somehow I never got the hype around Groq. They served models with fast inference speed - that sounded great in theory - and as a user I was looking forward to use them. But, when I did try, I discovered that they are quantizing the models underneath. And they dont even disclose it. So I stopped using them.<p>The whole thing never made any sense to me - but I guess AI hype is a thing.
    • adityashankar22 minutes ago
      I believe despite quantisation they were still extremely fast, which is still incredibly useful if you don&#x27;t need high precision&#x2F;accuracy (which is good enough for many use cases)
  • zamalek2 hours ago
    I&#x27;ve was part of an acquire-to-remove-an-annoyance, including having the better product that didn&#x27;t win for niche reasons - which vaguely matches the situation here. Yes, NVIDIA can wind down Groq (unambiguously a better product in this case) even though it doesn&#x27;t make sense.<p>That being said, there&#x27;s still a chance that NVIDIA engineering is in the process of stripping it for parts. Or the lawyers are - maybe they have too much momentum with GPUs and just want ASICs out of the market.<p>This kind of innovation stifling acquisition should have been blocked. NVIDIA is a serious monopoly threat.
  • ViscountPenguin8 hours ago
    I don&#x27;t really get the value proposition of groq as a user, the performance is really poor for the token price. Data centres on the other hand are becoming a commodity, and I don&#x27;t see any reason a priori to invest in groq specifically for something like that.
    • petesergeant3 hours ago
      &gt; the performance is really poor for the token price<p>That doesn’t match my experience or the numbers:<p><a href="https:&#x2F;&#x2F;openrouter.ai&#x2F;openai&#x2F;gpt-oss-120b?sort=throughput" rel="nofollow">https:&#x2F;&#x2F;openrouter.ai&#x2F;openai&#x2F;gpt-oss-120b?sort=throughput</a>
    • bluegatty8 hours ago
      Groq is considerably faster and better at inference, they have a totally superior product to Nvidia for inference based tasks, which will be the dominant concern in the future.<p>Plausibly, take all the Nvidia hype and multiply that by a factor and that&#x27;s what &#x27;Groq&#x27; could be worth.<p>And there is no real commodification - there&#x27;s Nvidia, Cerebras, Groq ... not many otheres.
      • TurdF3rguson8 hours ago
        &gt; they have a totally superior product to Nvidia for inference based tasks<p>They&#x27;re not really competing with Nvidia because 1) Nvidia owns their chips now, and 2) Nvidia is not really an inference provider.
        • bluegatty7 hours ago
          Groq is a slicon maker, the inference provider stuff is a path to market, it&#x27;s not really the reflection of their market potential.<p>Nvidia doesn&#x27;t own them or all their IP now, we don&#x27;t quite know the terms of the deal.
          • TurdF3rguson6 hours ago
            AFAIK the terms were the chip-making + talent stuff went to Nvidia, and the api provider stuff gets to keep existing separately.
      • 7thpower8 hours ago
        Define “totally superior”?<p>Was this comment created using quantized llama 3?<p>I love Groq, but across every single line break in your post there is a glaring issue that is easy to refute with in 15 seconds, even without 300t&#x2F;s of throughput.
        • bluegatty7 hours ago
          You wasted all of your commentary on snark and sadly unfunny humour, and yet still managed to add nothing.<p>Groq is more performant for the growing categories of inference-based tasks, wherein Nvidia&#x27;s advantage in inference depends bulk&#x2F;batch processing which will make up a smaller category over time, in relative terms.<p>The future of AI Silicon is inference, and the cost structure of AI data centres is constrained around the current necessity to have &#x27;high GPU utilization&#x27; otherwise, the cost &#x2F; amortization of the chips doesn&#x27;t work out.<p>That cost structure is a limitation of Nvidia architecture.<p>Groq serves a lot faster, and without the limiting batching requirement, which opens hosting arrangements common in most classical hosting scenarios aka without necessarily the high utilization requirements.<p>Groq has bespoke hardware, lack of CUDA, much lower memory desnsity obviously and they don&#x27;t have the deep distribution networks and leverage over TSMC that Nvidia has - but pound for pound, were we to be able to &#x27;fire up a server&#x27; for our inference needs, it would be Groq, not Nvidia that we&#x27;d turn to.<p>Were they not a later market entrant and didn&#x27;t have those barriers to entry, they&#x27;d be gigantic.
