31 comments

  • andix1 minute ago
    I'm wondering if all those GPUs ever end up on the second hand market for us to buy. Or if they will get refurbished multiple times and get used in second tier data centers until the chips die.
  • Animats3 hours ago
    <i>&quot;Secondary market data shows moderately-used 2 to 3 year old GPUs trading at 50% to 70% of new pricing under normal conditions.&quot;</i><p>That&#x27;s for good NVidia H100 units.[1] There&#x27;s a shortage of those. That seems to be the price after removal, cleaning, testing and refurbishing. Raw units removed from a shutdown will not be as valuable.<p>H100 units are available on eBay, but multiple sellers are using the same picture of a new unit in its original packaging, a bad sign.[2] Some even have pictures with the logos of a competitor.<p>[1] <a href="https:&#x2F;&#x2F;introl.com&#x2F;blog&#x2F;secondary-gpu-markets-buying-selling-used-hardware-guide-2025" rel="nofollow">https:&#x2F;&#x2F;introl.com&#x2F;blog&#x2F;secondary-gpu-markets-buying-selling...</a><p>[2] <a href="https:&#x2F;&#x2F;www.ebay.com&#x2F;shop&#x2F;nvidia-h100-gpu?_nkw=nvidia+h100+gpu&amp;msockid=3f8b9a1d8ff965f92c9d881c8e9664ed" rel="nofollow">https:&#x2F;&#x2F;www.ebay.com&#x2F;shop&#x2F;nvidia-h100-gpu?_nkw=nvidia+h100+g...</a>
    • lightbendover3 hours ago
      Who is buying these at 70% of new pricing given the sky high likelihood of them being shot? Maybe it&#x27;s safe to buy from small labs that went under quickly, but I can&#x27;t imagine a cluster that has been operating near its thermal limits for a couple years fetching that kind of resale.
      • christina971 hour ago
        Why would they be shot? Unlike the consumer cards that are basically factory overclocked to look good on benchmarks, the datacenter GPUs are designed to run at full tilt 24&#x2F;7 and survive for years.
      • threetonesun2 hours ago
        It was true of crypto GPUs too, although mostly people picking them up for gaming. Always seems high to me too but if you can get any guarantee of them not being on fire when they were pulled the bathtub curve keeps you pretty safe, thermal limits are limits for a reason.
        • rxyz2 hours ago
          crypto gpus didnt run at 100% power draw so the strain wasn&#x27;t that massive
          • azeemba2 hours ago
            Why weren&#x27;t crypto gpus running at 100%?
            • dantillberg2 hours ago
              Most crypto mining on GPUs would use 100% of memory bandwidth, but only a fraction of the compute available. This is a consequence of ASIC resistance of their mining algorithms -- custom silicon can only offer a modest benefit over GPUs if the hard part is memory bandwidth.
              • NuclearPM1 hour ago
                Why is that true? Can’t you just make more stuff parallel and shrink the ASIC chips accordingly?
            • fishgoesblub31 minutes ago
              Most were powerlimited to a degree to get the most hash&#x2F;watt out of them.
          • metalliqaz2 hours ago
            I bought a RTX 3070 off a miner when Eth went to proof of stake.<p>It was clean, cheap, and is still going strong for daily gaming.<p>In its working life it was undervolted and probably cooled better than in my rig.
            • BizarroLand1 hour ago
              I got an entire Prebuilt PC when POW ended from a miner. AMD 5950x, 64gb ram, 3090, 2tb SSD, all for $1700. This was 4 years ago when the 3090 was $1500 by itself, and with minor upgrades it&#x27;s still going strong today.<p>It&#x27;s wild to think that the system now is worth at least as much as I paid for it then if not much more than that. I saw a similar one going for $2500.
      • Arbortheus1 hour ago
        I feel like this is not as important as people make it out to be.
  • tyfon6 hours ago
    I half expect Nvidia to have buyback contacts like Ferrari with the larger customers to prevent a price crash when they all upgrade and to keep them scarce.<p>I hope not though, perhaps I can pick up a H100 in a few years if they get sold on the open market.
    • rickypp5 hours ago
      I worked at an org that had a substantial on-prem GPU datacenter. We transitioned to &lt;Big Cloud Provider&gt; with a substantial negotiated discount rate, with part of the contract being we would sell them all of our hardware and not purchase any more.
      • throwaway858254 hours ago
        How does this work? Is the OEM giving the cloud provider a big discount or is the cloud provider giving you a teaser rate to lock up your business.
        • wmf3 hours ago
          The list price of cloud is 10x higher than the price of hardware itself so that leaves room for some discounting.
