Oil is a commodity, which wouldn't bode that well for any current compute provider long term.
Nvidia's success in particular seemed to spawn a lot of new chip designs. Quite a few startups, but also AMD and Intel are some degree of serious with their training/inference offerings for datacenter.
Any kind of signal in current buying or selling behaviour on either the stock or its derivatives that might indicate the price is going to plateau or fall, such as (for example) a change in the market for shorts.. though this seems obvious and I don’t mean to be insulting, so I might be misunderstanding your question? If so please let me know.
If there are no signals which people can use as evidence in their models predicting market direction, what are people buying or selling on? Why would anyone read a quarterly report or have an investment thesis if it’s just a perfectly efficient market of spherical cows on a random walk, and analysis is a fools game as no evidence of where a company might be headed could possibly exist? It’s a very postmodern way of looking at things, I suppose.
Anyway I was just curious if the person had any facts on which they were basing their assessment that they could share. Seems not, which is fine. Opinions are allowed too, but I was curious since they stated it so matter if factly.
If it's news on a major website, it's priced in. So yes, we know it is priced in.
Obviously, it wasn't news 5 years ago, so no it wasn't then.
There is a gray area which is where professional investors who have an informational advantage make their money. But by the time something is mainstream news, that gray area is long gone.
CEG jumping 20% on the MSFT deal means it isn’t all priced in imo. Everyone has known there are rising energy needs and had bid up CEG. Then the Three Mile Island restart news hit and apparently it wasn’t already priced in
Nobody (well the market) expected Nvidia's revenue to start growing exponentially, though. Now it's forecasted to grow by at least 20% per year for the next 3 years. So presumably everything above that is not priced in.
Actually people did expect that, which is why the market does not accurately "price things in." People working in ML, Cuda, ~GPT2 were pretty confident in the eventual impact of AI and knew of Nvidia's monopoly on it.
There's a sort of midwit meme here where the naive take "oh this company is going to sell more next year, i should by stock at whatever price", -> "no, everything's already priced in" -> "if everything was priced in, you'd see a historic price graph as a straight line with a slope of interest rates"
> You don’t know if it’s priced in. Was it priced in 5 years ago?
It wasn't priced in 5 years ago, that is the point today it is priced it, people expected this. You should have bought when Microsoft invested billions in them 2019, today more investments like this are expected once these investments stop NVIDIA stock will drop since they were expected.
So saying now is the time to invest is dead wrong, NVIDIA is currently priced as if these investments will continue to ramp up so these things are already priced in. Only idiots thinks "positive event -> I must buy stocks!", no if the positive event was expected then the stocks wont move, if a positive event was expected but didn't happen then stocks go down, that is the situation we are in currently with NVIDIA.
How do you know it was priced in? You look at what people say they expect, and this is what people say they expected, hence we know it was priced in.
What if I were to tell you Nvidia's monopoly and the implications of AI were obvious many years ago? Many people on HN bought large amounts of Nvidia stock a while ago. Karpathy's "Software 2.0" was widely read here predating much of Nvidia's market cap increase:
It was "obvious" C3.AI was not a real AI company worth much. The market still "priced in" at $18 billion for months: https://finance.yahoo.com/quote/AI/
My hot take: the market is good at pricing in possibilities, but less so at pricing in certainty (or lack thereof).
When AMD launched the Zen architecture, everyone said they were on the up-and-up, and the market gave them a boost. The company's shares only started to climb much higher later on, when there were irrefutable revenue numbers. There was plenty of time to buy and make money. Someone with the domain knowledge like us can definitely have an edge in the market because we understand the product well enough to formulate better odds, then we can profit off the difference between our odds and the market's odds.
I legitimately get their names mixed up, and have had a hard time in occasion wondering why open ai would be controlled by someone who seems to really like doing fraud.
The bigger the raise, the more news and hype, the easier it will be to offload the stock to retail investors. Theres definitely a sweet spot here where I assume the large funds will want retail investors to start picking up the losses instead of them. I don't see this as a long path till IPO like Stripe or SpaceX but I could be wrong
The losses will just get bigger and bigger, and eventually there will be an appetite for retail investors to pick up the tab
The financial industrial complex will make money on transaction fees selling the stock to their clients, institutional or otherwise. They don't need it to actually succeed in the long-term - as it is the clients, people's money, on the line, e.g. incentives aren't aligned properly.
