Calvin Wang| March 24, 2024
Calvin Wang| March 24, 2024
How did Nvidia go from making high end GPUs for hardcore gamers to one of the most valued tech companies in the world? And what this sudden boom could possibly entail…
You might have heard all about Nvidia lately, a GPU making tech company that has everybody and their mothers rushing in to buy their stock. As I’m writing this on March 21st, it’s currently worth 911.41 USD on the market. With a market capitalization of 2.28 trillion, it looks like Nvidia is on its way to overtake the second most valued company in the world, Apple, who has a market cap of 2.64 trillion. (For those who don’t know, market capitalization is essentially the total market value of a company’s stock. So even though Nvidia’s individual stock is worth around 800 USD more than Apple, Apple is still considered a more valuable company.)
How did this happen? I’ll bet that a couple years ago, if you weren’t in the gaming community or interested in PCs, you wouldn’t know what Nvidia was and what they made. So what exactly led to this sudden burst in popularity and value? I’ll tell you what. AI.
AI isn’t just some magical code that programmers develop. It needs to be trained, almost like an animal. You know those annoying captchas that so many websites make you do? You’re basically training an AI’s image recognition for the company for free whenever you complete one. Even so, it is incredibly expensive to train AI! In fact, one of Nvidia’s most popular GPUs for AI training, Blackwell, comes in at 30k to 40k for one. Companies wanting to advance AI at the fastest rate need way more than one! Nowadays, I can search ChatGPT and I can use the bot immediately. Before, you might have had to wait for half an hour. That was because OpenAI didn’t have as many GPUs as it does today.
What does this have to do with Nvidia? I thought they were just a company where hardcore gamers bought high end 3500 USD GPUs! Also, what are GPUs? First and foremost, yes, Nvidia still does make high end GPUs for gamers. Don’t worry. But now, their data center production is what rakes in the big bucks. GPU stands for graphics processing unit, and it’s an electric circuit that can compute mathematical calculations at high speeds.
This is a handy little visualizer by @EconomyApp of Nvidia’s fiscal Quarter 4 2023 income statement. A fiscal quarter is a three month interval a company uses to report their earnings. There are 4 quarters in a year, and Q4 2023 was October 1-December 31, 2023. We can see that for Q4, more than half of Nvidia’s revenue came from “Data Center.” This represents the GPUs, CPUs, and DPUs that Nvidia sells. These processing systems are key to AI becoming as big as it is now.
So Nvidia is the creator of the best GPUs for AI to train on. Still, what’s with the sudden boom? You can find articles from 2 years ago talking about AI and microchips. Why did Nvidia suddenly experience such a huge boost in value now, when they’ve been making these GPUs for a while? Well, that would be due to their Q4 2023 income statement, the same one I just mentioned. Nvidia stock had been steadily rising in value, but its investors didn’t actually know how well Nvidia had been doing before the Q4 income statement was announced on February 21st, 2024. The days leading up to the statement actually had people scared. What if Nvidia didn’t actually make that much this time? What if, after the statement, the stock dropped because Nvidia didn’t live up to expectations? This fear actually led to a small dip (around 7 percent) from Feb 15-Feb 20. But despite the fear, Nvidia prevailed, revealing a total revenue of 6.1 billion and all sectors of the company with increased profits. So, their value spiked, and hasn’t stopped going up since. It’s been a month, and so far they’ve increased in value by around 30 percent. Essentially, this is because of how insanely successful Nvidia has become with their production of GPUs for AI.
It’s no surprise that AI is currently all the rage. I mean, look over at one of your classmate’s laptops during quiet work time and there’s a pretty big chance they’ve got chatGPT pulled up. Everybody has been hearing all about AI and the benefits (and downfalls) it provides and will provide. But AI isn’t only good for when you want to be lazy and don’t want to use your own brain to complete an LEQ prompt. It’s also what has been making the stock market reach record highs lately. The second a company announces an addition of AI to their product, their value increases. This sounds silly, but when wingstop announced an addition of AI to their chatbots at the beginning of March, they reached personal record highs in stock value.
All this sounds great, right?. Stock is doing well, people are profiting. However, this sudden rise in the market is looking a little too familiar…some people are starting to ask if this is really true economic growth due to technological innovation, or just an overinflated bubble, waiting for the perfect moment to pop.
First, what is an economic bubble? It’s essentially where the market faces a quick increase in value, particularly in the trading of assets. It’s usually followed by a “burst,” where the prices drop as quickly as they increased. It’s hard to identify a bubble in real time. Bubbles are also a little hard to define, as many economists have their own definitions with small variations.
So in simple terms, I’ve come to understand it as a situation in the market where a particular industry or product is grossly overvalued, usually due more emotion and excitability based trading rather than trading based on common sense and knowledge of the industry. The assets will trade at prices that are way bigger than than the asset’s intrinsic value. (Intrinsic value is the measure of an asset's worth based on objective risk calculation, in lieu of the current market price.) An example of this would be bitcoin. If you really thought about it, without even doing any risk calculation, what does bitcoin even do? Why would it ever be worth anything if it does nothing? The bitcoin bubble of 2020 showcased an asset that had very low intrinsic value trading at much higher rates. This bitcoin bubble popped, the assets decreased in value very quickly, and in retrospect, the whole rage over bitcoin was very silly.