          • to11mtm1 hour ago
            Yeah....<p>Where folks (often) get lazy is the resulting math over what the real bean-counters care about (but are too lazy to check often).<p>In a past life, I worked on costing models for a Cable&#x2F;Fiber contract house, to help the company decide &#x27;what was profitable to keep in house&#x27; versus &#x27;what do we subcontract&#x27; (sometimes that could even mean we just &#x27;rented&#x27; a machine and had a qualified operator using it, based on that employee&#x27;s hourly rate and expected L2R for taxes... so many spreadsheets...)<p>And from from my &#x27;I don&#x27;t know all the factors for this but I&#x27;ve seen how people screw up the big ones&#x27; view (and frankly, I&#x27;m guessing a lot of us have seen and dealt with the same category of &#x27;bad math&#x27; around outsourcing IT work...)<p>An on-prem data center means:<p>- You need to account for electricity costs - i.e. CA vs midwest electric rates.<p>- cooling and power backup capability - Smaller factor but real<p>- personnel cost - e.x. there&#x27;s probably cases where a smaller org could be better off with &#x27;on-site&#x27; server admins that have other roles based on local wages. Kinda case specfic but it&#x27;s a case.<p>- whatever the &#x27;space&#x27; holding the stuff costs<p><pre><code> - Sardonic take :Hey, let&#x27;s have another unused meeting room instead! (e.x. In the case of on-prem shops that simply fled to AWS in their migration from VMware) - the cost of licensing whatever is running - In defense of this, In one of my earliest IT lives, AWS handling the Oracle licensing for a DB was a *huge* win as far as making it as easy as possible to ensure whatever was going on we couldn&#x27;t have the Oracle licensing folks &#x27;ding&#x27; us on whatever infraction occurred between reviews (that could not be understood by the majority of the company, often including the accused. I was never guilty but I saw it happen to others.) - OTOH I know lots of folks who just want to be lazy about what they have to document. </code></pre> Still, IMO a lot of orgs don&#x27;t do the right math around these decisions, or just buy into the &#x27;Well trends can change&#x27; as though they can decide as an org they need to suddenly triple capacity in a month and it would be able to organically happen in the first place.<p>Frankly, the orgs that &#x27;might&#x27; need that either have their arch set up where they are in cloud, or they are onprem but can scale to cloud if needed in interim.
      • hamandcheese4 hours ago
        &gt; and not purchase any more.<p>why would anyone sign such a contract?
        • toast04 hours ago
          If you&#x27;re planning to run in clouds, committing to not buy hardware (during the contract term, presumably) isn&#x27;t a big imposition. Maybe you switch to a different cloud, and you wouldn&#x27;t buy hardware for that.<p>If you want to switch back to on prem, there&#x27;s probably a way to structure acquiring hardware so it doesn&#x27;t break the contract. Maybe you lease it, maybe the purchase happens through a related company, maybe there was no way for the contracted cloud to find out...
        • Shorel3 hours ago
          The usual very short term corporate thinking that maximizes quarter profits while bankrupting the company in the long term. IMO a very shortsighted decision, if not downright stupid.
        • collabs3 hours ago
          I don&#x27;t know if this was IT shrugging me off or if it was something real but some IT person at this big ISP I worked at told me that they cannot just buy an SSD — my windows box at work was running off of a hard disk in 2019 — and that there was some contract that said any computer hardware we bought had to be through HP or something like that and it takes many months it something like that.
          • protocolture33 minutes ago
            Was it a HP Finance thing? Because there are a shit ton of people who can transact through them now. We had HP finance, and it became a game of sorts to see what crazy shit we could get with it. Theres a local mob who sell computer parts who were happy to use HP finance. That said it definitely wasnt everyone and if you didnt do the legwork you could definitely be trapped.
    • nylonstrung4 hours ago
      There&#x27;s going to be a golden age of GPGPU compute in the next few years once A100&#x2F;H100 are fully obsolete for running frontier models efficiently and the price plummets<p>It will be perfect for stuff like GPU-accelerated query engines, &quot;classical ML&quot; and every other CPU-based workload that could conceivably be offloaded to GPU
      • throwaway8943454 hours ago
        GPGPU? General Purpose GPU? If embarrassingly parallel CPU algorithms weren&#x27;t offloaded to the GPU previously, why would the A100&#x2F;H100 price drop make a difference? We had cheap GPU in the past and we still left plenty of performance on the table with CPU programs because they were easier to build.<p>Is the idea that previously maintaining GPU programs was expensive whereas now AI makes it cheap? If so, I could buy that line of reasoning.<p>Maybe relatedly, I expect (hope) the hardware manufacturers will ramp up supply in the meanwhile which would also put downward pressure on GPUs. Right now though this hardware crunch is making me sad, not even because of GPUs but also because of general memory &#x2F; disk.
        • christina973 hours ago
          GPGPU programming has become significantly easier now and the payoff is bigger (better hardware), due to the immense investment in this due to ML&#x2F;AI.
          • NuclearPM1 hour ago
            What are the best tools for this?