At this point I don't think they see it as an investment, but rather as securing a previous investment. They aren't looking for a return on the 6.5B, but rather as a deployment of strategic capital to secure their previous investment.
A large amount of AI funding seem FOMO driven. Companies aren't seeing large returns on AI stuff, they are afraid of being left behind if somebody finds some sort of use for it.
In what way is GPT-5 missing? They never announced a release date, did they? And we had the o1 release just a couple of weeks ago, so why wouldn't they take more time working on improving GPT-5?
Beating all other models in latest benchmarks by a wide margin with a completely new approach, while having the only properly working audio chat with transformer isn’t enough? Not to even mention SORA if actually comes out someday.
I’m not an OpenAI fanboy but come on, credit where it’s due, they have so far been one step ahead of the competition from the start on every front.
I do research in other ML topics but I do follow the qualitative leaps they are pulling off continuously so far.
I'm not sure how relevant that is at this stage. Not loo long ago I read that OpenAI had ~$2B revenue, but are still losing money.
At this point I'd say this is a faith investment that either LLM/GenAI use is going to explode, and OpenAI is going to get a significant share of that, or that OpenAI will achieve a much more powerful and valuable AGI.
It seems corporate evaluation/adoption of GenAI is still early stages, and not at all clear who the winners are going to be if there is a lot of revenue forthcoming. How does OpenAI complete with free models from Meta, or open source ones, or pricing strategies from giants like Microsoft, Google and Amazon?
Ok so with $2B revenue is a 75x sales multiplier, which means a 54% CAGR for 10 years. They estimates $150B/y revenue within 10 years. Google is at $300B/y revenue.
Eventually they'll be put in robots able to gather real-world data live, and they'll have a practically infinite amount of video/audio/tactile data to learn from.
Sure, but this scales rather poorly. How many robots do you expect, and don't you expect some resistance from the public, for privacy reasons at the least?
Looks like the real winner here is nVidia, because they will be getting their investment money back as payment for GPUs using money invested by MSFT and APPL. Or is that view too simplistic?
Inference and training costs with $0 free AI models such as Llama which is good enough for AI solutions are all factors that are actually impacting OpenAI. What is not mentioned is that all these costs are quietly suffocating them whilst the company continues to suffer from a significant talent drain.
Companies like OpenAI cannot afford to continue racing their prices down to $0 and will just continue the pyramid scheme of raising billions to avoid bankrupting themselves over the said costs.
These fund raising rounds of A,B,C,D....X,Y,Z can't go on forever as these costs continue to suffocate the company. If OpenAI fails, there will be a horrific AI crash with hundreds of billions of dollars wiped out.
At this point, AGI could mean anything, from taking the majority of people's jobs to enriching themselves and never achieving 'AGI' in the first place.
It is entirely possible that 'AGI' could be a massive lie and was all about capturing the wealth for a few people at the company.
I jump between ChatGPT, Claude 3.5 and Grok2 on a regular basis, and I'd say its not "worse" in the general sense. I find it much better on anything sentiment-analysis related or for shortcutting customer research, and image generation is really good.
This seems crazy. It's a truly innovative company and all but they don't have revenue to justify this nor a moat given how many plausible competitors sprung up so quickly. Elon Musk just reached into his own pocket and made Grok appear.
OpenAI is looking more and more like a Ponzi scheme ego play from sama. They'll find ways to offload the losses to retail investors and everything valuable will end up in Microsoft's pocket in the end.
By "reached into his own pocket," you mean he leveraged engineering resources from Twitter, and then once it became clear that Grok wasn't a competitive model, raised $6,000,000,000 in capital from VCs and sovereign wealth funds [1] to pay for the additional compute necessary?
> In finance, the greater fool theory suggests that one can sometimes make money through speculation on overvalued assets — items with a purchase price drastically exceeding the intrinsic value — if those assets can later be resold at an even higher price.
They just need one more set of people past them to buy it. I know lots of uninformed people that would jump on it thinking they are “getting in early”.
I , for one, thank the investors for subsidizing AI and making it affordable for us. I am not sure my measly US$20 per month will be enough to cover the cost of my usage but I am grateful for the fun I am having while the music plays.
Misleading OP.
This is not news from this week. Last week or even older.
https://www.bloomberg.com/news/articles/2024-09-19/openai-to...