Is the mania for Nvidia stock and all AI assets a bubble as well? The day after Nvidia released the income statement, the S&P 500 hit record highs. (The S&P 500 is a market cap weighted index of the 500 top publicly traded companies in the US. Basically, it’s a measurement of the value of the 500 most valued companies from the US that anybody could buy into.) The wonderings of a potential “AI Bubble” shares many similarities to the devastating Dotcom bubble of the 1990, and even with the 1929 wall street crash that plunged the US into the Great Recession.
The dotcom bubble of the 1990s is likely unfamiliar to you, since you probably weren’t born until a little less than a decade after it happened. (I remember when it happened though, since I’m goated.) The dotcom bubble started due to the rise of internet-based companies. The web had opened to the public in 1991 and began to be used by the general public in 1993-1994, when websites for everyday use became available. Like most bubbles, it grew due to speculative spending, abundance of capital to invest in internet start-ups, and a failure for the “dotcom” companies to make a profit.
Record amounts of capital flowed in the market and by 1999, 39 percent of all venture capital investments were going toward internet companies. (Venture capital is a kind of private equity financing for start-ups with long term potential.) On March 10, 2000, the Nasdaq index peaked, and major tech companies of the time placed big sell orders on their stocks, leading to a panic from investors and a dip of around 10 percent. (The Nasdaq is an online global marketplace for buying and selling stock.) The investment capital was starting to dwindle with no returns from the profitless dotcom companies, and by the end of 2001, trillions of investment money disappeared.
Like the dotcom bubble, the current “AI gold rush” is based on technological advancement. Both are causing the market to reach record highs, and both are quite emotion fueled. Though unlike the dotcom companies, Nvidia is actually making a profit. When there’s a gold rush, the smartest thing to do is to sell shovels, and that’s exactly what Nvidia is doing. The way I see it, the problem is with non-tech based companies that are benefiting from the AI gold rush. Do fast food companies like Wendy’s and Wingstop really become much more valuable after adding AI to their products, or is that just goofy investing from people that get a little too excited from a mere mention of AI? How much more valuable are car companies after they add an AI chatbot? In my opinion, not at all. In fact, Nvidia and other big high tech companies like Microsoft, Apple, AMD, etc. aren’t safe either. To me, it feels as if all companies need to do is say “AI” and they receive a sudden bump in value.
Nvidia is incredibly profitable right now because they are one of the only companies that are able to make the best GPUs for AI. This is because they have deals with the Taiwanese microchip factories, namely TSMC, the top microchip manufacturing company in the world. TSMC makes the chips for Apple, Nvidia, AMD, Qualcomm, Sony, and more. They aren’t the only one as Samsung and Intel also make chips of the same caliber, but TSMC still outshines them due to their superior process and large numbers of highly skilled workers. Companies haven’t been building factories like TSMC’s due to the sheer amount of money that would have to be invested, but now, the potential profits could outweigh the potential risks. So, there was nothing stopping another group from making a factory on the same level as TSMC except for money, and it looks like there’s quite a bit of money to go around for AI now.
Nonetheless, around 90 percent of the world’s advanced microchips are produced in Taiwan. This is important to factor in, given the tense China-Taiwan relations right now. It’s dangerous for US-based companies to rely so heavily on Taiwan, since if Taiwan were to be invaded by China or imports were disrupted in any way due to political relations, that would be disastrous for the tech companies that rely so heavily on the production of chips in Taiwan. TSMC has acknowledged this, and is currently making a factory in Arizona. This simply highlights how vulnerable the West is to losing such an important manufacturer. If the West were to lose Taiwan, and the current AI craze is a bubble, it would be one terrible pop.
If there were to be a new company invented that can create GPUs/microchips of the same caliber at the same rate, it would decrease the value of Nvidia as well. Banking so much on Nvidia being the only successful company in terms of AI training is, in my opinion, a terrible idea. There will always be competitors rising, and a reason for many crashes in the past were due to people thinking that one industry or one company would hold all the profits forever and keep rising in value, when in reality it doesn’t play out like that. If Nvidia announces any kind of decrease in profits next quarter, no matter how small, I would bet money that it would lead to some kind of crash in the tech sector.
Right now there are protections in place for a potentially devastating market crash. If the market were to drop suddenly and drastically, a market-wide circuit breaker would probably be triggered. If the S&P 500 falls 7-13% from its closing price the previous day before 3:25 PM, the stock market closes immediately despite how far it may be from closing time. This allows investors to calm down and hopefully regain their senses.
Fingers crossed, even if there is a bubble, people are able to regain their senses quickly and the crash won’t be disastrous for the everyday American. Unfortunately, that looming threat of a recession persists, and perhaps a potential bubble burst in the market might just be what kicks it off. We’ll just have to wait and see.