The claim that AI products will never work might be valid. But NVIDIA don't need to care about that at the moment. What NVIDIA care about is the likes of AWS who are investing heavily into GPUs so that AWS customers can determine that maybe they don't need some model with tens of billions of parameters to make a search engine for an e-commerce product that has 100 listings.
> But NVIDIA don't need to care about that at the moment
But NVIDIA investors do, NVIDIA's current valuations relies on GPU's continuing to sell at premium prices at ever increasing demand for decades to come, otherwise they wont be close to make trillions in total future profits that their current market cap demands.
Couldn't agree more with you. I can't invest into Nvidias current pricing because it relies on expectational execution and macros that continue to drive demand with either no competitors eating up market share OR technology not changing on them.
It just seems like the current valuation is upheld by excellent performance and demand outstripping supply in the current moment. Throw in everyone being blown away by LLMs performance + lots of hype and people choosing the one company to put their bet on.
Premium prices is real sticking point for me. 100 billion is not lot to get say 30% that is 1 trillion of market cap. And this can happen in 3, 5 or 10 years... So there will be competition.
And on other hand question can also be total cost of ownership. If you have slower and more energy inefficient competing product. But the cost over lifetime is only say 75% it is not that bad deal...
If gen AI companies and gen AI initiatives within major players start shutting down, there might be a glut of graphics card flooding the market, though the supply might fuel another crypto bubble, so NVIDIA may not be in trouble after all thanks to market irrationality.
And I blame squarely business majors for their blind faith in assets whose price goes up with no explanation, they are the ones who always push the hardest for adoption of these fads with no understanding of what the actual value behind them is.
The blame may not lie in businesses having blind faith in the potential value of LLMs, it may be an issue with incentives. It seems like every major tech player leaned fully into LLMs and valuations ballooned because of the hype. There would need to be concerns over an even worse crash if/when LLMs don't match the hype, otherwise they're incentivized to take the cash up front and handle a smaller hit later.
Worse, those tech companies have been the only thing propping up the stock market, and arguably the economy as a whole. Collectively they may very well fall into the too big to fail category, whether because the feds can't allow them to lose all that value or because investors can't afford to properly value the companies when the hype dies.
I think the blame is pretty clearly on the big tech founders who have so much money that their mentality is "a billion dollar doesnt matter, all that matters is that my company isnt left behind." The founders incentives arent aligned with wall streets, so they have their companies waste all this money on GPUs for the <10% chance that they'll be able to create a product since failing doesnt cost them anything.
I'm actually not sold that big tech founders' problem is their views on money or being led behind. The underlying concern for both is a fear of to lose power and control.
I do agree that part of the massive investment in GPUs and LLMs at the biggest tech companies is fear of missing out. I also think a decent part of it is a lack of concern over real downsides. What will really happen when all the tech companies have to write off all those investments? They either get what is deserved, something like the dotcom bubble, or they all see a short lived decline in stocks that disappears quickly when they find a new hype train to jump on.
It's a different order of magnitude, though. Buying an H200 GPU costs upward of $30,000. The electricity cost to run it at full power for a year would only cost me $675 (700 watt TDP, 11 cents per kilowatt hour at my house.)
I feel they are selling to the same businesses who were convinced they needed BigData to manage their 100 row database and CloudComputing to manage their 100 page static web site.
>But NVIDIA don't need to care about that at the moment.
What? How does a company not need to care whether it's customers are sitting on a financial bubble and might not exist anymore in a few years?
Why would Amazon continue to massively invest in GPUs if they don't have customers for them?
ChatGPT isn't a multi trillion dollar product though, what they mean is that AI products doesn't today work well enough to warrant trillions of dollars in market cap that you see in NVIDIA etc.
It comes down to how many of these offerings can deliver value worth $X/month to fund the multi-billion dollar ventures behind them, where companies find that they are willing to pay that $X per seat for it.
There must be at least 6 major players among OpenAI, Microsoft, Google, Meta, Apple, Stability, and more that I haven't kept track of. NVIDIA has no way of knowing how viable all these businesses are in the medium term, they know how many GPUs they're buying while investment capital is flowing to them.
You are correct chatgpt is a great product. But if thats the only 'AI' product, or there is a handful. Those companies will focus on replacing NVIDIA ASAP. NVIDIA benefits from having a broad set of customers.
