Nvidia’s Blowout Forecast Sparks $260B AI Rally | Hacker News Reader

Nvidia’s Blowout Forecast Sparks $260B AI Rally

bloomberg.com

62 pointsfahrradflucht31 comments

I just can't get my head around this. Sure, Nvidia is in a good position with AI and Data Centre. I absolutely think they're in a great position and they're going to sell a lot of chips. But they're trading at 200 times earnings? I don't see how anyone could reasonably justify that. And not only that, but while yes, the earnings are good, surely the AI hype cycle has pulled forward a decent amount of demand, it's a race to acquire enough GPUs, but Nvidia are quite obviously going to be supply constrained. I don't know, I just find it difficult to understand how this price could possibly be justified.

My guess is that retail investors buy without thinking about the price and institutional investors play along.

For some reason, retail investors look at the price when buying a car, but not at the price of a company when buying a stock. At the same time, institutional investors focus on the near-term due to misaligned incentives. As Warren Buffet put it in 1985, institutional investors can't wait for a good long-term deal or people will start shouting "swing you bum" [1]. So, the whole system is in some kind of crazy frenzy of pumping up the price until it burst. If it bursts, the institutional investors are the first to leave or obtained their fees.

Long story short: I think you understand it perfectly well. Buying Nvidia at a 200 PE ratio makes no sense from a valuation standpoint.

[1]: https://youtu.be/T6HHwOoq9M4

To use a Drake lyric, retail investors are in it "for a good time not a long time". They don't care what the PE is. They don't care what the price is. The volitility is the entire opportunity. You jump into the meme stock and diamond hand your holdings, and just look it pumped 25% today. Its a gamble, that's why it makes no sense to traditionally minded investors like yourself and not gamblers who consider things like potential payouts multiple times over their initial bet, and are buying nvda calls right along side them betting on games through draftkings. If you had like $500 money you could stand to lose, why not role the dice? $500 in options for nvda yesterday would have made you like $10k today, or if you set a stop loss at say 20% you'd only be out $100 for that gamble. It almost makes less sense to sit on the sidelines considering these fixed losses and unlimited profit potential with call side options on a stock that's being actively memed. Have one of your runners pop off then all your picks from there can be made with house money.

> For some reason, retail investors look at the price when buying a car, but not at the price of a company when buying a stock.

This because these are very different processes from the buyer's perspective:

A retail stock investor primarily decides to spend $x on some ticker, and then they divide $x by the current share price to determine how many shares to buy (or they buy fractional shares if their platform supports it). So they can "own" NVDA by paying whatever amount they choose.

This is the opposite from buying a car, where the buyer has to pay the full sticker price all or nothing.

I don't think you can say that based on how many A100s you own. Perhaps with more A100s, you could confidently say what the PE should be, rather than simply claiming it's too high.

Investors are herd animals. If you were to invest in tech companies by what you read on HN, you'd often beat wall street investments by multiple years or months. I'll always remember reading a 2013 analysis of AMZN by Morgan Stanley that spent a few dozen pages comparing Amazon to Walmart and didn't make a single mention of AWS.

Similarly the AI thesis for NVDA has been clear for multiple years, even if it got obscured a bit by the Crypto thesis. Deep learning and ML has been an obvious trend that requires increasing amounts of GPU power since at least 2017. There are basically two companies manufacturing GPUs and Nvidia is one of them. It shouldn't take a genius to have picked that stock.

Frankly we should amend the old quote about the barber recommending a stock to be about Wall Street recommending a stock. By the time the finance bros catch on to a trend, it's too late for the tech bros to profitably invest in it.

I think there is a heavy quantitative induction mechanism in the markets, that is, a positive feedback mechanism from index funds trying to minimize divergence, or hedge funds trying to font-run those algorithms. So it doesn’t take much to set off irrationality, and not much exists to pull it back, because income investing basically doesn’t exist, and therefore value investing has little support.

Partial side note: obviously the FOMC has a lot to do with investment flows, but I’ve found that the tax system is just as relevant or more so in market distortions, but it is poorly understood, rarely observed, and never considered for alternative possibility, like the water that our economy swims in. For instance, the combination of 401k and double taxation killed income investing. Also, capital gains tax and long-term preference deters sale in a pump, and actually works the other way to prefer short-term selling of losses, so the net effect is to add to the positive feedback. Every tax situation is a comparison of speculative gain and loss on an efficient market hypothesis prior, versus a guaranteed tax loss.