          • dnautics6 hours ago
            is groq still using 6 racks to serve Llama3-70B or is that old news?
            • wmf5 hours ago
              The new chip isn&#x27;t out yet so that&#x27;s the only thing they could be doing.
      • imtringued3 hours ago
        Google has been releasing a new TPU generation every year since 2023 and the eight generation consists of a training and an inference optimized design.<p>Google&#x27;s eight generation TPU inference chip has 384 MB of on-chip SRAM vs 500 MB for Groq&#x27;s third generation LPU.
      • digitaltrees7 hours ago
        Haha. Groq is trash. It can’t be used for anything where reliable work is required.<p>Groq lasted how long in the social contract experiment? Libertarians always forget that rules and government aren’t tyrannical they are the mechanism to ensure bullies don’t destroy everyone’s freedom.
        • Renaud6 hours ago
          This is not about xAi.
          • digitaltrees6 hours ago
            Ahh, well I stand corrected. But Groq sold out where are our revolutionary chips? locked behind a monopoly. :)
        • SonOfKyuss6 hours ago
          Groq != grok
          • lukan5 hours ago
            I have the feeling quite some up and downvotes here did not take that into account.
  • caterama8 hours ago
    My company had a really terrible experience trying to use Groq, and I would NOT recommend anyone use their service if you need reliability. So many random errors, so many silly quirks.
    • mdp20215 hours ago
      Are you sure your post is about<p>&gt; <i>Groq, the AI chip company company that was acquired by Nvidia in December of last year, is raising $650M</i><p>? Could you provide details?
      • BoorishBears2 hours ago
        What&#x27;s so surprising?<p>I&#x27;d say they have a &quot;Tier 2&quot; inference stack, but that&#x27;d be Apple Silicon and AMD: their well below that.<p>In part because unlike those, no one can get their hands on the hardware, so they miss out on a massive amount of free development, testing, etc.<p>At this point there&#x27;s years of complaints about open weight models performing worse on their platform, tool calling makes it especially easy to tell since it&#x27;s so sensitive to these issues.
  • z3ratul1630717 hours ago
    as soon as i saw they switched to &quot;call us for quotes&quot; for the new models, i knew they are over.
  • internet_points2 hours ago
    Maybe Tom Ellis can give another AMA <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=39429047">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=39429047</a> so we can find out.
  • 0xbadcafebee5 hours ago
    The most bizarre thing here is the reporters. They intentionally misrepresented what happened. All the news stories from last year claimed Nvidia &quot;acquired&quot; Groq - and in the same story, quoted what <i>actually</i> happened, and pretended it didn&#x27;t. It&#x27;s like the journalists had some kind of group psychosis, pretending it was an acquisition. It wasn&#x27;t. It was a <i>really</i> white-glove product rental.<p>From the actual Groq PR release:<p><i>&quot;Groq announced that it has entered into a non-exclusive licensing agreement with Nvidia&quot;</i> - <i>&quot;As part of this agreement, Jonathan Ross, Groq’s Founder, Sunny Madra, Groq’s President, and other members of the Groq team will join Nvidia to help advance and scale the licensed technology&quot;</i> - <i>&quot;Groq will continue to operate as an independent company with Simon Edwards stepping into the role of Chief Executive Officer&quot;</i> (<a href="https:&#x2F;&#x2F;groq.com&#x2F;newsroom&#x2F;groq-and-nvidia-enter-non-exclusive-inference-technology-licensing-agreement-to-accelerate-ai-inference-at-global-scale" rel="nofollow">https:&#x2F;&#x2F;groq.com&#x2F;newsroom&#x2F;groq-and-nvidia-enter-non-exclusiv...</a>)<p>Nothing about an acquisition there. It says Nvidia is licensing it, and that others can too. The execs work for Nvidia to integrate it into Nvidia&#x27;s... something. And Groq the company remains the same as before.<p>There&#x27;s also no official source for the amount Nvidia paid for the tech, or two unofficial ones. Journalistically speaking, this is some bullshit.<p>Why&#x27;s the deal like this? No idea. Does it make sense? No idea. But it&#x27;s not odd that Groq is continuing to raise more money, because they never stopped being a normally operating company.<p>If you want an explanation for <i>why</i> Nvidia would do this deal, my best offer is here (<a href="https:&#x2F;&#x2F;openrouter.ai&#x2F;rankings" rel="nofollow">https:&#x2F;&#x2F;openrouter.ai&#x2F;rankings</a>): Of the top 10 fastest AI models, Groq is the provider of 4 of them. And of the <i>price</i> of those top 10 fastest AI models, Groq is #1, #2, #3, and #5. And you wonder why someone&#x27;s giving them a measly half billion dollars? They&#x27;re the fastest cheapest thing on the market. If you don&#x27;t understand the value of that, you really don&#x27;t understand AI.