This is an important point: what any professional investor (i.e. someone with a fiduciary duty to their company to attempt to make money through investing, rather than just gambling) says in public about their investments, is going to be far more about manipulating the market than observing the market. If they honestly expected their observations to exactly predict the future, publicizing them would just lose them all their alpha.
It's also about manipulating their clients' opinions -- say, in the direction of believing that staying with Elliott will, eventually, work out better than moving everything to ARK.
> If they honestly expected their observations to exactly predict the future, publicizing them would just lose them all their alpha.
Not if they're talking about positions in which they're already fully invested. In that case they are "talking their book".
I doubt you can sort out the noise/misdirection effectively to predict price movements, but if you can you should probably stop talking to us about it.
For the same reason that some people have been listening to them when the AI hype has been a gold mine.
And NVIDIA is worth ~3T so I am sure there’s few people that have been listening to hedge funds and their opinion about AI and NVIDIA for the past couple of years.
Short sellers care deeply about their reputation since they rely on outside money following their plays. If they consistently say "short this" and the stock goes up they run into big problems.
It's a "recent letter to clients" rather than to the general public. They often say what their thesis is. Whether they know what they are on about is debatable.
I believe Elliott is accurate that there is valuation bubble for key AI companies, however, I think it is important to note that this valuation has little to do with the actual performance and prospects of the companies and technology involved, particularly Nvidia, in the near term (2-3 years). The bubble exists from trader behavior -- the valuations have reached heights where stock investors are now rotating out of AI positions in significant numbers. This aggressive valuation growth spanned roughly a year, a duration significantly shorter from the capital purchases schedule for AI investments made by companies making significant AI data center investments.
Nvidia's quarterly earnings report later this month will probably indicate more growth -- the actual health of their business is immaterial and separate from the speculative valuations of their stock value. It is important to see the two as distinct and largely disconnected. Wall Street has little cognizance of the actual value of AI, and neither do we, Apple's consumer AI foray hasn't even been released -- yet the traders are done and investing elsewhere. Wall Street is extraordinarily myopic and divorced from the actual market time scales at play here.
Hardware is always a leading winner in tech super cycles followed by larger wins by software. NVIDIA was always going to lead market gains with AI before software catches up. It’s been this way with the personal computer, internet, and mobile super cycles.
And potential large market crashes. Yes Amazon and Google made massive gains, but very little of the tech stocks that were at all time highs in 2000 remain.
I intuitively believe that from the beginning, markets are usually overly optimistic on the dawn of a new tech and I'm anticipating a correction in perhaps a few months or a year. Anyone knows how to short a stock on that time frame?
“Financial derivatives” and “downside is limited” should never appear in the same sentence without strong conditions on what sorts of derivatives/ combos of them
I think people in the thread are mistakenly conflating the idea of buying options with writing options to sell, the latter of which can have unlimited downside
Let me give you a piece of hard earned advice. You should only ever short companies when you have information that other people have not realized yet that will immediately affect the stock price. How you do that without breaking the law is up to you to figure out.
Thinking "AI is a bubble" is not a good enough reason to short a company.
You can generally buy options contracts for up to two years in advance. If you've never done options before, make sure you do some reading. It's not like stock -- if you're wrong you can loose everything, and in some cases, loose more than you have. Certain options contracts have unlimited downside (As in negative gains).
So be careful.
But google for "options contracts". You'll need a broker and they will most likely make you get approvals for it. Usually you have to show some financial sophistication.
> if you're wrong you can loose everything, and in some cases, loose more than you have. Certain options contracts have unlimited downside (As in negative gains).
Only if you write/sell them. Buying an option you can only lose 100% of what you paid for it. (Which is the outcome most of the time.)
Your overall point of educating yourself on how options work before you jump in is well advised though; options are risky and way more leveraged than stocks.
Of course, but if you don't know what you are doing, it's hard to tell the difference, especially since selling options can make you a lot more money so it looks like the better play if you don't know what you are doing.
Many brokers support shorting - Interactive Brokers for example.
Unlike options you can sell short and hold the position for as long as you can fund it - cover the cost of repurchasing the shares if it goes up and pay dividends etc.
The trouble with shorting meme/viral/bubble stocks is the price may go a lot higher in the short term
In the UK you can spread bet on the market and stocks, i.e. Trading 212, otherwise any share broker should also be able to do it (probably at higher cost).
You may be right but even if so, timing the market is ridiculously hard, specially when shorting.
