It is very fascinating to see a kind of modern day version of the tulip craze play out in real time. What kind of revenue must OpenAI stabilize at to defend such a valuation? It’s not like they have any first mover advantage or other kinds of moats vs Anthropic & co.
OpenAI is currently the best in some major criterias, but not all. Gemini has much larger context window, for example.
And behemoths like Google/Meta/Apple can remain irrational with their AI spending longer than OpenAI can remain solvent. Unless Microsoft steps in, which they already did, kind of.
Eh, the tech is legit. But I guess the problem is to figure out how to make money in the long term. I can't really think of any similar products where something costs ridiculous amounts of money to develop, and then the product has a effective shelf-life of what...6-12 months? Before every other competitor has caught up.
I guess the big problem with AI/ML of the LLM type is that the data and compute is the gold, or fuel if you will. How does one commercialize a product in the long term, when each iteration becomes magnitudes more expensive than the previous, and the previous products become almost worthless within 12 months.
Oh yeah, keep growing your number of subscribers. But that also has a natural limit.
Maybe at some point, these AI models will be so integrated and omnipresent in everyone's daily like that some kind of oligopoly or monopoly arises, and they can leverage that.
>I can't really think of any similar products where something costs ridiculous amounts of money to develop, and then the product has a effective shelf-life of what...6-12 months?
To make a 10x return, OpenAI would need to be worth 1.6T.
So investors are thinking that there is a about ~10% chance of AI growing to about ~10% the size of the service industry (10% of service industries 16T/year is 1.6T, 10% chance => 160B).
While I'm not sure I agree, this math does actually make more sense than you'd think given the sticker-shock of 160B for a young company.
What I'm more curious about is the (as far as I know) non-public guesses as to which industries openai believes it will take bites out of. Software? Service? An entirely new industry space? Assuming its going to just grow productivity by itself, and not crush other industries, the valuation seems off. For $160B, the expectation is more likely "this will result in no longer needing hundreds of thousands of employees" - but are we replacing their salaries wholesale with OpenAI contracts? Doubtful.
On the other side of things, imagine OAI purchased something critical to their success from a near-monopoly producer. If they suddenly are devouring massive amounts of the economy, it will rationally attempt to increase its margins.
Not to say that there aren't ways around it. But you'd also expect crazy valuations to be going up the supply chain. And TSMC is wildly undervalued if you believe in OAI's ideal scenario. Market is inconsistent here.
And they are expecting us to pay more than the most popular streaming sites combined as monthly fee for ChatGPT? I won't. $40/month is too much imho
Looking at my API usages costs I am not even close to the current monthly cost
One thing that still surprises me is how quickly Satya and Microsoft have washed their hands clean of the OpenAI fiasco. They’re insiders. They must have known who Sam is, what was happening to the team, etc. Satya used his position to make a threat that he would destroy OpenAI, and in the process left Ilya and others out to dry in favor of Sam. Any no one seems to put any blame on Satya and Microsoft. I am not an insider, so this is all just my perspective, but it looks curious in the least.
It is very fascinating to see a kind of modern day version of the tulip craze play out in real time. What kind of revenue must OpenAI stabilize at to defend such a valuation? It’s not like they have any first mover advantage or other kinds of moats vs Anthropic & co.
Agreed.
OpenAI is currently the best in some major criterias, but not all. Gemini has much larger context window, for example.
And behemoths like Google/Meta/Apple can remain irrational with their AI spending longer than OpenAI can remain solvent. Unless Microsoft steps in, which they already did, kind of.
Eh, the tech is legit. But I guess the problem is to figure out how to make money in the long term. I can't really think of any similar products where something costs ridiculous amounts of money to develop, and then the product has a effective shelf-life of what...6-12 months? Before every other competitor has caught up.
I guess the big problem with AI/ML of the LLM type is that the data and compute is the gold, or fuel if you will. How does one commercialize a product in the long term, when each iteration becomes magnitudes more expensive than the previous, and the previous products become almost worthless within 12 months.
Oh yeah, keep growing your number of subscribers. But that also has a natural limit.
Maybe at some point, these AI models will be so integrated and omnipresent in everyone's daily like that some kind of oligopoly or monopoly arises, and they can leverage that.
>I can't really think of any similar products where something costs ridiculous amounts of money to develop, and then the product has a effective shelf-life of what...6-12 months?
This is most Hollywood productions in a nutshell
I was going to say AAA video games.
They are selling "AGI". That is, the dream they will get there first, I think?
Even their core team is no longer there. So investors are not the smartest for pushing it that high.
Some are betting on selling on the next round, not for long term investments.
Probably they have more things in store than they offer publicly?
$10bn annual profit within a couple years.
Doesn't seem too outrageous
Says who? They are currently estimated to lose $5B annually on $3.7B revenue.
Imagine for a moment that you are following the hype instead of the facts. Now you understand why investors dump their money on these kinds of things.
A 10% chance of making $100 costs $10.
To make a 10x return, OpenAI would need to be worth 1.6T.
So investors are thinking that there is a about ~10% chance of AI growing to about ~10% the size of the service industry (10% of service industries 16T/year is 1.6T, 10% chance => 160B).
While I'm not sure I agree, this math does actually make more sense than you'd think given the sticker-shock of 160B for a young company.
What I'm more curious about is the (as far as I know) non-public guesses as to which industries openai believes it will take bites out of. Software? Service? An entirely new industry space? Assuming its going to just grow productivity by itself, and not crush other industries, the valuation seems off. For $160B, the expectation is more likely "this will result in no longer needing hundreds of thousands of employees" - but are we replacing their salaries wholesale with OpenAI contracts? Doubtful.
On the other side of things, imagine OAI purchased something critical to their success from a near-monopoly producer. If they suddenly are devouring massive amounts of the economy, it will rationally attempt to increase its margins.
Not to say that there aren't ways around it. But you'd also expect crazy valuations to be going up the supply chain. And TSMC is wildly undervalued if you believe in OAI's ideal scenario. Market is inconsistent here.
Or there's a 0.1% chance it becomes a $160T company if they're the first to develop superintelligence
Is far faaaar lower than a 0.1% chance.
You can’t predict breakthroughs and key members of the OG engineering team left.
OpenAI definitely has some value. I work as a DevOps engineer and I use it quite a few times a day.
It's replacing a lot of my Google searches. I also see OpenAI used by people around me in the office.
So they definitely have first mover advantage and if they can keep up with competitors they can take a big chunk of the market.
Some value, of course. 157 billion… that is harder to justify.
Despite this, there hasn't been much movement in MSFT recently, despite Microsoft owning a big chunk of OpenAI.
https://www.wsj.com/tech/ai/openai-nearly-doubles-valuation-...
[dupe] Discussion: https://news.ycombinator.com/item?id=41722742
And they are expecting us to pay more than the most popular streaming sites combined as monthly fee for ChatGPT? I won't. $40/month is too much imho Looking at my API usages costs I am not even close to the current monthly cost
One thing that still surprises me is how quickly Satya and Microsoft have washed their hands clean of the OpenAI fiasco. They’re insiders. They must have known who Sam is, what was happening to the team, etc. Satya used his position to make a threat that he would destroy OpenAI, and in the process left Ilya and others out to dry in favor of Sam. Any no one seems to put any blame on Satya and Microsoft. I am not an insider, so this is all just my perspective, but it looks curious in the least.