        • girvo1 hour ago
          We&#x27;ve not had cheap GPUs with this much VRAM before, though. Might be an interesting change, though I also doubt it personally.
    • moffkalast3 hours ago
      The only reason a datacenter would ditch their H100 is if it becomes uneconomical to run them, with newer silicon providing much more power efficiency. When that happens they&#x27;ll look like a used V100 looks today: horribly inefficient, lacking modern data types and engines, requiring screaming server fans with weird adapters to not melt, way beyond end of life in terms of cuda support. Almost completely damn useless unless you really have no other alternative.
    • someguyiguess5 hours ago
      That should be illegal. Sounds like a very fraudulent business tactic.
      • rcxdude5 hours ago
        Fraudulent, I wouldn&#x27;t say so. Anti-consumer or anti-competitive? Sounds like.
      • rbanffy1 hour ago
        It’s common for car companies when they enter a new market. It removes uncertainty from the second hand market.<p>By doing that, you know upfront what the value of your used hardware will be at the time you decommission it. It removes a lot of the risk for buyers in a volatile market.
      • throw12345678914 hours ago
        Would you call a trade-in a fraudulent tactic?
        • InsideOutSanta4 hours ago
          Trade-ins are voluntary.
          • efficax4 hours ago
            so are buybacks. you choose to sign the contract. there&#x27;s no way they didn&#x27;t have an escape clause, although likely it meant not using the cloud provider anymore
            • InsideOutSanta3 hours ago
              <i>&gt; so are buybacks. you choose to sign the contract</i><p>Right, so they&#x27;re not voluntary.
              • anomaly_3 hours ago
                Literally what? Do you understand what is being proposed?
                • InsideOutSanta2 hours ago
                  Yes. &quot;You choose to sign the contract&quot; is not a reasonable argument, for two reasons:<p>1. There is one supplier, so you have no choice. 2. Even if you had a choice to sign the contract, this still means that it&#x27;s not the same as a trade-in, because trade-ins are always voluntary, but once you have signed the contract, a right of first refusal is not.<p>In general, the &quot;you chose to sign the contract&quot; argument is a poor justification for bad contracts. If the contract is bad, it is bad regardless of whether you chose to sign it.
                  • efficax1 hour ago
                    there are lots of cloud suppliers besides whichever big name cloud this is, you can even find other cloud providers in seattle
      • cyanydeez4 hours ago
        The Grift Economy places all legalities on the marks and their inability to form legal fights.
      • 2III75 hours ago
        Sounds like capitalism at its finest.
        • HPsquared4 hours ago
          Good old &quot;win-win-lose&quot;
    • sidewndr463 hours ago
      isn&#x27;t Ferrari the brand that requires any purchaser to be an existing owner? I could see NVIDIA going for something like that.
      • metadat2 hours ago
        Invite-only is only for special edition hypercars and halo models. Standard models can be purchased by anyone with the funds and desire.
      • driverdan28 minutes ago
        Porsche is doing that for RS cars.
    • echelon5 hours ago
      Wouldn&#x27;t that be crazy - a hobbyist market for H100s?<p>Maybe someone could start a business buying up and rehousing these.
      • rtkwe5 hours ago
        They&#x27;re pretty specific to the datacenter use case with no outputs and they need to be cooled externally, principally through the very loud high speed fans used in data centers. I suppose you could strap a fan to one and put it in a normal case or maybe make a dedicated after market cooler (like the water blocks made for water cooling cases).
        • jtolmar4 hours ago
          I think there&#x27;s a market for a home AI server that can run an LLM or video gen model behind a web frontend. Not literally a raspberry pi strapped to an H100, but something with lopsided enough specs that people joke it is.<p>And precisely because it&#x27;s such a huge headache to do yourself, I think a small company could make a nice business wrapping up used datacenter cards in that sort of server.
          • everforward30 minutes ago
            I&#x27;m doubtful they can, based on current inference prices. An H100 draws about 50W while idling, which is ~$8&#x2F;month at average US electricity prices. They also draw ~400W while active.<p>The electricity prices are relevant because if you paid $0 for your H100 and didn&#x27;t use it a single time, you could buy millions of tokens in inference just on the electricity it draws while idling. If you can&#x27;t keep that thing saturated through the night, you&#x27;re probably underwater overnight. Likewise, it&#x27;s too small to run even the frontier open source models so you need to be able to live with worse models.<p>Max power matters because you aren&#x27;t going to run many of those H100s before you blow breakers in most houses. Newer houses in the US are 15A service to non-kitchen breakers, so 1650W (that might be peak rather than continuous, not sure). If you&#x27;re plugging that into an existing run, you could maybe run 2 before you start blowing breakers? You can&#x27;t just plug 4 H100s into the wall in a normal house.<p>Maybe I&#x27;m wrong, though. I&#x27;d be curious, it&#x27;d be neat to run my own inference for something more than what&#x27;ll run on a 3080.
