On the other hand, my uncle lost a fortune investing in a wireless networking company in the 00s. He has worked an entire career in networking and has contributed to bleeding edge systems in the area. This company had excellent tech, but ultimately didn't succeed. A lifetime of savings kaput.
nvidia's p/e is also insanely high. "Wow they sure sell a shitload of product" isn't necessarily enough to justify the price. There are highly profitable companies that nevertheless see their stock price turn around.
>> There are highly profitable companies that nevertheless see their stock price turn around.
This happened with a local sports drink maker. It was branded for kids, more healthy, less sugar then Gatorade and energy drinks. Company took off, went public, stock went through the roof for about two years.
Blue skies, everything coming up roses, lots of articles in the local business mags and websites.
Then their supply chain dried up - a precursor to the pandemic and the founder even said in a startup presentation they were the canary in the coal mine and one of the first businesses in the state to suddenly have their product, packaging, and materials all just evaporate in a matter of weeks. Suddenly they couldn't get product into stores, stores eventually pulled their placement and within three months they were bleeding money horrendously while scrambling to find replacements. Something they were already working on, but soon enough every supplier they'd call had the same answer, they too had no means to ship stuff out and they too were dead in the water.
Then two months later the pandemic hit in full force and it was the death knell for the company. Delisted, and bankrupted, they closed up shop about 3 years after being a "can't lose" stock and company.
My family have all invested heavily in Nvidia and they're making good gains now, but I'm seeing the same thing you are - this can go bust very fast if Nvidia doesn't manage this really well.
The vast majority of financial issues would be prevented simply by following Buffett's advice to not lose money.
It's tongue in cheek, but he's correct. Suppose you have a great idea for investing. It doesn't matter how good it is... don't put your life savings into it.
If you do this you will make money. Investing is mostly about not losing money.
Counterintuitively, when I focus hard on not losing money, I become too risk averse, I fail to take appropriate risk, and my returns are stuck in the low single digits.
I can't count the times I sold great companies like Apple and Netflix and Tesla years too early because I was afraid to lose money and wanted to "lock in" a 50% gain.
By focusing on potential for high returns instead of not losing money, with a diverse portfolio of assets that don't correlate perfectly with each other, total returns are much greater even though individual bets can show big losses for months or years.
This is where a diversified portfolio makes sense. It's completely fine to take a "big bet" on what would amount to a years worth of returns/cash investment every now and then. But many great opportunities have already been priced appropriately, so the extra concentration of risk doesn't yield a great return.
If one was to be able to simply select the 50% of companies which perform better than the rest of the field. You would be in good shape.
THIS. As much as I wish I put all my money into NVDA when I bought it, I’m happy to have a lot of diversity in high tech IT. The daily swings are not nearly as bad and aim not biting my fingers all day long.
Sure, but not diversifying is also one of the most efficient ways to go broke. Which is something that diversifying will make much more difficult.
Also, full baloney. I was not diversifying for many years and it indeed made me great money (thanks MSFT). But when I started getting spooked and diversified, guess what?
I still ended up doing pretty well, even if it wasn’t on the same level as before (look up MSFT share price change between the start of 2017 and 2021). But it was so much safer and reliable, going broke wasn’t as much of a concern, and I knew I was much more secure in case of a downturn. Winning on risky triple digit percentage gains feel great, but I would rather take much safer diversified 50-60% gains over a 3 year period instead.
Not saying that those 50-60% gains are even close to what I would expect from truly safe plays. But safety and risk is a spectrum, and you have more choices than just “fully diversified super safe index funds” and “all-in on one single ticker.” You can adjust and make things diversified and safer than all-inning on a single ticker, while still maintaining some amount of risk that would allow for outsized gains.
I am not saying anything that you are attempting to dispute here... Your logic is sound and you can get rich doing it. But you won't get wealthy. And getting wealthy does carry higher level of risk, I would think that is common sense.
To me diversification goes against all logic because the rule #1 of investing should be that you as a investor KNOW what you are investing in. You can't tell me anyone investing in say S&P 500 has done extensive research on each every of the 500 companies. All they are hoping for is "hey, these are 500 biggest companies in the World, imma just put my chips here and hope for the best - history tells me that is probably safe bet."
On the other hand, you can do full-on research into a single or handful of companies and then put your chips there. You can't tell me that putting money in Magnificent-7 say 5 years ago was any riskier than putting money into S&P 500... and yet you could have gotten REALLY wealthy with the former and quite rich with the latter...