Hype drives narratives which drives valuations. Never underestimate hope or confidence as a multiplier influence.

Consumer ( or Crypto ) Business dropped by around 20%. While Datacenter is 15% increase. Even if both Consumer and DC were to double their Revenue, this will still be 100x P/E. And you can safely bet consumer wont double, and leaves those headroom for DC, i.e DC segment to grow by 2.5x. It will still be 100 PE.

Could they do that? Possibly. But I still think it is too much hype.

You can argue that this PE, at this market cap implies, it will eventually have the world's largest market cap (larger than apple...maybe 2X).

Would think OpenAI and others would tire of buying chips from NVIDIA at some point well before that and make their own (like Google does).

Yes, this is exactly the point - some people are making arguments about Nvidia's amazing software stack. Sorry, but Amazon and Google have some pretty fucking good software engineers, Google already tapes out TPUs. Oh, and those are Nvidia's biggest potential customers.

Honestly this stock market move makes it look like Google should just announce "hey we're putting the TPU on sale for general availability" they'd jump 25% overnight.

TPUs are available on GCP, but working with them has been frustrating.

My experience has been that it takes not insignificant effort to convert training scripts and then some weird unexpected bug takes a while to figure out, I’ve heard similar things from peers in academia.

Additionally, at least in early 2023, PyTorch had a substantial throughput reduction on TPUs so you’d probably need to use Jax (or god forbid TF) for efficiency sake.

Granted, I’ve heard some of the PyTorch XLA issues have since been improved.

Regardless, the H100 currently significantly outperforms TPUv4 in throughout on transformer loads, we’ll see what TPUv5 looks like to be fair but it’s not a given that Google/Amazon can outpace Nvidia in manufacturing and chip design when this is their core product and they also have amazing engineers + a large open source community building around CUDA.

I dunno if I'd trust Google hardware from a reliability and support standpoint.

It's a bubble for sure.

But as long as you're on the first half of the bubble, you're okay.

would it be good idea to short the stock though?

Probably not. When you short, your potential gains are limited to 100%, but your potential losses are unlimited.

"The market can stay irrational longer than you can stay solvent."

- A. Gary Shilling (https://quoteinvestigator.com/2011/08/09/remain-solvent/).

Naked shorting certainly leads to unlimited losses, but it’s not the only kind.

Simple example is that i can buy a put option in some company. If the stock price falls below that point, I buy shares and sell them to the option writer and make a profit. My maximum loss is limited to the amount I spent on the option.

Likewise the put seller has not taken unlimited risk - their risk is limited to the cost of the shares they’ve offered to buy at.

The most likely winner of such schemes are the brokers who sold two things instead of one. Buying great companies at fair prices is much easier and, in my opinion, safer. Another argument: how many short seller billionaires are there and how many long billionaires?

Writing options is an incredibly competitive market, margins are very tight. The quant firms writing them only work because of the massive scale, they lose almost as often as they win.

I worked at a large semiconductor company in 2013-16. It was very clear it was fucked. It took about 8 years before its stock price suffered any real headwind in its stock price. So who the fuck knows.

GPU makers really lucked out in the last few years. First from the crypto craze and now the AI craze.

I dunno if it’s “luck”. Nvidia in particular made a big bet more than a decade ago on GPU’s having strong aptitude for important general computation problems. Turns out they were right.

Agreed, if you were playing with computer vision and deep learning ~7 years ago, you knew the potential. Nvidia maintained their stranglehold on the software and hardware platforms. OpenAI (and all the great open source research published by others) really blew it open for Nvidia.

Not to mention the whole "video game" thing that people really like

They've been lucking out since the n64 was released.

The second seems inevitable from the first. Crypto was when people discovered the massive computational power of GPUs, and AI is when people figured out how to direct this power to do useful work.

https://archive.today/cmHCP

Must be the largest one-day market cap gain in the US history or very close.

I believe it beat Apple's one from 2022...

Where are the sales going to come from? is there a breakdown?

Data centers are over half of the revenue. Rest is mostly gaming.

https://s201.q4cdn.com/141608511/files/doc_financials/2024/Q...

Reply on news.ycombinator.com