            • smalltorch4 minutes ago
              Most houses have a 200amp panel and extra space for larger breakers.<p>A electrician can plop in a electric car charger for instance, that is a 40-50 amp circuit.
        • timmmmmmay4 hours ago
          This is also true of the older P100 and hobbyists do this stuff now, today
        • chasd002 hours ago
          &gt; Very loud high speed fans<p>if you haven&#x27;t heard a 5u server intended for a datacenter rack come to life it&#x27;s quite the experience. Sounds like a plane taking off.
          • gerdesj59 minutes ago
            &quot;Sounds like a plane taking off.&quot;<p>Helicopter. I live in Yeovil, Somerset, UK - there&#x27;s a helicopter factory just down the road. I had a IBM &quot;AS&#x2F;400&quot; or whatever they are called now in our computer room rack for a customer and it made nearly as much noise as everything else put together. It was clearly tuned for start up noise to impress because they would fire up in sequence, rise to a crescendo and then slow down in sequence to just a din instead of painfully loud.<p>A switch or PC server on boot will normally run up cooling fans instantly to max as a default protection mechanism until the &quot;OS&quot; has started and sensors read and then the fans will slow down to deal with the actual thermal load.<p>If you switch off your air conn, it gets noisy, quickly. Recently in the UK we are seeing routine temperatures around 30C and we broke 200 odd year records for temperatures a few weeks back. I know its even worse elsewhere but our infrastructure is not designed for this. Here we are at the same latitude as Calgary AB!
          • rbanffy1 hour ago
            My workstations are usually tower servers, which are the same design as their 4 and 5u rack counterparts.<p>Until the thermal management kicks in, they sound like jet planes. When the thermal management starts and assesses the required cooling, it’ll throttle down the fans to reasonable levels.<p>That is, until the moment you push the machine to its limits. When then happens, you might get back to the same levels of the boot time, but it’ll require you to push everything to the max - CPU, memory, storage (all 24 bays) and so on. For a normal user, there is a lot of room and it’s virtually impossible, even with a dozen of Teams windows open.
            • baby_souffle15 minutes ago
              Yes, but the line between &quot;reasonably quiet&quot; and 100% isn&#x27;t a step function.<p>If these do end up in Home Labs, it&#x27;s going to be a server rack in the basement and not the server rack at the other end of the room.
        • bayindirh5 hours ago
          Or you can secure yourself a DLC one and try to feed it with the correct regime (liquid composition, temperature and flow rate). I&#x27;d say good luck.<p>These things get hot and are fussy about their requirements.
      • wmf3 hours ago
        H100s (and the PCIe converter card you need) are available on eBay.
      • mschuster915 hours ago
        The GPU alone has a TDP of 700W, together with everything else (CPU, RAM, storage, fans) you&#x27;re looking at 1500W+. Depending on the country, that may be enough to saturate your home&#x27;s electricity uplink...
        • InsideOutSanta4 hours ago
          That&#x27;s actually less bad than I assumed. High-end gaming GPUs are already almost at 600W with peak usage above that. I assumed it was much worse than that; that seems absolutely feasible for running at home.
        • elictronic3 hours ago
          Never heard electrical uplink before but to those wondering Italy, India, and Japan all have requirements around 3kw. I was slightly surprised by this, but in the end if your running this in a tiny space with that low of power your already probably not buying used H100s.
        • holoduke5 hours ago
          Hell no. 1.5kw is half a socket capacity. A heat pump or many kitchen appliances are using much more. Our car charger uses more as well.
          • jermaustin15 hours ago
            That&#x27;s why they said &quot;depending on country&quot;<p>In North America we are on 120V, making a standard 15A outlet only 1500W max, and something like 1200W sustained. To use higher wattage appliances, we have to upgrade our outlets to 20A (2000&#x2F;1600W) or up our voltage to 240V, but that carries a different set of plugs and outlets as well.
            • datadrivenangel5 hours ago
              Usually the home service is 200+ amps these days which is 24 KW total across all circuits, though you&#x27;re right that most individual circuits are only 10-15 amp.
            • cyberax4 hours ago
              &gt; In North America we are on 120V, making a standard 15A outlet only 1500W max, and something like 1200W sustained.<p>It&#x27;s 1800W for short periods and 1500W sustained.
          • tyfon5 hours ago
            In the US they have a 120V system so it might actually saturate a normal socket over there. My PC room has a 16A fuse and 230V though, so should be plenty :)
        • httpz4 hours ago
          A typical hair dryer uses about 1500W. I guess GPUs are mechanically not much different from a hair dryer.
          • Crunchified4 hours ago
            Install it in your bathroom, use it like a restroom hand dryer.