> To me diversification goes against all logic because the rule #1 of investing should be that you as a investor KNOW what you are investing in.
I largely agree with what you say. However, diversification has degrees, and it doesn’t necessarily mean that you gotta spray and pray across the whole range of S&P500 to be more diversified than the “all-in on a single stock ticker” strategy. Examples:
* All in one single stock ticker - no diversification
* All in a few different stock tickers that are in the same industry sector (that you are knowledgeable about) - diversified businesses, but not diversified across industries
* S&P500 spray and pray - largely diversified
Option #2 is imo the solid middle ground, and it gels perfectly fine with your idea that you gotta know what you invest in. Yes, it is riskier than option #3, because it doesn’t account for the scenario where the entire industry sector experiences a downturn. But it is still diversified, still has the potential to make you wealthy, and is not nearly as risky as option #1 (but also not as capped as option #3).
Wealthy folks aren't diversified, because they have _control_ (and knowledge) over their particular investment (ie company founders).
Stock investors do not have the luxury of control, thus they must diversify.
And generally that's what the wealthy do. They go all-in on their own company, grow it to incredible returns, then use those returns to be invested in a diversified manner to grow further. Other stocks, realestate, angel investing, etc.
Most of the billionaires are like this. Or if you're Warren Buffet, you invested in a diversified manner, because he didn't control the companies he owned.
> Or if you're Warren Buffet, you invested in a diversified manner, because he didn't control the companies he owned.
you should check out Buffet's portfolio - he's not very diversified at all... If that was your portfolio someone would tell you you are nuts/gambler/...
It's impossible to lose money you don't bet. I don't understand this mentality. If you have a 30k portfolio... fine bet the 5-10k... even if you lose it you're fine.
> I can't count the times I sold great companies like Apple and Netflix and Tesla years too early because I was afraid to lose money and wanted to "lock in" a 50% gain.
Ah the sunk cost fallacy. Having an exit strategy is important. Never beat yourself up for an appropriate exit strategy.
FYI: This isn't a case of the sunk cost fallacy. The sunk cost fallacy involves continuing an investment due to the amount already invested, regardless of future potential. What you're describing is more related to regret aversion and hindsight bias.
At the same time, Buffet and Munger also think that diversification is dumb if you know you have a good stock, putting all your eggs into one good basket, is better than putting your eggs into multiple bad baskets.
Remember Zoom trading at a P/E of 400+ during the pandemic? They've lost 85% of their value since then, despite growing their earnings about 3-4x in 4 years and becoming the namesake for an entire generation.
Insanely high PE? On earnings day? Isn’t today the day that it adjusts for this new amount? I still don’t find the forward PE to be that bad for the best tech company in the world right now.
True. But another piece of knowledge about Nvidia is its competitors are waaaaaaaaay behind. They stand peerless.
Another piece of information is that CUDA software was provide free or cheaply to Universities doing LLM research I think. And the software is easy to use.
> its competitors are waaaaaaaaay behind. They stand peerless.
Its competitors are only way behind when it comes to software support. The hardware coming out of Intel and amd is, especially for its price, very capable. Given how much money is being invested in AI right now, I don’t see Nvidia’s moat lasting more than a few more years.
But every big tech compoany is handing nvdia billions a year now. Its not just amd working on amd's software anymore. Now theyve got armies of open source developers from around the industry ready to save their company a boatload if they can get the software working.
The number of companies that wish they were a SW company rather than a HW company is very long, and despite all their efforts once they get to a certain size the die is cast.
Getting good AI talent now is very costly. HW engineers are cheaper.
Nvidia has more SW than HW engineers for a reason and the transformation for that started slowly almost 2 decades ago and accelerated 2012 with AlexNet, the first public showcase of a NN running on GPUs. Jensen saw what that meant and transformed the company from that moment focusing on DeepLearning.
Nvidia isn't waiting for a market to develop but prefers to create markets by tackling hard and complex problems. It seems that Nvidia got lucky with AI but it was a long lasting preparation for Jensen.
We've been hearing that line for years, but HN comments are a valuable source of info on where NVDA might go in future. You're the only poster here I could name who is strongly pro-AMD in the AI space. Everyone else seems to put their toes in the water in the hope of being able to get an edge by avoiding NVIDIA's monopoly prices, immediately gets burned by trash software quality and runs away screaming "never again".