Number of earlier stories on same "closing in" topic:
OpenAI Fundraising Set to Vault Startup's Value to $150B
https://news.ycombinator.com/item?id=41514954
And even earlier
OpenAI in Talks for Funding Round Valuing It Above $100B
https://news.ycombinator.com/item?id=41514954
Time to invest in nvidia and Microsoft stocks I guess.
Because let's face it, that's where the money is going.
You are already late.
If ASI is on the horizon & is bound by compute, compute effectively is the new oil.
Oil is a commodity, which wouldn't bode that well for any current compute provider long term.
Nvidia's success in particular seemed to spawn a lot of new chip designs. Quite a few startups, but also AMD and Intel are some degree of serious with their training/inference offerings for datacenter.
That "if" is doing a lot of heavy lifting.
Is there any evidence they have already peaked?
What kind of "evidence" of future stock market movement are you expecting?
Any kind of signal in current buying or selling behaviour on either the stock or its derivatives that might indicate the price is going to plateau or fall, such as (for example) a change in the market for shorts.. though this seems obvious and I don’t mean to be insulting, so I might be misunderstanding your question? If so please let me know.
There's no such thing. If there was, market participants would start selling.
If there are no signals which people can use as evidence in their models predicting market direction, what are people buying or selling on? Why would anyone read a quarterly report or have an investment thesis if it’s just a perfectly efficient market of spherical cows on a random walk, and analysis is a fools game as no evidence of where a company might be headed could possibly exist? It’s a very postmodern way of looking at things, I suppose.
Anyway I was just curious if the person had any facts on which they were basing their assessment that they could share. Seems not, which is fine. Opinions are allowed too, but I was curious since they stated it so matter if factly.
Well, Microsoft certainly is running into some challenges at this point: https://www.sciencefriday.com/segments/microsoft-open-three-...
If nothing else, physics starts to become a limiting factor.
I legitimately want to live in a world where fundamental physics is the only limiting factor to our work. That sounds like the future.
The fundamentals of physics can change.
i want to live in a world where super science is invented and we get sci fi tech
Are you serious?
Its already priced in, things are still moving as expected. They could go up higher in the future, but not from this information.
You don’t know if it’s priced in. Was it priced in 5 years ago?
If it's news on a major website, it's priced in. So yes, we know it is priced in.
Obviously, it wasn't news 5 years ago, so no it wasn't then.
There is a gray area which is where professional investors who have an informational advantage make their money. But by the time something is mainstream news, that gray area is long gone.
CEG jumping 20% on the MSFT deal means it isn’t all priced in imo. Everyone has known there are rising energy needs and had bid up CEG. Then the Three Mile Island restart news hit and apparently it wasn’t already priced in
Nobody (well the market) expected Nvidia's revenue to start growing exponentially, though. Now it's forecasted to grow by at least 20% per year for the next 3 years. So presumably everything above that is not priced in.
6.5 Billion won't move NVidia's needle at this point. You need exponentially growing revenue from AI firms to drive increased spend on compute.
Actually people did expect that, which is why the market does not accurately "price things in." People working in ML, Cuda, ~GPT2 were pretty confident in the eventual impact of AI and knew of Nvidia's monopoly on it.
There's a sort of midwit meme here where the naive take "oh this company is going to sell more next year, i should by stock at whatever price", -> "no, everything's already priced in" -> "if everything was priced in, you'd see a historic price graph as a straight line with a slope of interest rates"
> Actually people did expect that
Obviously some did and they made a lot of money. So what? The market as a whole didn't.
> There's a sort of midwit
If the company grows by as much as it is expected to grow barring any external factors then yes, your "meme" certainly makes sense.
> You don’t know if it’s priced in. Was it priced in 5 years ago?
It wasn't priced in 5 years ago, that is the point today it is priced it, people expected this. You should have bought when Microsoft invested billions in them 2019, today more investments like this are expected once these investments stop NVIDIA stock will drop since they were expected.
So saying now is the time to invest is dead wrong, NVIDIA is currently priced as if these investments will continue to ramp up so these things are already priced in. Only idiots thinks "positive event -> I must buy stocks!", no if the positive event was expected then the stocks wont move, if a positive event was expected but didn't happen then stocks go down, that is the situation we are in currently with NVIDIA.
How do you know it was priced in? You look at what people say they expect, and this is what people say they expected, hence we know it was priced in.