            • rbanffy1 hour ago
              &gt; use it like a restroom hand dryer<p>One that runs continuously
  • chasil6 minutes ago
    Should I be worried that my plant is owned by Apollo?<p><i>No, because I am retiring next month!</i>
  • gnfargbl3 hours ago
    <i>&gt; The job of an operations team is to keep all of this in steady state. They know which racks run hot in summer, which cooling loops have been flaky since the last firmware update, which jobs to re-route when a node degrades but has not failed yet. None of that knowledge is written down. It lives in the team.</i><p>Hmmn. All of this information <i>should</i> live in the monitoring system, in which case any frontier model <i>will</i> be able to get to grips with it in short order. It feels like the author doesn&#x27;t really fully understand the changes brought about by the systems they are writing about.
    • Planktonne3 hours ago
      It&#x27;s generated prose; the author&#x27;s understanding doesn&#x27;t factor in, because they didn&#x27;t actually write it.
  • fancyfredbot5 hours ago
    The relatively slow depreciation of GPU value is an artifact of supply constraints. If you run fp4 inference and could choose freely between Hopper and a Rubin, the performance per watt would make the Hopper unattractive even if you paid zero for the hardware and only for the power.<p>You can&#x27;t get the Rubin, or even the Blackwell, so you will pay for the H100 but this won&#x27;t last if fabs ramp up capacity.
    • tracker14 hours ago
      Not to mention, physical limits to lithography are slowing down significantly... so tech will continue to evolve more slowly... it&#x27;ll never be the jump from 1080-1990 again, for example, even though 1990-2000 was pretty close, 2000-2010 much slower and since 2010 slower still.<p>What&#x27;s as or more weird is how much hardware is backordered, and how much live hardware is allocated, but waiting on facilities for operation. And how many facilities are years behind at this point already... all on various credit and dept swaps between all the involved companies... it&#x27;s not just a balloon, it&#x27;s a house of cards balanced on a balloon.
      • chuckadams2 hours ago
        The process node size in William the Conqueror&#x27;s time was really off the charts.
        • rbanffy1 hour ago
          Plotting that backwards shows a transistor would be the size of a continent, at least.
      • rbanffy1 hour ago
        &gt; since 2010 slower still<p>The increase in PFLOPS&#x2F;dollar has continued accelerating, a lot from process, but also a lot by simplifying the architecture- if you had placed an H100 worth of transistors on a CPU-like architecture, you wouldn’t reach the same peak performances.
      • cyanydeez4 hours ago
        supposedly, the next step is into fiber &amp; optics.<p>but I generally agree, people put a lot of faith in the exponential leaps vs the exponential space.<p>You tell them we&#x27;re not living on mars any time soon and they&#x27;ll bring up christopher columbus.
        • wyre2 hours ago
          How would fiber and optics help if the bottleneck is computation efficiency and not data transfer?
    • pianopatrick2 hours ago
      Are fabs going to ramp up capacity? Wouldn&#x27;t it make more business sense for them to just not ramp up capacity and enjoy the higher prices?
      • fancyfredbot51 minutes ago
        Yes they are going to ramp up capacity. It would only make business sense to do nothing if all your competitors were also doing nothing, which would probably require some level of illegal collision.<p>If everyone ramps up then in the best case everyone has the same sized slice of a bigger pie. So in theory it makes business sense. But the more realistic possibility is that you end up with oversupply, crash the market and everyone loses. This is what normally seems to happen with DRAM.
    • hinkley4 hours ago
      I wonder what shovels were worth after the Gold Rush faltered. Blacksmiths probably had all the scrap iron they could ever care for.
      • NohatCoder4 hours ago
        Nah, &quot;selling shovels&quot; is mostly a metaphor, the amount of iron that went into mining equipment was insignificant, beyond some local demand peaks. The majority of the business was consumables.<p>A key thing to understand about the gold rush is that it was not a major economic event, or at least nowhere near as big as the participants thought it would be, hence the tradegy.<p>The AI gold rush is different in that there actually is a mountain of &quot;shovels&quot; large enough to flood the global market quite severely.
  • narrator4 hours ago
    One data point: 512GB Mac Studios are selling on ebay for double what they were selling for new at the beginning of the year.
    • girvo1 hour ago
      My DGX Spark-alike is currently selling in my country for nearly double what I paid for it too. We live in a silly, silly world.
    • mtmail3 hours ago
      <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=49005798">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=49005798</a> &quot;I have an M3 ultra mac studio with 512 GB of memory. I want to sell it, [...] Given the high cost of the Mac studio ($20,000 or more)&quot;
    • wmf3 hours ago
      Anything with RAM in it is an appreciating asset.