    • tverbeure4 hours ago
      &gt; There&#x27;s also no official source for the amount Nvidia paid for the tech,<p>Did you honestly believe that this kind of deal doesn&#x27;t leave a trace in financial disclosures?<p>The February 2026 Nvidia 10-K has this:<p>Cash flows from investing activities: Groq. -$13B.<p>And this: &quot;Total consideration consists of $13.0 billion paid at closing and $4 billion, inclusive of imputed interest, payable within one year included in Accrued and Other Current Liabilities on our Consolidated Balance Sheets.&quot;
  • maz1b3 hours ago
    Deals like these are quite rare, would be nice to know more about how the funding worked, but for customers, what are they doing? Total silence from Groq for a very long time now.
  • xiphias25 hours ago
    I don’t understand one part of the licensing here: if it was just a license, can’t they relicense the software and hardware of LPU3 to AMD? Or hire new software and hardware people?<p>The new designs were their main asset besides the amazing talent that went to NVIDIA, not the remaining DCs.
  • markpotts1238 hours ago
    what&#x27;s the confusion. Groq offers a fast inference solution that is currently unique (why do you think Nvidia paid $8 billion to end-run around the SEC to acquire the technology). This is good news as it ensures that Groq customers can can be assured continuity to use their service.
  • andai7 hours ago
    Do they have any good models yet?
    • herrvogel-7 hours ago
      It’s not X AI, that would be Grok with a ‘k’.
      • jrflowers6 hours ago
        You pay them money to run models. Their website doesn’t list them as offering any models that were released recently. For a “pay for inference” provider, questions like “do they have Deepseek or Qwen 2.6?” are germane
    • mdp20215 hours ago
      The list is:<p>-- allam-2-7b<p>-- canopylabs&#x2F;orpheus-arabic-saudi<p>-- canopylabs&#x2F;orpheus-v1-english<p>-- groq&#x2F;compound<p>-- groq&#x2F;compound-mini<p>-- llama-3.1-8b-instant<p>-- llama-3.3-70b-versatile<p>-- meta-llama&#x2F;llama-4-scout-17b-16e-instruct<p>-- meta-llama&#x2F;llama-prompt-guard-2-22m<p>-- meta-llama&#x2F;llama-prompt-guard-2-86m<p>-- openai&#x2F;gpt-oss-120b<p>-- openai&#x2F;gpt-oss-20b<p>-- openai&#x2F;gpt-oss-safeguard-20b<p>-- qwen&#x2F;qwen3-32b<p>-- whisper-large-v3<p>-- whisper-large-v3-turbo
  • fontain8 hours ago
    I’m confused by the confusion. Groq licensed their technology (sold part of their business) to Nvidia for a large amount of money and distributed the spoils to their investors. Seems quite normal? But then the Axios article says…<p>“Existing shareholders will receive the remaining cash distributions and then have the opportunity to invest into a new company”<p>New company? But Groq still exists and continued to exist.<p>“The bottom line: Don&#x27;t be surprised if this becomes a new transaction template in the AI private markets.”<p>A transaction template? I don’t follow what was novel about this situation. The Meta not-acquisition-acquisition of Scale seems more novel.<p>I guess I feel like Zach’s confusion is because of the way Axios has presented what is happening to Groq. Looking at why actually happened with Groq, it seems like Axios are reporting it weird.<p>Unless Groq really is starting a new company in which case I am equally as confused.<p>edit: when announced last year it was announced as an asset acquisition <a href="https:&#x2F;&#x2F;www.cnbc.com&#x2F;2025&#x2F;12&#x2F;24&#x2F;nvidia-buying-ai-chip-startup-groq-for-about-20-billion-biggest-deal.html" rel="nofollow">https:&#x2F;&#x2F;www.cnbc.com&#x2F;2025&#x2F;12&#x2F;24&#x2F;nvidia-buying-ai-chip-startu...</a>
    • zachbee8 hours ago
      The interesting thing here isn&#x27;t &quot;how, logistically, is the Groq corporate entity able to raise more money?&quot;. That&#x27;s straightforward.<p>Rather, the interesting thing and the topic of most of the article is &quot;how, after Nvidia hired most of Groq&#x27;s team and licensed all their IP, did Groq manage to convince investors to invest in the remaining corporate entity?&quot;
      • fontain8 hours ago
        I thought you wrote the convincing explanation:<p>“One could argue that Groq’s datacenters alone could make them worth billions of dollars.”<p>Groq is a successful datacenter business with a high-revenue cloud product. That’s a compelling investment in its own right, right?
