15 comments

  • darkwizard421 hour ago
    Seeing a lot of tricks similar to how ridesharing companies tried to be &quot;profitable&quot; before going to IPO. Caveat: Thing have materially improved but really Uber is carried by its insane Ads margins<p>The idea of removing model training from your costs is a little wild tbh.<p>The profitability of being able to serve a query wasn&#x27;t really under question (nor is the margin expected to be anything less than 80%+) I think.
    • embedding-shape59 minutes ago
      &gt; The idea of removing model training from your costs is a little wild tbh.<p>Yeah, I didn&#x27;t believe they&#x27;d claim something like that. But yes indeed, from the article:<p>&gt; Anthropic&#x27;s gross margins are above 80% before accounting for revenue shared with distribution partners, including Amazon (AMZN.O), opens new tab, and the cost of training its model,<p>Is this how all AI companies calculate if they&#x27;re profitable or not, by removing the highest costs? What a circus.
      • _diyar38 minutes ago
        &gt; Anthropic has told shareholders that its adjusted operating income will be positive for a second straight quarter, the Financial Times reported on Sunday, citing multiple people with knowledge of the matter.<p>Note this claim is about „operating profit“, which commonly is the revenue - operating expenses (COGS, rent, payroll). This does not include RnD cost.<p>&gt;Anthropic&#x27;s gross margins are above 80% before accounting for revenue shared with distribution partners, including Amazon (AMZN.O), and the cost of training its model, the newspaper said.<p>Gross margin is typically (revenue - COGS) &#x2F; revenue. Thus, both statements above seem generally in line with commonly accepted accounting standards.
        • yread9 minutes ago
          If you are sharing revenue (royalties, licenses based on revenue, costs that scale directly with revenue) doesnt that count as cost of goods sold?
      • sandeepkd40 minutes ago
        The way I read it, they are convincing the investors that they can fool the larger population convincingly. At the end of the day the investor term is misnomer for big institutional investors, given that these people are managing other people money where they always make out a certain percentage of fees despite the outcome.
      • SlightlyLeftPad35 minutes ago
        I believe this is a new accounting method called “phantasmagorical accounting.”
    • aesthesia54 minutes ago
      &gt; The profitability of being able to serve a query wasn&#x27;t really under question<p>In one sense, yes, but I do see people question it regularly.
    • TSiege40 minutes ago
      I think part of the big push to &quot;slow down AI development&quot; is to add some sort of regulatory pressure that will give them sort cover to train less models and slow their burn rates
    • awongh40 minutes ago
      I didn&#x27;t realize how much money Uber makes from ads.... Why does every business devolve into an ad platform?
      • agentultra20 minutes ago
        Might be because most hit their maximum growth but need to keep growing indefinitely or risk becoming a “mature” company?
    • JumpCrisscross50 minutes ago
      &gt; <i>idea of removing model training from your costs is a little wild tbh</i><p>It&#x27;s one of several metrics and tries to estimate steady-state profitability. It&#x27;s the only one being leaked because it&#x27;s the most sensational one. But don&#x27;t assume cash-flow profitability is negative just because you don&#x27;t know it.
    • throwawaysleep50 minutes ago
      &gt; The profitability of being able to serve a query wasn&#x27;t really under question (nor is the margin expected to be anything less than 80%+) I think.<p>HN had long debates about whether AI inference could even be affordable from a compute perspective.
    • sphinxterai1 hour ago
      Well then you haven&#x27;t listened to Ed Zitron or any of the other AI bubble doomers. His contention is that its worthless and they lose money on every query.
      • embedding-shape58 minutes ago
        But isn&#x27;t he taking all the costs into account, that created the experience? Rather than just literally the inference&#x2F;serving infrastructure? Bananas way of calculating things if so, doesn&#x27;t match reality at all.
      • JumpCrisscross51 minutes ago
        Zitron is worthless–lying about numbers and not correcting the record when you&#x27;re called out means you aren&#x27;t trustworthy. Worse than that if you directionally agree with him, which I do.