I only dabble in AI stuff but have decades of experience doing quick surface-level quality checks of open source projects. I looked at some of AMD's ROCm repos late last year. Even basic stuff like the documentation for their RNG libraries didn't inspire confidence. READMEs had blatant typos in, everything gave off a feeling of immense lack of effort or care. Looking again today the ROCrand docs do seem improved, at least on the surface, I haven't tried it out for real.
But if we cast the net a little wider again, the same problems rear their ugly head. Flash Attention is a pretty important kernel to have if working with LLMs, maybe I'd like one of those for AMD hardware?
We're in luck! An official AMD repo with flash attention in it, great! Except.... the README says at the top:
Requirements: CUDA 11.4 and above. We recommend the Pytorch container from Nvidia, which has all the required tools to install FlashAttention.
Really? Ah, if we scroll down all the way to the bottom we can find a new section that says "AMD/ROCm: Prerequisite: MI200 & MI300 GPUs". Guys, why not just rewrite the README, literally the first thing you see, to put the most important information up front? Why not ensure it makes sense? It takes 10 seconds and is the kind of attention to detail that makes me think the rest of your work will be high quality too.
Checking the issue tracker we see people reporting that the fork is very out of date, and that some models just mysteriously don't work with it due to bugs. These issue reports go unanswered for months. And let's not even go there on the hardware compatibility front, everyone already knows what "AMD support" really means (not the AMD cards you might actually own) vs what "NVIDIA support" means (any device that supports the needed CUDA version, of any size).
I would never try to defend AMD with regards to them needing to catch up. Even talking with executives at AMD, neither would they. Nobody is trying to pull a fast one on this.
What has changed for certain, is their attitude and attention. I just got back from Dell Tech World. Dell was caught off-guard with this AI thing too. It is obvious the only thing that anyone is talking about now is "ai ai ai ai ai ai".
Give them a bit of time and I think they will start to become competitive over the next few years. It won't happen over night. You won't see README's fixed right away. But one thing that is for certain, they are all at least trying now, instead of pretending it doesn't exist.
Whether they will be successful or not, is yet to be seen. I wouldn't even know how to define successful. I don't think anyone is kidding themselves about Nvidia being dominant. But, I'm personally willing to bet on them selling a lot of hardware and working on their software story.
It's great that you're pushing AMD forward, no doubt about it! And I'm sure they'll make good progress now. Like I said, the ROCrand repository seems to be in much better shape now.
What is telling for me is that ROCm itself is on regular cadence updates. Not just small updates, but actual meaningful fixes and improvements.
Not only that, but it is all being done in the open, unlike their competition. Hotz demanded some documentation, they provided it and he still complained. Some people just can't find happiness.
Now, whether or not I am pushing them forward is yet to be seen, but at least I'm trying. By positioning myself as a new startup who's trying to help... that will easily garner all their support as well. As I said in another comment, why not let them try too?
First off, it’s a HW/SW solution and things like CUDA/NCCL/etc make a HUGE difference.
Second, the token/watt ratio of every other option is nearly an order of magnitude difference in real world tests. When you add in custom silicon like moronic Grok/Dojo and you see that there aren’t really any close competitors when using custom spins. That is money down the drain IMO. Best bet for most enterprises is to buy 25% AMD and 75% H100 if they can get it.
I think Blackwell is potentially a long term generational problem due to power limitations in most data centers for now.
Isn’t that like saying Apple’s iPads will only maintain a moat for a few more years because MS Surface tablets match in hardware and “only” lag in software? My general point is that the core software (and the software ecosystem built around it) is half of the product. NVIDIA isn’t going to stand still either.
You ignore a very important fact which goes above any cost and that is security and stability. A fast unstable or unsafe system belongs into the trash can in any enterprise.
B2C will first ask about security and stability.
Do you think AWS, Azure and GCP are the cheapest cloud offerings? Of course not, but why do they dominate cloud computing in B2C while price gouging everyone?
Because they offer something beyond price and that is security and stability as well as a reliable partner. They also offer support and capacity on a level which a startup CSP will never be able to offer.
This is also the reason why all AI accelerator competitors won't be a competition for Nvidia.
To beat Nvidia it's not only about beating CUDA, it's about beating Nvidia Enterprise AI suite with it's security offerings and support options. But enterprise business level SW is a level where AMD and others will never go to and will have to rely on Big Tech like MS, Amazon and so on to do that for them. But why should they if they have in-house solutions? Big CSPs developing their own AI accelerators shows you that they understand Nvidia's business model and are trying to compete head on because they understand that Nvidia is attacking them at enterprise level with AI enterprise solutions. And of course any enterprise using Nvidia enterprise SW will automatically use Nvidia HW.