Anything obvious is priced in.
What if I were to tell you Nvidia's monopoly and the implications of AI were obvious many years ago? Many people on HN bought large amounts of Nvidia stock a while ago. Karpathy's "Software 2.0" was widely read here predating much of Nvidia's market cap increase:
https://news.ycombinator.com/item?id=15678587
https://news.ycombinator.com/item?id=34881881
It was "obvious" C3.AI was not a real AI company worth much. The market still "priced in" at $18 billion for months: https://finance.yahoo.com/quote/AI/
If "obviously undervalued stocks" were a widespread phenomenon, active investors that beat the market in the long-term wouldn't be a rarity.
No. This is the same logic that I used when I did not buy Amazon and missed that ride. The market is not always rational.
My hot take: the market is good at pricing in possibilities, but less so at pricing in certainty (or lack thereof).
When AMD launched the Zen architecture, everyone said they were on the up-and-up, and the market gave them a boost. The company's shares only started to climb much higher later on, when there were irrefutable revenue numbers. There was plenty of time to buy and make money. Someone with the domain knowledge like us can definitely have an edge in the market because we understand the product well enough to formulate better odds, then we can profit off the difference between our odds and the market's odds.
I'd be very surprised if Microsoft wasn't one of the investors.
Just one more step for Altman in his Bankman-Fried journey.
I legitimately get their names mixed up, and have had a hard time in occasion wondering why open ai would be controlled by someone who seems to really like doing fraud.
Exactly.
The only difference between the two is that one has a business providing a useful service to millions and millions of people.
[dead]
How could this ever IPO? And if it cannot, who could ever buy it at a multiple of this?
The bigger the raise, the more news and hype, the easier it will be to offload the stock to retail investors. Theres definitely a sweet spot here where I assume the large funds will want retail investors to start picking up the losses instead of them. I don't see this as a long path till IPO like Stripe or SpaceX but I could be wrong
The losses will just get bigger and bigger, and eventually there will be an appetite for retail investors to pick up the tab
AGI will buy back the company.
The financial industrial complex will make money on transaction fees selling the stock to their clients, institutional or otherwise. They don't need it to actually succeed in the long-term - as it is the clients, people's money, on the line, e.g. incentives aren't aligned properly.
Why couldn’t it IPO? People would certainly invest at this valuation.
At this point I don't think they see it as an investment, but rather as securing a previous investment. They aren't looking for a return on the 6.5B, but rather as a deployment of strategic capital to secure their previous investment.
A large amount of AI funding seem FOMO driven. Companies aren't seeing large returns on AI stuff, they are afraid of being left behind if somebody finds some sort of use for it.
Apparently, investors do not see the loss of several leading engineers and the “missing” GPT-5 as a reason not to invest. Maybe they do have moat.
In what way is GPT-5 missing? They never announced a release date, did they? And we had the o1 release just a couple of weeks ago, so why wouldn't they take more time working on improving GPT-5?
Beating all other models in latest benchmarks by a wide margin with a completely new approach, while having the only properly working audio chat with transformer isn’t enough? Not to even mention SORA if actually comes out someday.
I’m not an OpenAI fanboy but come on, credit where it’s due, they have so far been one step ahead of the competition from the start on every front.
I do research in other ML topics but I do follow the qualitative leaps they are pulling off continuously so far.
Do tech giants really want to give Sam that much money and power - it looks like the answer is yes.
What will he spend it on?
6.5B is only 200k H100 chips
Go NVIDIA
Curious does anyone use the secondaries private platforms like https://forgeglobal.com/ to get exposure to OpenAI or similar pre-IPO tech companies?
With inflation being a thing the biggest [sum of money] always leaves me a bit sceptical as people often don't account for it...
The best iPhone of all times!
Same category of headline...
My guess is this investment was contingent on OpenAI going "for-profit".
If you can tell a good story, you can attract an audience with cash.
Way lower than the $100B I heard they were aiming for
The photo on the story looks like Altman is leaving with a backpack full of money?
Of course he's greasing his own wheels. The question is whether the investors will reap any benefits and whether Altman makes OpenAi profitable
What is the multiple over revenues/ ebitda?
I'm not sure how relevant that is at this stage. Not loo long ago I read that OpenAI had ~$2B revenue, but are still losing money.