I don't recall who said it, but if you put the same unexperienced investor in two paralell universes, one where they short and one where they long, the more likely outcome is that they will lose money on both.
The opening position is just one part of the equation.
You can continuously roll options contracts. I don't know if any broker offers an order type to roll automatically; otherwise I really like Alpaca's API (https://alpaca.markets/) to do it yourself.
AI might be a bubble, but Nvidia seems like they are in a decent spot going forward?
Programmers have gotten pretty good at hiding latency nowadays and Nvidia makes the most popular high-throughput devices, and has an utterly dominant software ecosystem for those types of devices.
Meanwhile Intel continues to fumble. AMD actually doing pretty ok on the hardware side, but their software ecosystem is not great. Who’s going to compete with Nvidia? Apple, lol?
Apple, and I think the other dark horse is Sony. The problem has always been ecosystem and these are the only 3 platforms that anyone is in organically because they actually want to be there.
Being a compiler target for someone else’s ecosystem is not a moat. Being a compatible clone of someone else’s ecosystem is not a most.
Nvidia has spent 15 years building out cuda and seeding it into key places, at enormous expense. Apple has the best laptop hardware on the planet and a significantly underappreciated degree of traction among tech enthusiasts and makers that is going to pay long-term dividends if apple can simply avoid rocking the boat.
Sony, similarly, has a successful history and culture of proprietary ecosystem and API usage, and studios are onboard with using it. PS5 Pro might as well come with an RTX sticker given the headline focus on RT and AI upscaling, and I think people are underestimating the degree to which this is going to give Sony an early-mover advantage in shaking out applications of AI in the gaming space. You know the nemesis system patents… they’re well-positioned for squatting patents on key ideas that shake out early.
If an AI bubble exists and it bursts, Nvidia will loose a massive amount of customers. I have no doubt that they can survive, but they will still suffer from a massive financial hit.
>Who’s going to compete with Nvidia?
You know what is the surest way for a market leading company to go bankrupt? Pretending that the current situation is inevitable and no competitor will ever try to challenge them.
AMD is working hard on the software side of things. A great example of this is how when llama3 was released it was day zero was optimized for mi300x. Another example is the cadence / value of ROCm releases.
The total GPU producer valuation ballooned 10x the past few years, even if NVIDIA captures the entire market the valuation still isn't warranted unless the market massively grows going forward.
At the end of the day compute is a commodity. The cloud hyperscalers will buy whatever product 1. meets their needs, and 2. is cheapest. As soon as a competitor releases a product that meets their needs its a race to the bottom in terms of pricing. I just cant see a world where Nvidia never sees meaningful competition.
I mean it depends what you mean by decent spot going forward.
If you mean that their equity will continue to grow and trade at its current valuations I find that one tough to swallow (i.e. I wouldn't buy it right now). Right now the P/E isn't terrible but its under the assumption that their chipsets continue to be the dominant chipset and that all the major tech companies keep buying them at the volumes that they currently do - seems unlikely to me.
Now a good place to work for a company that will be around for a long time -- sure safe bet there.
NVIDIA must be feeling pretty invincible right now considering the last time they were in a bubble and it burst they just land right on top of a new bubble.
To me, we're obviously in a bubble. Everyone assumes that it's going to continue to take economically disrupting amounts of compute to train larger and larger models into the foreseeable future. They think that the need for compute will always go up.
Just like they thought that the price of real estate would always go up, until it didn't in 2006 or so.
We're in for a correction, a strong correction, in hardware demand. Someone will figure out a better algorithm, a better way to do transfer learning, or maybe a better architecture for AI chips that greatly increases performance.
At the rate things are going, we'll be using 2 bit arithmetic (0,1,infinity,-1) for models and all of the huge multiply-accumulate arrays won't be quite so useful.
>AI, it added, was in effect software that had so far not delivered “value commensurate with the hype”.
Sober assessment in my opinion. It is obviously hard to make predictions about the future, but AI, as it exists now, has not revolutionized the world. It definitely makes me some percentage points more productive, but while the overall effect of a widely spread small productivity enhancement might be large, that doesn't means the technology lives up to the revolution it promised.
A lot of companies are marketing products based on very shallow work and shallow understanding of the technology. The engineers and data scientists who know what is going on at these companies are silently coping with having business leaders who have been told about the limitations and are selling snake oil.
