4 ms·

So what is the end game here? IPO on the stock market for over $500 billion or stay private like Stripe, SpaceX or Bytedance? Nobody can buy them without bankr

So what is the end game here? IPO on the stock market for over $500 billion or stay private like Stripe, SpaceX or Bytedance? Nobody can buy them without bankrupting themselves many times over, unless the big companies buy them (very unlikely) Or is OpenAI just going to keep raising and raising more and more? Is OpenAI even profitable?

htrp 2y ago
Allegedly yes (excluding all of the existing AI researcher costs).
- mechagodzilla 2y ago
  Wait - they're 'profitable, excluding all of their employee costs'? Does that just mean they're charging more than the cost of their electricity to generate the tokens?
  - mikeyouse 2y ago
    Presumably this would mean also earning more than the amortized cost of all the hardware they’ve purchased too which is a much higher bar than just electricity costs.
- edgyquant 2y ago
  If they were why would they raise? They’ve already achieved scale.
  - asadm 2y ago
    they need upfront $$ to fund GPT6 and beyond.
blackoil 2y ago
IPO is one option. Merger is another. Companies don't need to buy them in cash but in stocks.
Bjorkbat 2y ago
https://www.theinformation.com/articles/why-openai-could-lose-5-billion-this-year https://www.theinformation.com/articles/why-openai-could-los... If you take this at face-value, they probably aren't profitable (to the tune of a $5 billion loss). I don't like idle speculation, but having actually read this their analysis seems pretty sound. Their numbers on training costs and hosting actually came from an anonymous insider. Impossible to say how reliable this anonymous insider is, but planting false, exaggerated numbers to outwit competitors seems a bit too much like a hare-brained scheme, hence why I opt to go with the numbers. If anything, companies tend to downplay costs when lying. The one thing that caught me as a little bit crazy was spending $1.5 billion on employees. They have ~1500 employees, so that's $1 million per head. That seems insane, but then again, AI researchers are a valuable commodity these days. Even if the cost of employees was $0, you'd still be looking at estimated losses in the billions.
lijok 2y ago
Kind of a moot question. They’re not building rockets or payment rails.
ilaksh 2y ago
Sometimes I think some companies are playing a different game. Where profit was never really a concern. The goal is just to keep showing growth and raising. You just have to cash out before the music stops or The Singularity happens. One aspect of The Singularity is the rapid acceleration into an unpredictable point of no return. Maybe the financial model mirrors this?
- quonn 2y ago
  > One aspect of The Singularity is the rapid acceleration into an unpredictable point of no return. One aspect of reality is that governments actually like to govern predictably and those with sufficient military power make sure they can.
nwoli 2y ago
Regulatory capture, monopoly prices when the earth is dependent on them and no new companies can form
paxys 2y ago
> Nobody can buy them without bankrupting themselves many times over, unless the big companies buy them (very unlikely) There are lots of M&A structures. A company can in fact buy another that is bigger than itself. Plus Microsoft will easily be able to afford a $100B acquisition with some creative financing, or maybe even outright cash.
- ThrowawayB7 2y ago
  Heaven forfend. It would be Nadella's equivalent of Ballmer's attempt to buy Yahoo.

htrp 2y ago

Allegedly yes (excluding all of the existing AI researcher costs).

mechagodzilla 2y ago
Wait - they're 'profitable, excluding all of their employee costs'? Does that just mean they're charging more than the cost of their electricity to generate the tokens?
- mikeyouse 2y ago
  Presumably this would mean also earning more than the amortized cost of all the hardware they’ve purchased too which is a much higher bar than just electricity costs.
edgyquant 2y ago
If they were why would they raise? They’ve already achieved scale.
- asadm 2y ago
  they need upfront $$ to fund GPT6 and beyond.

blackoil 2y ago

IPO is one option. Merger is another. Companies don't need to buy them in cash but in stocks.

Bjorkbat 2y ago

https://www.theinformation.com/articles/why-openai-could-lose-5-billion-this-year https://www.theinformation.com/articles/why-openai-could-los... If you take this at face-value, they probably aren't profitable (to the tune of a $5 billion loss). I don't like idle speculation, but having actually read this their analysis seems pretty sound. Their numbers on training costs and hosting actually came from an anonymous insider. Impossible to say how reliable this anonymous insider is, but planting false, exaggerated numbers to outwit competitors seems a bit too much like a hare-brained scheme, hence why I opt to go with the numbers. If anything, companies tend to downplay costs when lying. The one thing that caught me as a little bit crazy was spending $1.5 billion on employees. They have ~1500 employees, so that's $1 million per head. That seems insane, but then again, AI researchers are a valuable commodity these days. Even if the cost of employees was $0, you'd still be looking at estimated losses in the billions.

lijok 2y ago

Kind of a moot question. They’re not building rockets or payment rails.

ilaksh 2y ago

Sometimes I think some companies are playing a different game. Where profit was never really a concern. The goal is just to keep showing growth and raising. You just have to cash out before the music stops or The Singularity happens. One aspect of The Singularity is the rapid acceleration into an unpredictable point of no return. Maybe the financial model mirrors this?

quonn 2y ago
> One aspect of The Singularity is the rapid acceleration into an unpredictable point of no return. One aspect of reality is that governments actually like to govern predictably and those with sufficient military power make sure they can.

nwoli 2y ago

Regulatory capture, monopoly prices when the earth is dependent on them and no new companies can form

paxys 2y ago

> Nobody can buy them without bankrupting themselves many times over, unless the big companies buy them (very unlikely) There are lots of M&A structures. A company can in fact buy another that is bigger than itself. Plus Microsoft will easily be able to afford a $100B acquisition with some creative financing, or maybe even outright cash.

ThrowawayB7 2y ago
Heaven forfend. It would be Nadella's equivalent of Ballmer's attempt to buy Yahoo.