  • acd5 hours ago
    Used GPU hosting for pension funds.<p>You go to ebay search for a used GPU. You get a price.<p>Neither is used servers a new thing or used routers. There are established used server companies.<p><a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;The_Emperor%27s_New_Clothes" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;The_Emperor%27s_New_Clothes</a>
  • ponkpanda3 hours ago
    Very difficult to sense check that substack post without access to the TLB credit agreement.<p>There are likely management service agreements from xAI proper -&gt; SPV to cover precisely what the author talks about. Clearly, xAI could play games but without seeing the docs (which are not public), it&#x27;s very difficult.<p>This article&#x27;s basic point is right though. On the other hand, the LTV of this deal was approx 50% debt-financed (not too high; very much depends on the &quot;V&quot;). At 12.5%, it&#x27;s not as if its being priced as a high quality asset.<p>Overall, substack post was too bearish. The wider point is that there&#x27;s a lot of froth tied to what has now become systemically opaque - namely the circular deal flow that every hyperscaler, nvidia, neoclouds and friends are now engaged in. When the proverbial hits the fan, that stuff will be difficult to price and find few willing buyers with the competence to underwrite.<p>The systemic issues are the bigger concern than one specific deal imo.
  • cmiles84 hours ago
    All indications are there will be a lot of repossessed GPUs appearing on the market before too long. Likely to be messy for a while but will open up a lot of possibilities when it’s easy to get your hands on some secondhand GPUs.
    • tomaskafka4 hours ago
      This sounds like weapons market after the crash of soviet republic. Never has been a better time to get a functioning tank or parts of nukes.
  • latchkey5 hours ago
    I like Meg a lot a human, but Meg is all doom and gloom. Every single post she makes is about how GPUs fail [0] and now she&#x27;s onto how financing is a big thing just waiting to crash and &quot;nobody knows what a used GPU cluster is worth&quot;...<p>Actually, we do, people offer them to me all the time. A used box of MI300x is $257k. &quot;There is no GPU futures market&quot;... actually there are a few of them that people have pitched to me.<p>This article is a lot of words from someone who isn&#x27;t actually buying or deploying compute. My point is... take it all with a grain of salt.<p>[0] <a href="https:&#x2F;&#x2F;x.com&#x2F;meggmcnulty&#x2F;status&#x2F;2040851080066859386" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;meggmcnulty&#x2F;status&#x2F;2040851080066859386</a>
    • cgyvbunji5 hours ago
      Somehow at any given moment every possible topic to discuss on HN belongs to either the set of &quot;it&#x27;s amazing and nobody can say anything bad about it&quot; or the set of &quot;this thing sucks and nobody is allowed to say anything positive about it&quot; and I never have ANY idea which one any given topic will be in on any given day.
      • fragmede5 hours ago
        latchkey is being restrained. He runs a data center filled with AMD GPUs. He&#x27;s got a lot more insight to the business of it than the post does.
        • cgyvbunji4 hours ago
          That&#x27;s great, their comment should get more attention then.<p>I was complaining that it&#x27;s obviously incorrect that nobody knows what used GPUs are worth, not about latchkey.<p>I have NO idea why anyone is upvoting a post titled &quot;nobody knows what a used GPU cluster is worth&quot;, that is a WILD claim.
          • latchkey4 hours ago
            she makes a lot of wild claims for clicks and gets to the front page with them... sigh.
        • owebmaster2 hours ago
          In 2008, was the opinion of a banker more &quot;insightful&quot; than they opinion of journalists, bloggers and normal people talking about the imminent subprime crash?
          • CookieCrisp2 hours ago
            Cherry picked your example there a bit
      • fragmede4 hours ago
        latchkey is being restrained. He runs a data center filled with AMD GPUs. He&#x27;s got a lot more insight to the real, lived experience of such a business than the post seems to have.
    • tharmas4 hours ago
      Aren&#x27;t the AI Datacenters for the Digital Control Grid, so govt contracts?
  • somat3 hours ago
    Isn&#x27;t it worth exactly what you can sell it for. a few ways to do this.<p>slow and awkward, best market match: The auction. sell to highest bidder.<p>faster and more customer friendly but poor market match until a lot of units sold: The store. guess price, adjust up or down to reach sell frequency desired.<p>fast and good market match but takes a knowledgeable customer base: The reverse auction. Start with price too high lower it over time until it sells.
    • charcircuit1 hour ago
      &gt;adjust up or down to reach sell frequency desired.<p>&gt;Start with price too high lower it over time until it sells.<p>These are the same strategy.
  • Zigurd4 hours ago
    Time to dig up the spreadsheet models of what dark fiber is worth.
