A few months ago, I wrote a post about how Nvidia had grown into a juggernaut thanks to its dominant position in the AI ecosystem: as the world’s largest supplier of GPUs— chips that are used to create and power AI applications like ChatGPT— Nvidia is making an insane amount of money.
I wrote that post back in March, when Nvidia was trading at around $95 per share, with a market value of around $2.4 trillion. It had already doubled since the start of 2024 and risen sevenfold from where it was at the start of 2023.
Today, it’s up another 36% from those heights and it’s neck-and-neck with Apple and Microsoft as the world’s most valuable company, with a market cap in excess of $3 trillion.
The stock just keeps going up and up as the company continues to exceed already sky-high expectations.
The numbers are staggering. Nvidia is currently expected to earn $67 billion of profits this year, which is up from the $50 billion that analysts expected the company to earn at the start of the year.
And things might get even better for Nvidia in 2025.
The company’s next generation Blackwell AI chips are scheduled to ship later this year, and that could lead to another surge of demand.
Analysts are currently penciling in $91 billion of profit for the company next year, but those estimates are likely too low, and Nvidia will probably end up making well over $100 billion when all is said and done.
These enormous numbers are a reflection of the fact that Nvidia’s customers just can’t get enough of its AI chips—chips that are seen as the most cutting-edge when it comes to AI.
And of course, there’s been a ton of excitement surrounding AI, ever since OpenAI released ChatGPT back in November of 2022.
Nvidia’s largest customers— Microsoft, Alphabet, Meta, and others— are all extremely excited about AI and what it can do for the world and their businesses.
They’re racing to buy up loads of chips from Nvidia, so they can use them to incorporate AI features into their existing products and to create completely new AI-powered products.
I don’t think we’ve seen this much excitement about a technology since the internet boom— and eventual bubble— of the 1990s.
And that’s what makes me believe that we could be in the early stages of a bubble today. Think about it— all of the ingredients are in place for an AI bubble to form.
You have a world-altering technology that has the potential to revolutionize many parts of our lives. That’s created an environment of intense excitement that’s causing people to want to invest in this technology.
Over time, that excitement might build and eventually turn into a frenzy, drawing in huge sums of money as everyone seeks to capitalize on what seems like an unstoppable trend.
Handful of Winners
Bubbles share a lot of similarities, but they’re all different as well. What I think makes this potential bubble unique is that there are only a handful of winners or perceived winners (stocks that investors think will benefit from the AI boom).
Almost all of the money flowing into AI today is related to the build out of the infrastructure that’s needed to power AI applications—specifically, AI data centers, which are made up of high-speed networking equipment, specialized cooling systems, and of course, Nvidia chips.
Nvidia is taking the lion share of the money that is going towards this AI infrastructure build out, with around half of every dollar spent going to the company.
Sure, there are some other winners, like Dell and Supermicro, which are taking Nvidia chips and other components and packaging them together into powerful AI computers; Arista, whose networking gear links those components together; and Vertiv, which is supplying power and thermal management systems to the data centers that house AI computers.
There’s also Broadcom and Marvell, which are helping their customers create custom AI chips, as well as AMD, the second-largest maker of GPUs behind Nvidia.
But Nvidia is far and away the dominant player in this AI ecosystem. The company’s AI chip market share is estimated to be upwards of 95% thanks to its advanced chip designs and accompanying software.
Other than Nvidia, the only company making significant sums of money from AI today—meaning more than $5 billion per year—is Broadcom.
(Even OpenAI, the company most consumers are familiar with when it comes to AI, is only generating annualized revenues of around $3.4 billion today—though that number is growing fast).
And that makes this situation different than other bubbles, like the Covid bubble of 2020 and 2021, where there were a ton of different winners or perceived winners—from electric vehicle makers to software companies to biotech firms— or the internet bubble of the 90s, where any company that put .com after their name was bought up by investors.
That’s why I think this could end up being the biggest bubble of all time. Investors who want exposure to AI have a limited number of choices. So, all of the money that wants to chase the bubble has to flow into a small group of companies, and primarily, Nvidia.
As I mentioned earlier, Nvidia is already one of the biggest companies in the world, but it could reach unimaginable heights if this bubble thesis plays out.
Is it a Bubble Yet?
