Nvidia Becomes First Company to Reach $4T Market Cap
cnbc.com
cnbc.com
What could actually drag Nvidia down and make them spend decades in the dark like Cisco still does? So far the two things I've come up with are: (a) general disillusionment in AI and companies not being able to monetize enough to justify spending on GPUs. (b) Big companies designing their own chips in-house lowering demand for Nvidia GPUs.
I don't think Nvidia can counter (a), but can they overcome (b) by also offering custom chip design services instead of insisting on selling a proprietary AI stack?
There’s a reason why only TSMC and to a lesser extent now Samsung and Intel are the only serious players in top-end semiconductors. You can’t just buy the machine and print chips, the amount of iterative tuning and know-how required to get good yields is immense. Weirdly, the actual bottleneck seems to be the availability of what can almost be described as “master craftsmanship”. But it’s not enough either to hire a couple of masters, it’s the collective institutional knowledge built up over >50 years.
And, of course, TSMC is not worth nearly as much as NVidia either even if they manufacture all their hardware.
Exactly this, as otherwise there would be nothing stopping ASML opening up its own competing fabs next door with their cutting edge machines 12 months before they sell them to their customers for maximum profits, the same way Bitcoin ASMIC mining companies did with their chips.
Or at least some Dutch/EU company in the area doing it, but nobody else can do what TSMC does at the cutting edge. For context, EU's most cutting edge fabs will be the Dresden TSMC one at 12nm.
But also also, it most likely will for the next 10 years.
The monetary push is very LLM based. One thing being pushed that I am familiar with is LLM assisted programming. LLMs are being pushed to do other things as well. If LLMs don't improve more, or if companies don't see the monetary benefits of using them in the short/medium term, that would drag Nvidia down.
Nvidia has a lot of network effects. Probably only Google has some immunity to that (with its TPUs). I doubt Nvidia will have competition in training LLMs for a while. It is possible a competitor could start taking market share on the low end for inference, but even that would take a while. People have been talking about AMD competition for over two years, and I haven't seen anything that even seems like it might have potential yet, especially on the high end.
Also, I think the market has to expand beyond LLMs to areas like robotics and self driving cars (and they need to have real success) for Nvidia to maintain this valuation. I don't think only LLMs are enough because I don't see code assist/image generation/chatbots as a massive market.
And that gives customers the most flexibility. Nvidia dominates training and is highly competitive in inferencing. At the same time, SW improvements speed up single node and networking performance. H100 released 3 years ago is today several times faster than it was on release with constant SW updates.
Customers who buy Nvidia for training today can use the older GPUs from Nvidia for inferencing later. And Nvidia supports even V100 still in SW updates and speed improvements. And since all is based on the same SW ecosystem, it allows for more seamless operations for customers. You can mix different Nvidia GPU clusters but you can't easily mix Nvidia solutions with other solutions.
That is also why Nvidia has always been dominant and that's flexibility. NVFP4 is a good example of what they do to stay ahead. And it is even supported by Hopper so any old customer can use Nvidia's new format to further improve model training performance. Suddenly old Hopper clusters become more valuable with some SW releases by Nvidia.
Nvidia has a track record which no competitor can match. Going with Nvidia is no mistake today while going with any competitor is a risky bet. If you spend billions, you think twice about making bets.
(a) can happen, but Nvidia has some buffer. All these companies promised their shareholders huge massive gains if only they embrace the AI revolution. They will be extremely resistant to admit they can't make any money out of it and will keep the charade going for as long as they can. It won't be like a cliff, so Nvidia has time to adjust and handle it.
The market cap is nonsense though, it's just hype. I never put any real weight on them.
They would need to make more offering design services, giving away their design secrets in the process, than selling a product protected by a massive moat. This would be slaughtering the goose.
If you can present an example of someone executing this strategy as a pivot from an extremely high margin proprietary product and increasing their market cap as a result I would be very interested to read about it.
Admittedly, I haven’t thought this out, but I can’t find compelling reasons why they shouldn’t do this.
Are these companies already more valuable than the VOC at its height, when it owned entire countries? Is that where we're headed?
Billionaires aren't exactly buying more eggs when they have more cash, they're cash is competing for ownership of assets.
One scenario is that maybe a smaller NN is enough for most tasks but you have to train it in a smart way (search, creating feedback, reasoning).
It's a long shot but maybe GPUs won't be the best hardware for the job. It's a pure speculation though.
(c) NVidia isn't all about LLM, it has robotics, embedded, vision.. This is going to be huge as generic robotics hits.
But its the LLM frenzy that made their market cap 20x in a few years, the other things didn't.
