Close Menu
    Wall Street

    The AI Trade is Dead, OpenAI and Anthropic undershot the moon, pressure private equity

    For those who don't work in the tech industry, the hype for cutting edge models is already completely dead. We got a taste of what it was capable of with Fable, but it turns out it's mostly just good at making stuff that already exists like Minecraft and Tinder clones. It was kinda neat for the 2 weeks it was freely available on plans, but ultimately none of these businesses are willing to pay ludicris token prices just to use a slightly better model. Companies have already switched back to subscription models, they'd be stupid not to, it's free compute. Even if they were to remove subscriptions entirely, it wouldn't help as people can just switch to open source models instead. It's a no win situation for Anthropic and OpenAI.

    This means that the only AI sector that was going for mass compute, LLMs, will definitely not be able to ever pay back all that compute they spent. That's 2.5 trillion down the drain.

    This is where you come in đŸłâ€đŸŒˆđŸ»'s. Private equity, Softbank, Amazon, Microsoft have to hold this massive bag of debt they spent on these worthless models and they want you to buy them from them. Expect the market to trade sideways or slightly down until this crashes, they are trying to slowly offload shares without crashing the market. Their goal is to hold out until the Anthropic IPO so they can short their own shares to cover. They want to offload their shares on you at a 2 trillion evaluation and then rug pull your 401k.

    Positions:

    Late 2027 SNDK, NVDA, PLTR, TQQQ, and ORCL puts.

    Other good options are NBIS, MU, and CRWV

    The AI Trade is Dead, OpenAI and Anthropic undershot the moon, pressure private equity
    byu/sfw_supdood inwallstreetbets



    Posted by sfw_supdood

    32 Comments

    1. RichSpecific4490 on

      You are right
.but don’t underestimate the money printer. Best of luck, hoping you win. I have some skin in the game myself, let’s make some money.

    3. yeeyeehair16587 on

      This is the type of post where we keep prodding for screenshots and find out that op only has like 400 bucks in puts

    4. mneymaker on

      I was tald Compute=Revenue. And I need a lot of compute to break even

      Now stfu Michael, the leatherman will speak the Word of God

    6. Ble_h on

      >Ludicris tokens

      I didn’t know Luda was involved with AI. I think he infamously said “why you all up in my ear, talking a whole bunch of shit I ain’t trying to hear”

    14. lightning_Mcking29 on

      lol “AI trade is dead” while the hyperscalers keep buying GPUs and the labs keep raising.

    15. sfw_supdood on

      To add, SPACEX just tested out this strategy for Anthropic’s IPO. The early investors can’t sell their shares for like 9 months – 1 year, but they will short their shares right away at IPO, that’s what SPACEX did. Add in a rule where IPO retail investors can’t sell without a $50 fee and a staggered release, and that’s their plan to make you a bag holder.

      Don’t fall for it.

    16. Concept-Plastic on

      MU is making what $500m/day right now and newer reports say compute will be stressed further into 2028-2029

      Sounds bearish as f, im in bro, lets buy puts

    17. GetDeepSignal on

      Bro, definitely doesn’t work in tech industry if you don’t see the clear potential of these llms. And computer is gonna be used for next frontier models which will eventually help with physical AI like robots. Claude code is indeed doing amazing things by speeding up software, agree there’s a little bit of saturation, but companies like Mag 7 want to play long game, not option gamble lol.

    18. Correct_Mastodon1112 on

      Compute is gonna be needed regardless of the model. Passing on the hyperscalers and frontier start ups makes sense. Shorting compute, not so much.

    19. LoveNewton_Nibbler on

      Ill bank on having a runway of 2+ years of 🍊 doing everything he can to slant the game towards the tech giants. Nvda calls for me personally

    20. TwoDurans on

      Chasing the newest and best model costs far too much money and there isn’t any path to making that back. Some investors are getting spooked, others are doubling down and buying on dips. Remains to be seen who will be right.

    21. intelligentx5 on

      It’s dead? Zoom out lol. Every fucking post like this after a 1-2% drop just makes me laugh

    22. Lundhlol on

      Guy thinks AI is even close to a finished product, that or been reading too many Burry tweets.

    23. arter_dev on

      OP is right. I build agentic harnesses & workflows for a living. Unfortunately I have to live / eat / breath the intersection of LLMs & developers.

      Devs are not giving two shits about which model they’re using. Whichever is cheapest and governable in a harness. There’s no loyalty. It’s a total commodity. They bounce around from providers almost by the week. This is 100% a railroad situation. The only devs sticking with the same providers reliably are those employed by enterprise, which just sets us up for an identical structure of IBM mainframes back in the 80s. It will be a blip of rev in a sea of debt.

      I have 5 providers wired into my own harness, but at work I’m stuck with Claude.

      Positions: all VOO.

      edit: not to mention, we are also at the peak of capability the general public can be trusted with due to jailbreak risk (which is inherent in how these models work and will never go away).

    25. Bagholder147 on

      Antrohropic actually makes money. And Oracles backlog is higher than it’s market cap. It’s undervalued. They had to leverage the stock to buy WBD.

    28. DarkR124 on

      If AI is dead trading will be least of our worries. Going to take like 40-50% of the global economy with it lmao

    29. justawhiteguy5324 on

      There was a guy in like 2022 or something who said “NVDA” can’t go any higher and put 50k in puts for like 6 months out. Then watched NVDA 4x that year
 good luck big dawg

    31. Ogrelander on

      No doubt AI markets can hold value for a while yet, but ultimately they’re locked in a battle with China’s cheaper, state backed alternatives.

      The quality and cost comparison will attract a lot of users and revenue away from US AI products. They will need to hold their prices artificially low or bite the bullet and charge more, and that’s were valuation will decline imo. 

      Tesla went big on hype and vibes, and have a rabid, loyal fan base despite not really being a car company anymore, then you see how many BYDs are on the streets these days?

    Leave A Reply