Now, don’t get me wrong—I’m not saying Nvidia is in a bubble today. You can make the case that at 35x next year’s estimated earnings, the stock isn’t particularly expensive given the company’s growth rate and given the likelihood that Nvidia will easily beat the current earnings estimates.
But there are others who believe that Nvidia is already in a bubble because its earnings are being fueled by what they consider to be AI hype.
This is a view that was recently laid out by analysts at Goldman Sachs in a widely-read research report in which they argued that the large companies that are buying up loads of Nvidia GPUs to power their AI applications will end up losing money on those investments because AI isn’t as revolutionary as they think it is.
Another bearish view that’s been going around is this idea that even if AI is revolutionary and there ends up being many successful AI products in the future, Nvidia won’t be able to sustain its super high profit margins.
Last quarter, the company kept 57 cents of every dollar in sales as profit.
That’s an unusually high profit margin that’s well above the company’s historical profit margins, as well as the profit margins of other large tech companies.
Even if Nvidia remains the leading provider of AI chips in the future, increased competition may lead to lower profit margins over time, which could weigh on the total earnings of the company.
So, those are two bearish views that suggest that Nvidia could be in a bubble already.
Personally, I don’t buy the first argument— the idea that AI isn’t revolutionary. It seems to me that every time we have a new, groundbreaking technology, there is a group of people who dismiss it based on its capabilities at the time.
It’s the equivalent of downplaying the internet in its early days because it was too slow or because it was used to browse rudimentary websites and send text-based emails.
For whatever reason, the skeptics don’t want to or can’t foresee what the technology might become in the future.
If you’ve watched the rapid progress of AI just over the past year and a half, you can see how quickly this technology is progressing and, with just a little imagination, envision the incredible things it could enable in the future.
Not to mention, AI isn’t just about creating new products that generate new revenue streams for the companies that are investing in it— it’s also about the cost savings that companies can harvest by making their workers way more productive.
So, with that said, while I’m not concerned about AI’s revolutionary potential, I am concerned about the margin compression story for Nvidia. It’s hard to imagine the company maintaining profit margins of 50% or more indefinitely. And if those margins eventually come down, that could be a headwind for the stock.
Room to Run
But I don’t think any of this matter over the short-term—at least the next six months, maybe even longer.
I think there’s enough hype surrounding AI and momentum in Nvidia’s stock, that it can continue rallying from here.
This is more of a “gut feeling” type of call. After all, bubbles are driven by human emotions—excitement, hype, and FOMO— and you can’t really quantify those things.
My sense is that we haven’t reached the peak of the euphoria. History suggests that bubbles can grow to be much bigger and last much longer than you would expect.
There’s also the potential for the rally to broaden out beyond just Nvidia.
Though far from the gains for Nvidia, we are seeing rallies in AMD, Marvell, and other semiconductor companies that are helping build AI infrastructure along with Nvidia.
We’re also seeing strong returns for Microsoft, Meta, Apple and other megacap tech stocks that are using that infrastructure to create AI-powered products for their customers.
And at some point, if the AI hype intensifies further, we might even see smaller, more speculative AI-related companies begin to surge. Just like any company with “.com” in their name soared during the internet bubble, we might see the same happen for any company with “AI” in their name.
We also might begin to see more AI-related startups begin to IPO.
Riding The Momentum
Anyway, this is all speculation on my part. There’s no guarantee that we will see things play out the way I’ve laid out, or that there will even be an AI bubble.
From an investment perspective, what I’ve been doing is overweighting companies that would benefit from an AI bubble, but that are attractive stocks in their own right.
Nvidia, of course, but also stocks like Apple, Alphabet, and Meta.
If it becomes clear that an AI bubble is forming and valuations become extreme, I’ll probably start to take profits on these positions.
So, I’m riding the AI momentum for now, but I don’t recommend that anyone necessarily follow suit. This is an active, relatively high-risk strategy.
But it’s still in every investors’ interest to follow along to see what happens. Today, a handful of stocks dominate the S&P 500, so even if you own an index fund, your returns will be heavily influenced by movements in these stocks.
I wrote a post several weeks ago about how I don’t think that the large weightings of these megacaps is necessarily a risk for the overall stock market, but if we get an AI bubble and valuations for these companies reach extreme levels, then I might have to reevaluate that view.
I’ll talk more about the valuation for the broader stock market in future posts.
Thanks for reading!