Yes, and they have the second big wave coming. This time they have no close competitors.
BTW, without LLMs Russian drones are using latest NVidia for navigation and targeting.
That's what people don't get. We're still primarily in AI research mode. The race in LLM training isn't about making money, it's about a R&D race and whoever gets a better product faster than competition.
Therefore Nvidia won't fail, because we're far far away from any AI production mode since the computing for that would need decades to install. Imagine how much compute power you would need for 24/7 assistant inferencing in real time for every person on earth. We're are just scratching the surface. For Nvidia to fail at this time would be the same as that the world would stop on AI research lol.
Imagine, you have an industry where 80% of all companies' / private investors R&D money becomes revenue of a single company. That's basically Nvidia's position in a nutshell.
Basically, Nvidia could do $1 trillion revenue on world wide R&D budgets alone, no need for their customers to make money yet.
Not convincing, another cheaper and/or faster hardware producer can replace NVidia. Something like quantum or optical optimized for LLM inference.
No. We hate them and we want them to have competitors, aka. "we must have a choice".
I miss ATI, Matrox, S3, 3dfx.
Some other competitor like Samsung releases something unexpected they've been building in secret.
i.e. maybe you need “two hits” to become this big, separated by ~2 decades
For nvidia, it was graphics and then programmable GPUs (CUDA)
For Apple, it was GUI desktops and music players/phones
Google is up there, but I’d argue it’s closer to “one hit”, and limited by the founders stepping back and turning the company into an investment conglomerate, rather than being mission-based
When the founders leave, efficiency and creativity seem to be slowed by competing factions of upper management, often working at cross purposes
I’d say that in the best cases, institutional knowledge can build over 2 decades, but it’s also very possible to lose it
Microsoft - Programming Language (1975), Operating Systems (1981), Office Suite (1983)
Meta - Facebook (2004), Instagram (2010)
I would argue microsoft is unique because of how badly IBM screwed up.
EEE and owning the soul of every customer
Facebook and cool zuck
And even Warren Buffett said that it's unusual for a company to succeed in two different areas, yet Amazon did it.
I would say Meta had only one hit -- social networking. Instagram and Whatsapp were good acquistions, but I think of them more as monopoly plays. Zuck has been investing in VR and AI, looking for the next big thing -- but like Google, he doesn't really know what it is. I'd say it's more driven by a desire to be big and powerful and profitable, than by mission (although of course every case is a mix)
I'd argue that Google also has one hit. Chrome and Android are impressive, but they are not businesses on their own. Self-driving cars didn't turn into a great business.
Google invented transformers, and OpenAI turned it into a product. But neither one got the product + business combo right. OpenAI is very far from profitable, and Gemini looks like a ChatGPT clone to me.
Microsoft is a weird case -- I'd argue that they were second in most of their hits:
- Windows - the GUI desktop was second to Apple
- Enterprise software - Excel was second to Lotus; Word was second to WordPerfect; databases were second to Oracle
- Cloud - Azure was second to Amazon
- Bing - was second to Google
I think Microsoft got to $3T by operational competence -- by cloning and embrace and extend, as opposed to nailing a new hit product + the associated business.
I would give them credit for the early BASIC and OS, but I don't think that's how they reached $3T. Whereas Google basically reached $2T with search+ads, and Meta with social networking.
---
To summarize my off the cuff theory, which I admit you can poke holes in:
- Apple, Amazon, Nvidia: nailed a hit product + associated business combo TWICE
- Google, Meta: nailed it ONCE (and they are younger companies, so to be charitable, they may not have had time)
- Microsoft: a special case
https://en.wikipedia.org/wiki/List_of_public_corporations_by...
..there are some other interesting companies to consider:
Saudi Aramco, Broadcom, TMSC, PetroChina, Berkshire Hathaway
If you build large eCommerce like Amazon then of course you need huge IT infrastructure for it. Amazon was kind of forced to build the IT infrastructure and in that process Bezos saw another business opportunity because not every smaller company can easily build large IT infrastructure but many companies need it. And just as AWS built "SW services" for Amazon eCommerce it also became the foundation of the cloud business.
I'm not suggesting this is all luck, Nvidia has executed very well and their early investments in programmable GPUs really paid off as a result, but a lot of their insane valuation now is due to crypto and then LLMs which are basically two back to back once in lifetime goldrushes where Nvidia happened to be the best positioned shovel seller.
You can run a company well to prepare to ride such a wave should it appear, but you've also got to be born with horseshoes up your ass for this to work out as well as it has for Nvidia
I find this pretty crazy given that (at least for now) NVidia is the most successful startup of all time. Imagine the millions of other entrepreneurs, many of whom worked just as hard, who completely failed in the process.