  • keeda1 hour ago
    A relevant article I found from an industry insider (which could indicate bias but also relevance): <a href="https:&#x2F;&#x2F;www.whitefiber.com&#x2F;blog&#x2F;understanding-gpu-lifecycle" rel="nofollow">https:&#x2F;&#x2F;www.whitefiber.com&#x2F;blog&#x2F;understanding-gpu-lifecycle</a><p>Which has this anecdotal data point:<p><i>&gt;... when I left Paperspace in mid-2024, our M4000 GPUs, nine-year-old GPUs, were still consistently utilized at near-total capacity. That’s not a typo. Nine. Years. Old. Still booked, still working, still generating revenue.</i><p>Also, I won&#x27;t claim to understand accounting, but in general it seems it is advantageous to <i>accelerate</i> depreciation schedules for high CapEx industries because they lower taxes: <a href="https:&#x2F;&#x2F;leyton.com&#x2F;us&#x2F;insights&#x2F;articles&#x2F;what-is-accelerated-depreciation-and-why-is-it-a-tax-advantage&#x2F;" rel="nofollow">https:&#x2F;&#x2F;leyton.com&#x2F;us&#x2F;insights&#x2F;articles&#x2F;what-is-accelerated-...</a><p>As such you&#x27;d assume these cloud providers to want faster depreciation of their GPU assets rather than slower? I suppose in this case they do have an incentive to show bigger revenue numbers, but there seems to be a trade-off here that is not being discussed?
  • 1-64 hours ago
    The GPU cluster&#x27;s RAM is probably worth more than the flops these days.
  • jmyeet2 hours ago
    This is a good article. I&#x27;ve been curious about how this is going to play out. A couple of data points:<p>1. An enthusiast had a project to get a V100 working on his PC [1]. This was a ~$10k GPU 10 years ago. It&#x27;s now sold for scrap;<p>2. The A100 came out in 2020 and cannot run a large model like DeepSeek v4 Pro. It can run Flash. You need a 16xH100 cluster to run Pro and that&#x27;s a ~4 year old GPU and AFAICT 8xB100 or 4xB200;<p>3. We&#x27;re about to roll out R100&#x2F;R200s.<p>I&#x27;m surprised that NVidia is moving to a 1 year product cycle (per this article) because the big question I&#x27;ve had is what&#x27;s that going to do to existing investments in GPUs. Why? Because if 4xR100 can do the work of 32xH100 then that&#x27;s a massive advantage in performance-per-Watt, which I think is going to be the only metric that ends up mattering.<p>In addition to raw power, new capabilities are developed and come online. For example, certain smaller, more efficient quantization methods just didn&#x27;t exist on older hardware.<p>Oh, another thought from this: a 9% annual failure rate just goes to show you how ridiculous the idea of orbital data centers really is. Orbital DCs were always just a pump-and-dump scheme for SpaceX&#x27;s IPO.<p>Currently it gets expensive to run models larger than ~31B locally. You start to need some pretty expensive hardware. That&#x27;s going to change. I don&#x27;t expect we&#x27;ll be running 1T+ models on a Macbook Pro within 5 years (at reasonable inference rates) but I think people today will be shocked at what&#x27;s being run locally in 5 years and that&#x27;ll easily be 100-200B+ models.<p>[1]: <a href="https:&#x2F;&#x2F;www.hackster.io&#x2F;news&#x2F;hacking-a-server-grade-nvidia-gpu-into-a-home-desktop-674ada7032a7" rel="nofollow">https:&#x2F;&#x2F;www.hackster.io&#x2F;news&#x2F;hacking-a-server-grade-nvidia-g...</a>
  • Zarathustra303 hours ago
    Is the 30-50% of face value realistic for &quot;Liquidation Value&quot;? If one organization has to liquidate, sure. But if the bubble pops and many groups have to liquidate at once? Owners will be lucky if they can dodge the recycling fees.
  • grim_io5 hours ago
    Nothing, because those &quot;GPU&#x27;s&quot; are special proprietary hardware and are not what most people are capable of plugging in into anything.
    • cgyvbunji5 hours ago
      People are selling adapter boards to plug some data center GPUs into a regular pcie slot.
  • semiquaver6 hours ago
    Maybe this article has something useful to say but the painfully LLM-generated prose is too distracting to make it evident.
    • mrhottakes5 hours ago
      Did an AI make all the text and kerning so ugly? Why do web sites look like this now?
      • aqfamnzc4 hours ago
        Isn&#x27;t this just Substack?
    • ijidak5 hours ago
      How do you know? I feel that it&#x27;s just poorly edited.<p>In my experience, AI is easier to read than this was.
      • Game_Ender5 hours ago
        Some sentences feel pretty AI like:<p>&gt; These are not catastrophic events. They are the steady state.<p>&gt; There is no GPU futures market, no standardized residual value curve, and no way to lock in a forward rental rate. The premium is is the price of underwriting in the dark.<p>The headings are also AI like, a lot of essays before usually did not have titled sections but now they do and they all feel like these.<p>In addition the diagrams themselves look pretty AI generated.
        • owebmaster2 hours ago
          Soon we will have people asking if these comments are AI-generated.
  • tim-tday3 hours ago
    I’ll take “people who sell used hardware” for 500 Alex.