At this point I'd say this is a faith investment that either LLM/GenAI use is going to explode, and OpenAI is going to get a significant share of that, or that OpenAI will achieve a much more powerful and valuable AGI.
It seems corporate evaluation/adoption of GenAI is still early stages, and not at all clear who the winners are going to be if there is a lot of revenue forthcoming. How does OpenAI complete with free models from Meta, or open source ones, or pricing strategies from giants like Microsoft, Google and Amazon?
Ok so with $2B revenue is a 75x sales multiplier, which means a 54% CAGR for 10 years. They estimates $150B/y revenue within 10 years. Google is at $300B/y revenue.
Llama is good enough and I don’t have to give my data to a very sketchy organization. Happy to never use OpenAI again.
Not today, but I have a feeling that 3 years from now things will look very very different.
I'd say Llama has improved at a much faster clip than anything OpenAI has released, so probably in 3 years what I said will be even more true.
All these models train on the same data (sort of) and so all will pretty quickly reach the same asymptote.
I agree, I don't think there's a moat and this is going to be a commodity feature.
Eventually they'll be put in robots able to gather real-world data live, and they'll have a practically infinite amount of video/audio/tactile data to learn from.
Sure, but this scales rather poorly. How many robots do you expect, and don't you expect some resistance from the public, for privacy reasons at the least?
For me, Llama is definitely not good enough yet. At least, the one I can host isn't.
Try Lambda Labs if you want to host a bigger model.
Gemini is also very cheap and comparable and works well as a drop in replacement for OpenAI models.
Looks like the real winner here is nVidia, because they will be getting their investment money back as payment for GPUs using money invested by MSFT and APPL. Or is that view too simplistic?
This is a definitive bubble.
Inference and training costs with $0 free AI models such as Llama which is good enough for AI solutions are all factors that are actually impacting OpenAI. What is not mentioned is that all these costs are quietly suffocating them whilst the company continues to suffer from a significant talent drain.
Companies like OpenAI cannot afford to continue racing their prices down to $0 and will just continue the pyramid scheme of raising billions to avoid bankrupting themselves over the said costs.
These fund raising rounds of A,B,C,D....X,Y,Z can't go on forever as these costs continue to suffocate the company. If OpenAI fails, there will be a horrific AI crash with hundreds of billions of dollars wiped out.
They're racing to AGI.
Just because they say it doesn’t mean we have to believe it.
Exactly.
At this point, AGI could mean anything, from taking the majority of people's jobs to enriching themselves and never achieving 'AGI' in the first place.
It is entirely possible that 'AGI' could be a massive lie and was all about capturing the wealth for a few people at the company.
Believe what they do; not what they say.
For perspective, xAI raised an almost similar amount at a pre-money valuation of $18B. https://xcancel.com/elonmusk/status/1794975702812426595
Do they have anything to show for it?
A significantly worse model that says naughty words.
I jump between ChatGPT, Claude 3.5 and Grok2 on a regular basis, and I'd say its not "worse" in the general sense. I find it much better on anything sentiment-analysis related or for shortcutting customer research, and image generation is really good.
This seems crazy. It's a truly innovative company and all but they don't have revenue to justify this nor a moat given how many plausible competitors sprung up so quickly. Elon Musk just reached into his own pocket and made Grok appear.
OpenAI is looking more and more like a Ponzi scheme ego play from sama. They'll find ways to offload the losses to retail investors and everything valuable will end up in Microsoft's pocket in the end.
By "reached into his own pocket," you mean he leveraged engineering resources from Twitter, and then once it became clear that Grok wasn't a competitive model, raised $6,000,000,000 in capital from VCs and sovereign wealth funds [1] to pay for the additional compute necessary?
[1] https://x.ai/blog/series-b
If six billion dollars is all it takes to build a competitor, OpenAI doesn't deserve a 100 billion+ valuation.
Really stretching the definition of "competitor" here.
> In finance, the greater fool theory suggests that one can sometimes make money through speculation on overvalued assets — items with a purchase price drastically exceeding the intrinsic value — if those assets can later be resold at an even higher price.
They just need one more set of people past them to buy it. I know lots of uninformed people that would jump on it thinking they are “getting in early”.
I , for one, thank the investors for subsidizing AI and making it affordable for us. I am not sure my measly US$20 per month will be enough to cover the cost of my usage but I am grateful for the fun I am having while the music plays.