Can't read the article, but... yeah I think most people looking at this with any amount of logic realizes that.
I feel like we are in a very rude awakening for a crash very soon, we have already seen multiple AI products come out and fail horribly.
In my view there are 4 possible outcomes:
- It is not overhyped but we realize that the current pace of things is un-sustainable for a variety of reasons and investment is decreased.
- It is mildly overhyped but we realize that the vast majority of what these startups are doing can be done by Google, Microsoft, or Apple with just a minor change to their system prompt. And its moved to on device processing threatening Nvidia.
- It is seriously overhyped, beyond AI startups failing we have a major issue like someone dieing or some other major scandal caused by AI doing something and no human verified it. We go back to recognizing what the technology actually is and what its limitations are.
- We hit a wall with the current implementation and it does a decent job at things, but progress just stalls and hype dies down.
I just don't see a positive outcome given the many issues that have come up and it being shoved into places as a source of authority and time and time again, it fails.
Edit: Thank you for the archive. Yeah that aligns.
> Regardless of hype or utility, these things are ridiculously expensive to run at any sort of scale.
I probably should have been more clear, that is what I was referring to it being unsustainable. Cost, Energy use, Environmental impact, etc. All the things that are coming out recently about the "hidden cost" of AI.
Costs go down quickly. Intels new CPUs releasing end of year have an iGPU equivalent to a 3060 and the entire chip has a tdp of 30 watts. Another 5-10 years and the same will probably be true of todays top of line gpus.
In 5 years, everyone will have CUDA-like solutions. That erases Nvidia's lead. Okay, Nvidia is a bubble.
AI overhyped? What?!?! Has this person used LLMs at all?
Without a doubt, LLM AI is the most useful thing to happen to me. Probably on-par with than the entire internet due to how quickly I can get a response and how detailed/nuanced/niche it can be.
I expect AI to eliminate the need for most diagnosing doctors and pharmacists. If we can't get that back in GDP, its a medical corruption problem. That example alone should be proof that AI is not overhyped.
A single human trying to remember 25 year old outdated knowledge from medical school is better than AI?
Rofl
Anyway, I own a medical clinic, and chatGPT is literally better than my practitioners at diagnosis. I have seen patients cured after 2 years of incorrect diagnosis because we entered their symptoms into GPT3... yeah.. that was GPT3.
That is such a delusional claim and such an obvious lie.
Even if it were true you could easily outperform a chatbot by basic medical knowledge. If doctors do not have that they are quacks and should not be practicing medicine.
Elliott - not as smart as he thinks. NVIDEA is indeed growing explosively, and that growth will indeed moderate, and NVIDEA improves cost/yields it will expand geometrically and decline to a linear rate - others will hop in and share.
As for AI, it enjoys the same geometric growth - that will moderate individually, but applications will proliferate.
This has years to run, maybe decades...
Yeah. The product isn’t AI, the product is Jensen - and he’s a legitimate business savant. He has an uncanny ability to see where the market could be going in 5-10 years and execute iteratively to get there.
Right now you are seeing the RT groundwork finally take root - cerny is talking about how “surprised” they were about studios being so excited about RT. And sure it looks better, but the part cerny missed is that with game budgets ballooning, studios are eager to save wherever they can, and it turns out RT lighting can save about 25% of the development cost of a game.
So now you are seeing a pivot to RT-by-default. We literally already have multiple games that do not even have a non-RT fallback renderer, on non-RT hardware they just do software-raytracing with shit-quality BVH and low ray count done on the cpu side instead of the gpu. Traditional raster fallback is essentially dead in the UE5 era.
So far it’s metro EE, then Fortnite (even on series S!), Alan Wane, Avatar: Pandora, etc.
The writing is on the wall, he did it again. That’s why Cerny is surprised, that’s why PS5 Pro might as well come with an RTX logo (the headline feature is the massive improvement in RT, plus AI upscaling), and that’s why AMD finally grumbled and is pushing forward a proper implementation in RDNA4.
They already work. AlphaFold is just transforming an industry. And runway will be a game changer for Hollywood. Udio and Suno are disrupting the music industry. And this are only some of the innovations that came up in the last 2 years. We live in interesting times.
btw. yesterday I came across withpretzel.com, this is a revolution for working interactively with supercomputers and jupyer.