Once SW is more spread than HW then it dictates where the direction goes. If MS releases Windows 12 only for ARM then Intel and AMD are immediately screwed and they can't do nothing about that. No enterprise in the world cares if their CAD system runs on x86 or ARM as long as it can be used for the intended use.
You make a lot of assumptions about what I understand.
If I am in charge of a data center I had better understand the impact of security and stability as well as the qualities of vendor relationships on my costs or I probably won’t be in that role very long.
You, on the other hand, apparently have never managed an enterprise ISA transition, or even cross-compiled software. The idea that Microsoft would just do that and that it would work is naive in the extreme. CAD software is compiled first for an architecture, and then generally within an operating system. It is all interconnected and interdependent.
I don’t really see the software ecosystem catching up that quickly is the thing. Sure, we have some support for hardware like Google’s TPUs and Coral but the field’s practitioners and researchers are oftentimes so behind the curve of general systems work like dealing with the nuts and bolts of libraries and package management that anyone trying to compete against NVIDIA will need to spend a lot of time investing there rather than yet another group of ML engineers that shudder at the thought of packaging and distributing their software to the public and supporting frameworks for years with partnerships and continued investments doing a lot of work that’s basically toil and extremely undesirable for said ML engineers.
Isn't CUDA basically hidden behind pyrorch and others?
I know that it's touted as the key competitive advantage, but it seems to stem from the fact it actually works, unlike others.
Still great advantage, but not a lock in. If competitors get their act together, couldn't they just replace CUDA with another API, all hidden somewhere in the sw stack?
For training workloads? Yes. For inference? Nope. Their competition is very technically strong there. The lack of market attention (pun intended) is utterly baffling there. And before someone inevitably chimes in about CUDA - for inference you don’t need it. Ask Google for example.
The forward PE is not insanely high in today’s market, and this is the PE figure investors care about. It’s not “oh they sell a shit load of product” it’s “oh sell a shit load of product with ridiculous high margins and they’re growing rapidly YOY”. Will NVDA correct at some point? Of course.
It's more that a high P/E means that the company itself needs to make a lot more money for the long-term shareholders to not be holding an investment that doesn't pay off in the end. You can still make plenty of money speculating on it as it's going up (so long as you sell), or short-selling it as it's going down, but generally speaking it's a sign that the business is being valued the assumption that it will grow by a lot, which is riskier (for example, Tesla's valuation more or less requires that they eventually monopolise the car market or enter and dominate some similarly large market. You can make an assessment yourself as to whether that's likely to happen if you want to make a prediction for how Tesla's value is likely to change over the long term).
A company with a high P/E is by definition not making much money for its shareholders, relative to the size of their investment.
Unless you mean that the share price may appreciate. That's absolutely a thing, but it's a dangerous game. Of course plenty of people have made fortunes this way; people have also lost fortunes; I think the advice to steer away from such companies is basically a statement about risk.
Or it means that earnings are lagging the price increase. They just announced a 629% increase in earnings from a year ago (461% non-gaap) and it seems to be accelerating.
You need to “get it” because it’s really important.
We’re all in here arguing about PE ratios of tech companies reaching 100x. Is that too much? Who knows. For the best tech company in the world? What is the limit?
But for other companies like Tesla, their PE was once 1000x. That’s crazy town.
PE is the first number you should use for comparing two stocks to determine value vs risk.
People get too caught up on the listed/current PE, but what matters is forward PE. The stock market is about the future, and growth company PEs as listed are always going to be behind.
High PE companies arent making tons of money for their shareholders, thats what a high PE means. Theyre making less per dollar of share than the average company
Some people care about PE. Some people don’t. It’s just one metric that represents past performance. If it’s high, the market believes the stock price will go up.
No, P/E being high means the stock market thinks the company's income will grow by a lot in the future. The price may go up or down as those expectations for the future change (a high P/E stock can announce record profits and have the stock go down, if in fact the stock was being priced on the expectation that the profits would be higher than announced).
nvidia's p/e is also insanely high. "Wow they sure sell a shitload of product" isn't necessarily enough to justify the price. There are highly profitable companies that nevertheless see their stock price turn around.