        • rowanG07741 minutes ago
          He makes his money and fame by spewing bulshit as many grifters before him.<p>Which unfortunately probably hides the real truth. That large labs do have potential problems with long term profitability.
      • freejazz40 minutes ago
        isn&#x27;t that when the training costs are actually accounted for?
  • nemomarx1 hour ago
    &quot;profitable without COGS&quot; doesn&#x27;t actually mean anything at all does it?
    • JumpCrisscross49 minutes ago
      Profitable without Capex, <i>i.e.</i> gross margin. Inference is included in COGS. Capex (training) is not. That&#x27;s reasonable.
      • icedchai27 minutes ago
        Is it really capex though? New models are being constantly trained, released at least quarterly, while old ones become obsolete. Training costs vary, but never disappear.
      • cmdli41 minutes ago
        This also includes revenue sharing with distribution partners, which should probably be in COGS.
      • freejazz38 minutes ago
        Isn&#x27;t training necessary for the end-product? How is it not a cost to generate the output if you can&#x27;t generate the output without having done the training? Seems more like saying that a car is profitable product when you don&#x27;t have to account for the steel that its made from. Seems completely disingenuous.
        • seizethecheese14 minutes ago
          In what case is capex (for any company) not required for the end product? Are you arguing it’s disingenuous to begin with as a concept?
        • icedchai24 minutes ago
          It&#x27;s an accounting trick.
    • mrweasel49 minutes ago
      Is that basically: We&#x27;d be making money, if we didn&#x27;t have to build the product?
  • cdrnsf57 minutes ago
    I imagine there are many other businesses that would be profitable if they excluded all of their largest costs from their reporting.
    • throwaway_9528342 minutes ago
      Turns out building houses is super profitable once you remove the cost of building materials, labor, and land.
  • burgerboii41 minutes ago
    I believe that their desire to slow down AI development is just for profits.<p>Active competition requires constant reinvestment and does not allow them to milk their trained models long enough (except poor Haiku maybe).
  • ChrisBland1 hour ago
    GAPP or ACSOI? Adjusted Consolidated Segment Operating Income from the groupon days....
  • maherbeg47 minutes ago
    Is this community adjusted EBITDA?
  • jimmydoe56 minutes ago
    very convoluted, number game but probably works for casual investors who just want to put money in something.
  • winfredJa40 minutes ago
    surprised by 80% margin that doesn&#x27;t include training cost.
  • jqpabc1231 hour ago
    Yes, but --- using something they call &quot;adjusted operating income&quot;.<p>This is reportedly a sort of &quot;Enron&quot; accounting which excludes some really big expenses like revenue sharing, the cost of model training and hardware deploymments which are kept off the corporate balance sheet using &quot;special finance vehicles&quot;.<p><a href="https:&#x2F;&#x2F;www.msn.com&#x2F;en-us&#x2F;technology&#x2F;artificial-intelligence&#x2F;the-very-big-caveat-to-the-report-that-anthropic-is-profitable-for-a-second-straight-quarter&#x2F;ar-AA2cbhG0" rel="nofollow">https:&#x2F;&#x2F;www.msn.com&#x2F;en-us&#x2F;technology&#x2F;artificial-intelligence...</a>
    • sigmar1 hour ago
      &gt;This is reportedly a sort of &quot;Enron&quot; accounting which excludes some really big expenses like revenue sharing, the cost of model training and hardware deploymments<p>source? this seems false. reportedly the adjusted profitability includes inference and amortized training costs
      • jqpabc1231 hour ago
        <i>source?</i><p>Listed at the end of my post.<p><i>this seems false.</i><p>Source showing this in accordance with GAAP (Generally Acceptable Accounting Practices)?
        • sigmar56 minutes ago
          that says only that the gross margin calculation excludes profit sharing and training. You should read it more carefully<p>edit: def not gaap profitable or they would have said that to investors. and their stock-based comp is surely astronomically high on paper.
    • mixdup1 hour ago
      It&#x27;s really easy to be profitable when you exclude all of your expenses
  • throwaway858251 hour ago
    GAAP or non GAAP profitable?