I recently read in the https://news.ycombinator.com/item?id=44495428 (it is absolutely good article) and the first comment there which I wanted to paste here is:-
It's so, so , so hard to walk the line between persistence (which leads to glory) and stubbornness (which leads to more time following already wasted time.)
-Bruce_511 (HN)
So sure preparation meets opportunity but don't be stubborn just cause you feel like you are prepared.
Work hard but honestly, if you walk the line like bruce said (then persistence will result in you doing the hard work too tbh that's my philosophy right now)
So Everybody just need to figure out this line b/w stubborness and persistance especially in the startup world. But I think we need a book written on this topic tbh, any suggestions anybody?
Not random, but always in the places where the big and easy to get money is. You're not gonna get much VC funding in Botswana or Bangladesh no matter how good your idea is and how smart you are.
That's why so many sacrifice everything to go live in the bay area even if it's an overpriced shithole full of homeless, drug addicts and feces.
BUT one thing I didn't know was that, from the very beginning, Nvidia was not a graphics or gaming company. In college, I thought of them as the company that made graphics cards.
Rather, they were looking for applications of fundamental semiconductor technology, and gaming turned out to be the most promising one at the time.
I learned that from one of the Acquired podcasts; it might have been this one:
https://www.acquired.fm/episodes/nvidia-the-gpu-company-1993...
I'd say it's analogous to Amazon -- Jeff Bezos was actually not interested in books, per se. He was interested in Internet retail, because he saw a rocket that was taking off. The "meta" was that he understood that an online bookstore could inherently have more selection than a physical bookstore. And that he could build this vertical slice business, and expand from there. This strategy succeeded wildly, and he was misunderstood for a long time.
Jensen said something like "Intel and Nvidia are like Tom and Jerry -- whenever I see Intel in a market, I pick up my chips and run !!!"
So crypto and AI were applications of semiconductors that Intel was not occupying. If you wanted to make money in semiconductors 2000-2015, you would probably go for desktops, servers, phones. But Jensen explicitly understood that his smaller company had to find other markets.
So he made a bet, that turned out to be correct. From that perspective, I would not call it luck. All the big companies have been lucky, to some extent, but they haven't become as big as Nvidia.
The market for compute is endless, and Nvidia makes huge efforts to commoditize the software side of things so people can buy hardware.
For example, I'd argue it's a thing that's NOT in Google culture.
Google culture is more about technical advantage, scalability, etc. These kinds of values are typically enforced by employee performance evaluations and so forth
So that is helping Nvidia reach $4T, and all the other tech companies
They also probably have good ways of offsetting the weakness of the dollar, whereas regular citizens and investors don't
Nvidia would need to move on the order of 4,000,000,000 units to hit $4T in revenue, more than triple that to realize $4T in profits. Even if the average per-unit costs are 2-3x my estimated $1k, as near as I’ve been able to tell they “only” move a few million units each year for a given sku.
I am struggling to work out how these markets get so inflated, such that it pins a company’s worth to some astronomical figure (some 50x total equity, in this case) that seems wholly untethered to any material potential?
My intuition is that the absence of the rapid, generationally transformative, advances in tech and industry that were largely seen in the latter half of the 20th-century (quickly followed with smartphones and social networking), stock market investors seem content to force similar patterns onto any marginally plausible narrative that can provide the same aesthetics of growth, even if the most basic arithmetic thoroughly perforates it.
That said, I nearly went bankrupt buying a used car recently, so this is a whole lot of unqualified conjecture on my part (but not for nothing, my admittedly limited personal wealth isn’t heavily dependent on such bets).
In other words, why should it matter to me what the company's profit margin or asset base or what not is actually worth when I make money if the stock number goes up?
If you own a slice of nVidia’s shares at a current P/E of 37, after a year, they’ve earned 2.7% of the value and you still have the same stake as you did before. That’s pricing in further growth and upside in earnings from here (otherwise, you could buy US treasuries at a better price), but doesn’t seem outrageous to me.
Disclaimer: I don’t directly own any $NVDA; I do own mutual funds that own some.
A year ago both its trailing and forward P/E were higher. So the stock is relatively a bargain compared to what it was a year ago.
The price implies that revenues and profits are expected to continue to grow.
> My intuition is that the absence of the rapid, generationally transformative, advances in tech and industry that were largely seen in the latter half of the 20th-century (quickly followed with smartphones and social networking), stock market investors seem content to force similar patterns onto any marginally plausible narrative that can provide the same aesthetics of growth
I wouldn't disagree with this.
