The biggest company in the world and the first to ever have a valuation of over £4 trillion is Nvidia
However, can this valuation be justified even though it is the clear leader in artificial intelligence?
Over the last three years, Nvidia has emerged as one of the stock market's greatest success stories.
Both institutional and do-it-yourself investors now favor this company's stock. Since the launch of ChatGPT in November 2022, Nvidia's share price has increased by about 900 percent, making the semiconductor giant the first £4 trillion company in history and practically synonymous with the artificial intelligence (AI) industry.
However, despite being accompanied by equally explosive earnings growth, these explosive gains have always drawn some skepticism. The idea that Nvidia might be overpriced is gaining traction as critics become more willing to voice their opinions that the AI bubble may be about to burst.
To put things in perspective, Nvidia's market value was £4.06 trillion on October 3.
Karim Moussalem, CIO for stocks at Selwood Asset Management, notes, "That's more than the whole of UK listed companies." "With the help of investment headlines or the announcement of new products, Nvidia can increase its market capitalization by hundreds of billions of dollars in a single trading day.
The price/earnings (P/E) ratio of Nvidia is approximately 30 times its projected earnings and 54 times trailing earnings. Considering the Magnificent Seven that define the AI trade, these aren't all that unusual for large tech stocks.
The investment specialist at Orbis Investments, Rob Perrone, stated, "It's difficult to argue that Nvidia's valuation is completely crazy." "This company has been doubling its revenues and generates 70 percent returns on capital. Wall Street believes it will continue to grow at a rate of 30 percent for some time to come."
From where does Nvidia make money?
Since Nvidia made almost a third as much money as the entire index last year (£84 billion compared to the FTSE 100's combined 192.9 billion, or roughly £260 billion), you might wonder if it matters that it is worth more than the whole UK stock market. It is also expanding much more quickly.
In recent years, Nvidia's revenue and profits have skyrocketed, making it a massive company. Nonetheless, a reasonably grouped group of businesses have contributed to its revenue growth.
In its most recent quarterly results, Nvidia's accounts receivable balance was made up of two customers, totaling 33 percent.
According to Nvidia's reports, two customers, identified as Customer A and Customer B, contributed a total of 39% of the company's Q2 revenue (23 and 16 percent of revenues, respectively). Essentially, one customer accounts for nearly a quarter of Nvidia's revenue, while its two biggest customers account for two fifths.
With Customer A and Customer B accounting for 30% of revenue in Q1 (16 percent from Customer A and 14 percent from Customer B), this revenue concentration is growing.
The combined revenue from Nvidia's next four biggest clients was 44%, which means that six clients make up more than four-fifths of the company's total revenue.
As original equipment manufacturers (OEMs) are likely to purchase and distribute Nvidia chips, these direct customers may represent a more varied end-user base than the statistics alone indicate. However, Colette Kress, the chief financial officer of Nvidia, stated in her Q2 statement that about half of Nvidia's Data Center revenuewhich was 88 percent of total revenue for the quartercomes from big cloud service providers.
Even though Nvidia itself seems to have a loyal customer base that is driving up spending, Perrone asserts that in order to evaluate the long-term viability of the growth projections incorporated into its valuation, you must look downstream.
He states, "It doesn't really seem like a lot of people have figured out how to make money off of this stuff yet (apart from Microsoft)."
Nvidia's circular economy.
The main query, then, is where Nvidia's clientsdirect or indirectget their funding.
"A strikingly circular flow of capital" is what Moussalem calls attention to.
"Nvidia helps finance OpenAI's and other big model developers' acquisitions through equity and structured deals, sells the chips that power them, and then reaps the rewards as every new announcement of larger training runs or data center projects feeds the perception of limitless growth."
For instance, Nvidia's recent investment in OpenAI will allow it to purchase more stock in the company that creates ChatGPT for each gigabyte of Nvidia hardware that OpenAI uses.
FT Alphaville may have provided the best summary of the deal.
How much of Nvidia's £100 billion in chip sales revenue would go back to the company?
"I don't know."
Coreweave is another example, having gone public earlier this year. It purchases Nvidia GPUs and then leases them to businesses, including Nvidia, who wish to temporarily increase their processing power. Apparently, Nvidia has a £2 billion investment in Coreweave. Essentially, Nvidia is simultaneously a Coreweave supplier, customer, and investor.
According to Moussalem, "every press release drives the stock higher, which in turn supports additional capital raises and even larger spending commitments." "This creates a powerful feedback loop that looks more like late-cycle financial engineering and less like traditional business," he claimed. It is also known as a Ponzi scheme or vendor financing.
The merry-go-round of capital amongst AI's major players is "not the most sustainable source of financing," according to Perrone, who is more formal.
In the end, he told BFIA, "you need money coming in from customers, from outside the ecosystem."
Would the price of Nvidia's stock drop?
Moussalem claims that the current state of affairs is "eerily reminiscent of the late-1990s .com boom."
The prevailing belief in unending growth that underpinned that boom came to a sudden halt.
Nvidia's revenue could plummet if a major cloud service provider determines that investing in AI infrastructure is no longer worthwhile. As of right now, there is no indication that they intend to do so anytime soon.
However, early signs are present that could eventually alter that opinion. Right now, AI's economics appear to be in jeopardy. By 2030, AI companies will need to make £2 trillion in revenue to finance the planned expansion of data centers, according to the Bain and Co. 2025 Technology Report. However, the expected productivity gains that AI is expected to produce will still fall £800 billion short of this amount.
Recall that although cloud hyperscalers and Nvidia are (extremely) profitable companies, the AI developers who pay them are not. To encourage adoption, OpenAI, Anthropic, and the majority of other AI start-ups continue to give away the majority of their extremely expensive products for free.
"OpenAI is burning cash," Perrone claims. When Meta (formerly Facebook) had as many users as OpenAI now has, it was making money.
"You need enormous numbers of revenue, not in the hundreds of billions but more like the trillions," said Perrone, in order to justify the amount of investment being made.
For AI to be viable, the number of paying customersthat is, businessesmust grow significantly over the course of the next five years. Nevertheless, why should they, considering that 95% of companies investing in AI aren't seeing any return on their investment, per a recent MIT study?
Nvidia's current valuation is ultimately based on a large stack of cards. For the foreseeable future, it is anticipated that its profits will continue to rise quickly.
Perrone wrote, "When expectations are high, so is risk," in an article about the AI competition's sleepy winners.
According to Moussalem, "we are in the midst of a bubble, and a total retail-driven frenzy."
Even though he is more cautious, Perrone thinks that investing in AI presents "better opportunities" and opportunities than purchasing Nvidia stock. Buy the Ammo-makers: How to Find Value in the AI Wars is one article in BFIA magazine that highlights some of these concepts.
See More AI.
Leave a comment on: Are we overpaying for Nvidia?