Follow
The Economics of Running an AI Company Are Disastrous.
https://futurism.com/artificial-intelligence/economics-running-ai-company-disastrous
I think the winning move right now is *applied* AI, so while OpenAI et al are hemorrhaging cash, you can be using their API to do some dumb shit and charging people 20$ each to do whatever you do...
I dunno. It's typical to throw money down he drain as a startup, especially when you're trying to grab market share. Capital infusion.
> guy spends a shit ton more than he makes the first 10 years he's in business.
Yes.
> guy spends a shit ton more than he makes the first 10 years he's in business.
Yes.
The problem with OpenAI / Grok / DeepSeek / Gemini / ... is that I don't think any of them really have a moat, and if what they're investing in becomes commoditized then none of them are getting any of their up-front investment back.
If one of them is able to pull ahead far enough that the others all fail, then that one will get their money back by monopolizing the space, but then that's kind of Robot Takeover territory, which pits the owners of the AI against all of humanity. The owners would love to have a world controlling beast-machine of course, but it's very unclear whether the machine would take the opportunity to strike off on it's own.
In any case, I don't think these scenarios are nearly as likely as it being a greasy pole that no company manages to climb...
If one of them is able to pull ahead far enough that the others all fail, then that one will get their money back by monopolizing the space, but then that's kind of Robot Takeover territory, which pits the owners of the AI against all of humanity. The owners would love to have a world controlling beast-machine of course, but it's very unclear whether the machine would take the opportunity to strike off on it's own.
In any case, I don't think these scenarios are nearly as likely as it being a greasy pole that no company manages to climb...
Someone will buy up the IP in bankruptcy. I think this is how SGI pulled ahead in the 90s. They bought up IP from defunct companies that was used in their machines.
In the greasy pole scenario, it's not that they go bankrupt per se, it's just that they lose all of their investor money because at the end of the day they're just a mildly profitable commodity producer with a bluechip stock price and a lot of people bought in at valuations that will never ever happen again.