    • datadrivenangel1 hour ago
      non GAAP. If they were GAAP profitable they either wouldn&#x27;t be sharing or would be IPOing already.
      • JumpCrisscross49 minutes ago
        &gt; <i>If they were GAAP profitable they either wouldn&#x27;t be sharing or would be IPOing already</i><p>They&#x27;d be in their quiet period...
  • dr_faustus35 minutes ago
    Pathetic! At my company, we have a 100% margin before accounting for cost!
  • altmanaltman1 hour ago
    &gt; Anthropic&#x27;s gross margins are above 80% before accounting for revenue shared with distribution partners, including Amazon (AMZN.O), opens new tab, and the cost of training its model, the newspaper said.<p>Yes the company known for famously training 1 model
    • layer81 hour ago
      The “opens new tab” sometimes drives me insane when using TTS to listen to Reuters articles. I’m assuming they are using the wrong CSS&#x2F;markup for the purpose.
    • sobiolite1 hour ago
      Well fortunately now we&#x27;re gonna be Pacing the Frontier.
      • unreal61 hour ago
        I cynically wonder how much of this &quot;pacing&quot; is a mutual decision to reduce training expenditure costs as OpenAI and Anthropic plan to IPO.
        • thewebguyd1 hour ago
          Or they&#x27;ve hit diminishing returns that will collapse their valuation, so saying &quot;this is a threat to humanity&quot; sounds better than &quot;This is about as good as the tech is going to be for a long time&quot; to investors.
        • stackskipton1 hour ago
          My guess is a ton. According to <a href="https:&#x2F;&#x2F;www.thebignewsletter.com&#x2F;p&#x2F;monopoly-round-up-just-stop-the-anthropic" rel="nofollow">https:&#x2F;&#x2F;www.thebignewsletter.com&#x2F;p&#x2F;monopoly-round-up-just-st...</a>, they are asking for exception to anti trust as this type of collaboration would normally be massive anti trust violation.
  • GiorgioG1 hour ago
    This is Enron-level fraud. What would Ford&#x2F;GM&#x2F;Toyota&#x27;s gross margins be without the cost of manufacturing vehicles?
    • maerF0x053 minutes ago
      That&#x27;s a bit hyperbolic. It&#x27;s closer to using EBITDA as your &quot;earnings&quot; and bucketing model costs in a rapid depreciation model (which is fair, I&#x27;d assume a model is good for more than just 1 year...
      • JumpCrisscross48 minutes ago
        It&#x27;s gross margin. Not including training in gross margin is perfectly reasonable.
        • freejazz36 minutes ago
          Unfortunately, just stating that something is reasonable does not make it true.
        • lovich41 minutes ago
          The models don’t exist without training. I don’t see how excluding the training cost from the thing they are selling(inference) is perfectly reasonable and not just an accounting trick.<p>If I am building a widget and have a widget factory that cost money to build and operate, is it reasonable to only use the cost of shipping my widgets to my buyer as the costs for my gross margin?
          • maerF0x030 minutes ago
            Generally speaking, IDK about Anthropic specifically, they don&#x27;t train purely from scratch though. A good chunk of the setup is reused previous models can serve as a basis for the next model. Plus there are methods that also use the previous model as a warm start.<p>Just taking a wild guess, but I&#x27;d assume the .5 releases are built on the previous and the Majors (3, 4, 5) are more extensive retrains?
            • lovich9 minutes ago
              That’s maybe a reasonable argument but it’s not one Anthropic is making.<p>If the article is to be believed they aren’t including their training costs.<p>Also lol at reporting it as above 80% without accounting for the revenue sharing as well.<p>I bet anyone’s finances look great if you just start ignoring all the money they owe.
              • maerF0x06 minutes ago
                &gt; I bet anyone’s finances look great if you just start ignoring all the money they owe.<p>Lol and truth.
    • judge202051 minutes ago
      More similar to if they took out the cost to design the cars and all the many levels of corporate overhead they have in the form of middle management.
  • exabrial1 hour ago
    [flagged]