That means if we hold constant the profit earnings, if you bought the whole company at its current valuation ($4tr), it would take you 37 years to break even.
Is this reasonable? Depends on sector and growth potential. To me, this is a "fair" valuation and not overly inflated based solely on existing earnings.
We use it as a snapshot in time to check our sanity and to allow us to compare apples to oranges.
That said, you could have made the same statement about AAPL or MSFT 20-30 years ago, and you would have been dead wrong.
As someone who bought NVDA in 2016/2017 and held till now, I'm very happy with the way I applied my software knowledge to profit where I won't have to work again.
Risk taking is best done in domains where you have an edge!
As it went up, I occasionally sold some and took profit, but not much, till now.
I saw how important and productive machine learning was, and it seemed like NVDA had an excellent CUDA moat, so that was my thesis for that bet. In the last couple years I sold some and bought a house.
This is obviously a story with survivorship bias. But anyway my superpower isn't technical knowledge, its the lack of emotion and bias towards non-action when markets tumble. As you probably know, time in the market is better than timing the market.
Or 16 years at 5% inflation.
Fortunately, I opted to pivot towards ratio of total equity, the per-unit activity was a very rough attempt at moving away from abstractions, and that is obviously one of the many flaws in such an exercise.
I already noted the profit margins are incredibly unstable, so I don’t trust the reported figures where they quadrupled inside of a decade. I’m not suggesting it isn’t real, only that it isn’t possible to pin that 55% down as sustainable for any significant period of time, certainly not the 30-50 years is it would take to realize $4T of value at their current pace.
Estimates for next year are $300b. Sources are rumors about the supply of Blackwell and Rubin GPUs for 2025 and 2026 with estimated ASPs.
Nvidia is selling everything they produce so insiders at supply chains can easily tell how much revenue Nvidia will roughly make.
So quite lopsided. That curtails my expectations even further, as this represents a lot of initial infrastructure investments whose long-term expenditures (and viability) remain to be seen.
1. https://www.marketscreener.com/quote/stock/NVIDIA-CORPORATIO...
B300 has been installed a week or so ago. Good luck to AMD to catch up...
They make big bucks on the premium end of their chips. Those contracts are typically on the 8-figure range, I would think they easily have thousands of them around the world.
Even Jensen has implied[1] that the consumer GPU market (i.e. gaming) holds a minor share of revenue these days.
1: Citation needed, I know. I mean comments like "we are not going to abandon gamers, etc...".
That's to say nothing of all the other products and services they build. I just visited their website, clicked on "solutions" at the top, and there's waaaay more there than just desktop GPUs. And its worth noting that NVIDIA doesn't manufacture or sell any of the down-market NVIDIA-based boards.
Given NVIDIA's role in data centers, I think the 4T market cap is, while probably still somewhat inflated by speculation, not so inflated as to be a bubble ready to pop.
I also see where the reasoning here contradicts the reality. If we assume Nvidia only sells $1000 gpus and moves a few millions a year, then how did it received $137B in FY2025? In reality they don't just sell GPUs, they sell systems for AI training and inference at insane margins (I've seen 90% estimates) and also some GPUs at decent margins (30-40%). These margins may be enough to stimulate competition at some point, but so far those risks have not materialized.
Nvidia is going to sell >5 million Blackwells this year and will do $200b in revenue with that alone.
Nvidia has a high net profit margin of >50%. If Nvidia would make $4 trillion in revenue then they would have >$2 trillion in net profit. Then the market cap would easily be 5-10x higher than today because otherwise Nvidia would be the cheapest stock in history of all time.
Market cap is also a very bad indicator as it doesn't really tell how much money was really invested into the stock. Market cap is just a product of shares * prices. For example, I bought Nvidia shares in 2016 for a certain amount. These shares are >100x more valuable today but I didn't put any extra money into them. So 99% of "my" market cap was simply created by traders pushing up the stock price.
If tomorrow, the majority of Nvidia stock holders decide to sell and all stocks are sold then I guarantee you that never ever will $4 trillion be traded because if there is a strong sell move then the stock price will drop like a rock and the last sellers will get a fraction of money as they have based on todays market cap. We might be lucky to see $500b of trading volume.
Some examples:
- Palantir - valued at 337B USD (more than Meta less than 3 years ago !), with a revenue of 2B and net incomre 500M
- Gamestop - valued at 10B USD, with a quickly declining revenue of 4B and net income 100M (they lost money most of the time in the past 10 years)
- Coreweave - valued at 74B USD, on 1B revenue (growing quickly), 300m net loss, and very discutable accounting
- Tesla - valued at 1000B USD, on 100B revenue (which is now collapsing), 7B net income which will most probably turn into a less starting this quarter (no more ZEV credits) or next quarter (no more 7500$ subsides)
- xAI - valued at what, 100B USD ? On probably 0 revenue and huge losses.