        • Ardren8 hours ago
          &gt; Groq Launches European Data Center Footprint in Helsinki, Finland<p><a href="https:&#x2F;&#x2F;groq.com&#x2F;newsroom&#x2F;groq-launches-european-data-center-footprint-in-helsinki-finland" rel="nofollow">https:&#x2F;&#x2F;groq.com&#x2F;newsroom&#x2F;groq-launches-european-data-center...</a><p>That sounds like they are renting racks in a Equinix data centre. Do Groq have 4 data centers worth billions?
    • bluegatty8 hours ago
      There&#x27;s nothing normal at all about the Nvidia Groq deal, it&#x27;s hard to read in terms of what it means. A straight licensing deal would have been easier to ingest.
      • fontain8 hours ago
        I could be completely off the mark but I thought the non-exclusive license was necessary because Groq’s datacenter business uses the technology already? Nvidia acquired the assets but Groq needed to retain rights to use the technology for their own product.
        • bluegatty7 hours ago
          The deal was probably structured the way it was due to concerns over regulatory approval.<p>These &#x27;we get your executives&#x27; type of deals - aka Windsurf - are new, weird thing in M&amp;A.
        • wmf8 hours ago
          They could have sold the IP then licensed it back. The nonexclusive part was purely a fig leaf to dodge antitrust.
  • haeseong5 hours ago
    [dead]
  • trippsydrippsy3 hours ago
    [dead]
  • vladsiu7 hours ago
    [dead]
  • iririririr8 hours ago
    [dead]
  • ai_fry_ur_brain9 hours ago
    [flagged]
    • dang9 hours ago
      Please don&#x27;t post shallow-indignant comments. The article raises an interesting question. If the discussion terminates in an angry cliché before it even gets started, that&#x27;s a boring outcome.
      • ai_fry_ur_brain8 hours ago
        Well dang, that&#x27;s sort of my goal. I think the masses should be angry and should be highly polarized (regardless if its cliche) against the groups of people investing billions in chat bots while millions can&#x27;t afford food and medicine.
        • unmole7 hours ago
          Take your middlebrow slacktivism elsewhere. HN is not the place for it.
        • appplication8 hours ago
          I’ll also agree with you, this is not a topic that needs nuance, it’s as straightforward as fuck the billionaire class.<p>If you want nuance, the obvious answer to this is that the rules that apply at our level do not apply to them. Raising money is an inevitability and does not require any fundamental basis other than the name behind it.
    • pezezin9 hours ago
      [flagged]
  • digitaltrees7 hours ago
    [flagged]
    • Renaud6 hours ago
      Groq =&#x2F;= Grok
  • fareesh5 hours ago
    i like groq but models seem to have stagnated - looks like the company isn&#x27;t focused on b2c anymore?
    • ares6235 hours ago
      Are stagnating models really a problem? Because if so, then how is &quot;if we stop training we are profitable&quot; an acceptable answer to the profitability question?
      • dkersten2 hours ago
        Groq don&#x27;t train their own models, they serve open ones. They&#x27;re a few versions behind the frontier open weights models. The top model they serve is GPT OSS 120B, released last summer.