  • nikanj2 hours ago
    Essentially nothing, fractions of a penny on a dollar, because players who could afford paying real money won&#x27;t risk the crusty old hardware - so you&#x27;re limited to buyers who still need a massive cluster but don&#x27;t have AI infinite money glitch enabled
  • BenFranklin1005 hours ago
    Would it make economic sense to strip it down and sell for parts? That’s how it’s done now for older data centers, where the obsolete equipment is sent off to China, stripped down for parts, and sold on the secondary market. I’ve picked up several older but still useful RAID hardware cards off eBay this way.<p>I’m mainly interested in getting some DDR4&#x2F;5 and RTX5090s on the cheap :).
  • aslkalska5 hours ago
    so you&#x27;re telling me you can&#x27;t use 1 or 2% percent of 5 billions dollars to rebuild a team that runs GPU clusters for like a couple of years ?! and the guy that borrowed billions from these banks would want to mess up that relationship for what? I mean he is a stupid narcissist but not to that degree. This whole article makes no sense to me.
  • Joel_Mckay4 hours ago
    Given the price for a rack of bc-250 after the crypto hype cycle, the expected value of the hardware will be around 5% to 10% of the original retail price.<p>Without other market influences, that is a &gt;90% expected discount when the over-provisioned market must inevitably self-correct.<p>If the Market follows what Samsung&#x2F;SK Hynix did to the South Korean exchange this week, than the &quot;AI&quot; bubble will hit harder than the dot com crash.<p>I like the Shrek Movie correlation theory, as they always happen just before Debt-backed investors get hit hard... And the new film is due out in 2027. =3
    • owebmaster2 hours ago
      &gt; If the Market follows what Samsung&#x2F;SK Hynix did to the South Korean exchange this week, than the &quot;AI&quot; bubble will hit harder than the dot com crash<p>Can you tell us more about this? Or some link
      • HAL300022 minutes ago
        Samsung and SK Hynix together account for around 60% of the Kospi&#x27;s (SK stock exchange) market capitalization.<p>Over the past few weeks Kospi index has tumbled 25% since its June peak, resulting in a $1 trillion wipeout and its chipmaker duo have both lost at least 30% of their value. There have been days of near 10% plunges followed by sharp rebounds driven entirely by shifting confidence in whether AI spending is sustainable.
      • Joel_Mckay25 minutes ago
        Don&#x27;t worry about it... I am more focused on the bizarre Shrek film timing phenomena, and pondering whether the pattern will hold again. =3<p>Patrick Boyle gives a summary of the situation, but not the underlying Shrek issue:<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=nJtL9MBVj48" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=nJtL9MBVj48</a>
  • DiabloD34 hours ago
    GPU clusters have, largely, no actual value.<p>If anything, you might have to pay to have them disposed of, they don&#x27;t really have any meaningful used eBay market outside of the randos that want to do high end extreme local inference in their basement.<p>Also, as for RAMmageddon, the inference SBCs that all of the AI bros bought don&#x27;t have DIMMs, they&#x27;re not even the right chip: its all GDDR and LPDDR. The only DDR DIMMs being consumed are for regular non-inference machines that help run the business and service infrastructure behind the scenes.
  • mertleee1 hour ago
    [dead]
  • busymom04 hours ago
    &gt; Silent data corruption (SDC) is the most expensive, where a faulty GPU produces wrong answers without crashing anything, which means a multi-day training run can complete normally and the resulting model weights are quietly poisoned.<p>Anyone know how these get caught ultimately?
    • wmf3 hours ago
      Usually silent data corruption isn&#x27;t caught. If a file fails to open you might realize it&#x27;s been corrupted.
  • timmmmmmay4 hours ago
    [flagged]
    • mlyle4 hours ago
      That still doesn&#x27;t tell you what GPUs will be worth in 5 years, because it depends upon what inference demand is like and what the alternatives are. What will an hour of NVL72 be worth in 2029?<p>(And other things, that we know partially but not fully-- like what failure rate for current generation parts will be under this loading).<p>So we have big uncertainties about the revenue, moderate uncertainty about the proportion of the asset that will survive, and some uncertainty about what operating costs will be. It&#x27;s difficult to turn this into a residual value.<p>Finally, the whole &quot;operating the big facility&quot; thing is not likely to be plug-and-play for a new technical team following a default. How much outage&#x2F;disruption ensues?
      • munchler4 hours ago
        Nobody knows what <i>anything</i> is going to be worth in 5 years.
        • mlyle3 hours ago
          Strictly true, but finance being able to predict this pretty closely is how the entire modern economy works.
        • throwaway858254 hours ago
          If that was true the insurance and loan industries wouldn&#x27;t exist.
  • pier255 hours ago
    so when will xAI default on its debt?
    • wmf3 hours ago
      Never? They&#x27;re currently renting out those GPUs at a profit.
      • owebmaster2 hours ago
        Their debt expects a much (much!) bigger margin tho
  • cgyvbunji5 hours ago
    If nobody knows what a used GPU cluster is worth it means nobody is doing anything important with GPUs - time to short everything. Do you believe it?