Timeline
Post
Remote status
Fetching contextโฆ
This post replies to something we haven't fetched yet. We'll try to pull in the missing thread context in the background.
Goal๐ฏ๐ป๐ฎ๐ช๐ด๐ฎ๐ป 
Nvidia is the real winner here, you are correct. Until it all collapses and they have a ton of stock they can't move.
Replies
22
I can't wait to pick up an H200 for $2000.
Me at the data center Ewaste pile when it all comes tumbling down.
Skunkle: Mediocrity SpeedrunnerI read eWaste as Ee-wahs-tay. Apparently, all cylinders are NOT firing today.
IT'S NOT MY FAULT! HE CAPITALIZED THE E AND NOT THE W!
More specifically, until they can't pay back the money they borrowed to increase base production to meet temporary demand.
@Goalkeeper @caekislove @cjd Inference demand is gonna feed NVDIA for a century. If you want to run open source, you still need GPUs with a ton of VRAM. Whenever someone wants AI anywhere (except China), NVDIA will make money.
I think what this reveals is labs are on the bottom of the infra stack, and treat themselves like the top of the value chain instead of the rock-solid foundation. Right now, theyโre arguing that they need to be given a de facto duopoly on development until they can get big enough that their AGI does everything and then they can pay back their investors. In reality, Microsoft is fine using OA or Kimi for copilot, it barely makes a difference to them. Kimi is way cheaper btw, because they can run it on their own hardware without paying an API fee to OA.
I think what this reveals is labs are on the bottom of the infra stack, and treat themselves like the top of the value chain instead of the rock-solid foundation. Right now, theyโre arguing that they need to be given a de facto duopoly on development until they can get big enough that their AGI does everything and then they can pay back their investors. In reality, Microsoft is fine using OA or Kimi for copilot, it barely makes a difference to them. Kimi is way cheaper btw, because they can run it on their own hardware without paying an API fee to OA.
And the little side-quest is the fact that you can take questions and answers from one AI (which they OBVIOUSLY have to give you) and use it to train another one - and since they're all starting from roughly the same corpus, filtering out evolutions that don't produce the same result is going to end up giving you roughly the same AI they did - so... It's the only SaaS that is absolutely trivial to pirate...
All the hard work is already done. Definitely something they didn't account for.
AI industry got SO MUCH of this going on...
It's like a giant bucket of crabs.
They aren't going to be selling surplus AI GPU stock at a profit. They are going to lose their ass on this deal when the bubble pops. It won't kill them but it is going to hurt.
@Goalkeeper @caekislove @cjd what bubble is the question. If AI is better but less exclusive, profit moves to the infrastructure layer. If youโre serving inference, you make more money from commoditization. NVDIAโs the nigga selling chips. To them, this just means more demand for infrastructure. The pressure is on the labs because theyโre struggling to monetize as the industry matures.
Too much supply + not as much demand = prices crash. They have already spent billions of dollars on R&D and manufacturing because of what the big AI firms promised them. The only question is how much demand is going to fall not if it's going to fall.
I have a feeling a lot of this investment money is just on account of ChatGPT social engineering the fuck out of high net worths...
@Goalkeeper @caekislove @cjd Theyโll still sell inference. Acquisition of labs by hyperscalers is a serious possibility. NVDIA makes serving inference more efficient. Theyโre not going anywhere.
I never said it was going to erase them from the market. I said it was going to hurt.
Until promises made are promises kept it's a bubble. It's all speculative at the moment. They are running out of runway. It's either going to take off or slam into the wall sooner rather than later. Nvidia isn't going to wait forever. Investors are getting antsy.
@Goalkeeper @caekislove @DarkGura @cjd If the labs really hit a critical point, do you know what'll happen? Anthropic gave Amazon a huge amount of equity in exchange for compute. OpenAI made a similar deal with Microsoft. Amazon and Microsoft ain't running out of money anytime soon. They're already making money off inference, which has crazy margins btw.
Notice that Microsoft is experimenting with adding Kimi K3 to Azure for copilot? Moonshot is OA's competition why would they do that? It gives Microsoft leverage over OA, and it puts pressure on their business. Maybe they're introducing competition to their subsidiary so they can acquire them outright. Does Microsoft want to own equity in an independent lab that's trying to summon the god-machine, or do they want the product?
If there's a little blip in the stock market, does that matter in the long term?
Notice that Microsoft is experimenting with adding Kimi K3 to Azure for copilot? Moonshot is OA's competition why would they do that? It gives Microsoft leverage over OA, and it puts pressure on their business. Maybe they're introducing competition to their subsidiary so they can acquire them outright. Does Microsoft want to own equity in an independent lab that's trying to summon the god-machine, or do they want the product?
If there's a little blip in the stock market, does that matter in the long term?
@DarkGura@detroitriotcity.com @caekislove@caekis.love @Griffith@clubcyberia.co @Goalkeeper@nicecrew.digital @cjd@pkteerium.xyz "this time it's different" lmao
@Griffith @caekislove @Goalkeeper @cjd nvidia, intel, amd and anyone else that can make those chipperinos. Vulkan inference is growing
@Goalkeeper @caekislove @cjd Not just stock, their entire business model has moved to AI and their personal computing side has become less than 10% of their revenue.
If AI falls though they will need a massive restructuring that will involve gutting at least half the company, I can't think of a single example of when a company fell that hard and didn't collapse.
If AI falls though they will need a massive restructuring that will involve gutting at least half the company, I can't think of a single example of when a company fell that hard and didn't collapse.