Something can be significant and overhyped at the same time. I don't have a strong opinion about any of the companies you listed but if you look at the markets they are targeting they aren't very big. Runway, for example, seems to be targeting a similar market as Industrial Light and Magic. Its parent company (Lucasfilm) was purchased in 2012 for ~$4 billion. You need a lot of billion dollar companies to justify the current AI hype, and if AI is used incrementally to serve existing markets (as all those examples are), then most of its success stories will be smallish.
Maybe in aggregate all those moderate successes justify the valuations of Nvidia and the cloud providers but so far there is no massive AI use case, just a lot of little improvements.
Nvidia risk is all about the stability (or not) of their monopoly.
How quickly or effectively will model creators be able to make their vertically integrated custom training hardware is an open question and determines how you should bet imo.
Nvidia is trying to make themselves a vertically integrated product with a necessary software layer on top, but it’s not obvious if that will succeed or if it’s enough of a moat to hold on to the monopoly long term with the incentives as they are.
Intelligent software is such an obvious capability shift that its dismissal as overhyped is dumb. The web was known to be a big deal in the 90s but it was still hard to bet on and in 99 you would have lost a lot of money in most cases. Does that mean the web was overhyped?
Enormous value will be created and captured - Nvidia has a monopoly now, if you think it’s resilient you should buy but it’s unclear how resilient it really is and it could drop in the short term either way. Major shifts are hard to bet on.
Even obvious bets like Amazon in 99 you’d have to have waited a long time to 10x your initial investment.
My personal take was I bought Nvidia when this seemed obvious a year ago and sold it when it hit 3T. It seemed clear it’d capture all the value on training in the short term, but I’m less certain about the future and 3T is a lot.
Yes, of course the most advanced microchip maker in the world, why would there be continued demand for the most advanced microchips made ever?
-please sell i need to reup
> Many of AI’s supposed uses are “never going to be cost efficient, are never going to actually work right, will take up too much energy, or will prove to be untrustworthy”, it added.
Uhh... I was with them on the "bubble" and "overhyped" from a valuation perspective. Without having done the proper math of a valuation model, it does seem (based on gut/experience) that the expected value here is baking in too much certainty when this is still early innings...
But I wouldn't go as far as saying products will never work. It's really hard to make predictions, especially about the future.
Isn’t that kind of the point? Nvidia and OpenAI are clearly seeking to become market-makers by promoting AI hype to the masses: mostly the masses of tech wantrepreneurs who will borrow oodles of dollars and spend them on hosted compute accelerators and paid model / LLMaaS licenses. If that segment succeeds at replacing lots of unwanted tech employees, they all win. If it doesn’t, they’ll move on to the next hype bubble. That’s what these companies do for money, not B2C products like glorified chatbots.
Probably the most famous "activist investor" after Carl Icahn.
They like to call out bad management teams, stealthily build up sizeable stakes in companies until they hit a threshold (5%) at which point they must disclose their position, then they demand a seat or two on the board of directors and start trying to stir up shit to change the company so that it improves its return to shareholders. They often succeed.
Which is to say they have something of a contrarian view on a lot of things. It must be a cool gig, tbh. Beats being the CEO's yes-man which is the case for all sorts of advisor roles..
>It will change society in the same way as Internet and social media.
>Many jobs and professions will be eliminated.
What jobs and professions were eliminated by the internet and/or social media? I can think of jobs eliminated by computers (computer was originally a job title, after all), but none that were entirely eliminated by the internet.
The internet wasn’t designed to take peoples jobs. It was designed to disseminate knowledge and connect people.
AI is explicitly designed to replace workers. To make artificial people who don’t eat, drink, sleep, take breaks, go on strike, ask for a raise, take maternity leave, or disobey.
That's what most people fails to understand: most real layoffs today are due to mere automation, people simply have mostly no more needs to go to a physical bank office, they use internet banking, the bank does not need the physical front office anymore, there is no AI in the game, but headlines say the opposite.
AI is a bubble, something will remain, far resized, but the main issue here is that we are about to loose our software primacy because some big giants want an untenable model to keep their profit instead of keeping the desktop model with the user at the center, not the service at the center with "endpoints" scattered around. Unfortunately DUE to certain commercial practice most people do not know nor understand the power of IT so there is no real push against certain models.
You do remember the dotcom bubble, right? :D
Something can change the world, but still be a bubble.
They are promising AGI, super intelligence or the singularity in less than 10 years. That's the bubble.