Same goes for OpenAI, SpaceX etc, and I'm not even starting to talk about crypto (yeah, Dogecoin has a 'market cap' of 27B USD...).
We are living almost unprecedented times in terms of US stock valuations.
Not longer than 1 year ago, Tesla was still officially projecting 50% YoY growth until 2030.
Amazon also had a p/e over 100 and a p/s over 10 in its earlier history.
You also have to look at the market size these companies are addressing while having few competitors.
Tesla for example is addressing "driving cars" with its self-driving project. There are 2B cars on the road. If we value driving them at $5 per day, that is $5 * 365 * 2B = $3650B in annual value. Capturing 10% of that would be $365B per year. At a p/e of 30 that is over $10T - ten times Tesla's current market cap.
Now Amazon has a market share of 40% when it comes to ordering things.
And even if Nvidia had won that contract, the Dreamcast ultimately failed. Nvidia was close to destruction multiple times in its early years.
That said, I would be wary about buying shares of any company tied to AI right now.
Very few people scrambling to throw money into 'AI stocks' have any idea about technology. When the music stops it's going to be ugly.
This is a technology that will reach 90% usage for almost everything people do, so there is still so much more growth to go.
Edit: scrolling through the entries in the Wikipedia page here (https://en.wikipedia.org/wiki/List_of_public_corporations_by...), it seems likely that this is the highest valuation in real terms (ie, adjusted for inflation)
If so, the total addressable market of Nvidia might be pretty big.
Let's take the human body as a point of reference. The weight of the human brain makes up about 2% of the human body.
Earth weighs in at about 10^25 kg. 2% of that is about 10^22 kg.
All computer hardware on planet Earth weighs what? Maybe a billion computers times 10 kg? That would be 10^10 kg.
So we still would have to up that by a factor of 10^12.
Still 99.99999999% of the way to go.
It is far, far smarter to design the chips and leave the manufacturing to others who will have the expertise and take on the risk.
AMD tried, realized it was completely pointless and split off into Global Foundries. The giant Intel struggled incredibly hard with the 10nm node and has now lost significant ground because of it.
Leading edge silicon is ruthless. You can burn hundreds of billions of dollars a year trying to catch up and if your product is even slightly worse it may as well be worthless - all the cost with none of the demand.
Going into older nodes is far more forgiving, but you can't make GPUs with that.
China has some of the best technical universities in the world. Many of the names you see in the U.S. AI/ML space are Chinese — and a significant number have returned to China, seeking new opportunities. The world has changed.
By 2035, China could be on par with TSMC, or even surpass it. Once, there were Nokias and Kodaks — and like them, TSMC will eventually face real competition. Nothing is eternal. TSMC doesn't possess secret knowledge that others can't obtain or special brains that others cannot match.
China’s future security and competitive edge depend on becoming self-sufficient in advanced node production. They now have all the necessary ingredients, more than anyone before, to challenge and possibly overtake TSMC. It's only a matter of time.
If so, then why aren't we paying them 30% of our revenue like in the AppStore?
Apple did so, and paid so much money they accounted for a quarter of all of TSMC's reveue in 2024; over 20 billion dollars just from Apple alone.
Also Nvidia could even afford to burn 100s of billions of dollars since they are becoming the most profitable company in the world with probably passing Apple this year.
BUT then Nvidia would compete with TSMC and that is a problem because Nvidia is building up a cash and supply moat. Why should Nvidia build TSMC competition if they can simply buy 95% of the packaging supply at TSMC so competitors like AMD and others struggle to get supply? By booking everything at TSMC, Nvidia can easily keep their market share even if competitors improve their products.
This is a huge differences to many industries where most companies have their own production. Imagine there would be only 1 car manufacturer and all other car companies would only design. And now 1 car company would book 100% of supply from the manufacturer. What would the other car design companies do? They would get out of business even if they design better cars.
Nvidia has done this before and that was in the 90s. By speeding up design and releasing new products and more products 2-3x faster than any competitor. The result was that from 90 competitors in 1993, there was only 1 left in 2003 for Nvidia in gaming GPUs.
And Nvidia is doing that again by speeding up their roadmap cadence and as well as booking all supply. Nvidia is crashing competition not only with a great product but by removing their competitors' option to place their product in the market.
At the very least, Apple would stop them. They'd never allow anyone to stop production of their precious Apple Silicon.
The duck is usually so well behaved!