Since the advent of the Internet some thirty years ago, investors have almost always had a buzzy trend or innovation that caught their attention. However, none of these other next-big-thing trends have come close to what the Internet did for corporate America.
But after a long wait, professional and ordinary investors are now enthusiastically convinced of the potential of the artificial intelligence (AI) revolution.
While estimates vary, as you might expect from a disruptive technology, analysts at PwC expect AI to add $15.7 trillion (Yeswith a “t”) to the global economy through various consumption-side benefits and productivity gains around the turn of the century.
No company has benefited more directly from the hype surrounding AI and its seemingly limitless possibilities than Nvidia (NASDAQ: NVDA).
The rise of Nvidia is unlike anything we’ve ever seen
As the page turned to 2023, Nvidia was a $360 billion company on the verge of becoming one of America’s top tech stocks. But by the closing bell on Aug. 28, 2024, it was worth $3.09 trillion. In June, it briefly became the most valuable public company shortly after completing its historic 10-for-1 stock split.
It doesn’t take much digging to understand why Nvidia at one point added over $3 trillion to market value in less than 18 months. The company’s H100 graphics processing unit (GPU) quickly became the chip of choice in AI-accelerated data centers. It is, in effect, the brains that power generative AI solutions, facilitating the training of large language models (LLMs) and driving split-second decision-making in AI-driven software and systems.
Demand for Nvidia’s hardware has completely overwhelmed supply. Producing the must-have AI GPU has given Nvidia a staggering amount of pricing power. While Advanced micro devices (NASDAQ: AMD) sells its MI300X AI GPU for prices between $10,000 and $15,000, while Nvidia’s H100 typically costs between $30,000 and $40,000.
Nvidia’s CUDA software platform has also played a major role in its success. CUDA is the toolkit that developers use to build LLMs and squeeze the most compute power out of their GPUs. CUDA is an indispensable tool that has helped keep Nvidia’s customers loyal to its ecosystem of products and services.
The end result of this seemingly classic expansion was that the company’s reported revenue and earnings completely exceeded the consensus of Wall Street analysts for six consecutive quarters.
But despite this success, one little-known performance metric seems to confirm that Nvidia’s best days are over.
The first consecutive decline in two years for this operating measure portends trouble
As I said earlier this week, I wouldn’t be shocked if Nvidia were to beat Wall Street’s consensus revenue and earnings per share (EPS) estimates for its second fiscal quarter (ended July 28) — which is exactly what it did. It’s not unusual for analysts to be conservative in their estimates and set a low enough bar for companies to meet.
But key figures, such as turnover and net profit, only tell part of the story.
At Nvidia, the only performance metric that tells a more comprehensive story about where things are headed is gross margin. Personally, I prefer to use adjusted gross margin, which strips out acquisition-related costs and stock-based compensation; but either gross margin or adjusted gross margin will work fine for this discussion.
During its fiscal first quarter (ended April 28), Nvidia’s adjusted gross margin expanded to a nearly unthinkable 78.35%. Over five quarters, it’s up nearly 14 percentage points, a reflection of the company’s premium pricing for its AI GPUs.
However, Nvidia also reported an adjusted gross margin of 75.5% (+/- 50 basis points) for the second fiscal quarter in its first quarter report. If it were to reach this range, it would mark the first consecutive quarterly decline in adjusted gross margin in two years.
After the closing bell on Wednesday, August 28, Nvidia reported its long-awaited fiscal second quarter operating results, showing adjusted gross margin declining 320 basis points to 75.15%. While that’s within the range of what the company had forecast three months earlier, it’s on the low end of what was expected.
Additionally, Nvidia’s fiscal third quarter guidance predicts the possibility of further gross margin contraction, with an expected adjusted gross margin of 75% (+/- 50 basis points).
Although adjusted gross margin is still increasing substantial Looking at the situation 18 months ago, it seems that the tide is turning, and not for the better.
Competitive pressure and history work against Nvidia
While Nvidia is selling more H100 chips and CEO Jensen Huang has indicated that demand for its next-generation Blackwell GPU architecture remains strong, the lion’s share of adjusted gross margin growth comes from the scarcity of AI GPUs and the company’s outlandish pricing power.
The first problem is that the scarcity of AI GPUs will diminish over time. AMD has ramped up production of its MI300X, and it doesn’t face the same supply constraints as Nvidia from leading chipmakers. Taiwanese semiconductor production (NYSE:TSM)As external competitors enter the market and ramp up production, Nvidia’s pricing power will steadily diminish.
It’s also highly likely that vendors will want a bigger piece of the pie. Taiwan Semiconductor is in the process of significantly expanding its chip-on-wafer-on-substrate (CoWoS) capacity, which is a necessity for packaging the high-bandwidth memory needed in AI-accelerated data centers. Increasing its CoWoS capacity will likely result in higher costs on Nvidia’s part to incentivize production.
And it’s not just external competition that the AI giant has to worry about. Nvidia’s four largest customers by net revenue (all members of the “Magnificent Seven”) are developing in-house AI GPUs for use in their high-compute data centers. The H100 and Blackwell GPUs, while retaining their compute-performance advantages, won’t be enough to deter these top customers from leveraging their in-house chips and denying Nvidia valuable data center “real estate.”
The other big problem working against Nvidia is history. At no point in the last 30 years has there been a next-big-thing technology, innovation, or trend that has avoided an early-innings bubble. All technologies take time to mature, and artificial intelligence seems to be no exception to this unwritten rule.
While Nvidia has received large orders from its top customers, the vast majority of companies investing in AI lack a clear game plan. Even Meta platformswhich is one of Nvidia’s four largest customers by net revenue, has no plans to meaningfully monetize its AI investments anytime soon.
This apparent lack of direction, combined with the company’s adjusted gross margin guidance, essentially confirms that Nvidia shares have peaked.
Should You Invest $1,000 in Nvidia Now?
Before you buy Nvidia stock, here are some things to consider:
The Motley Fool Stock Advisor team of analysts has just identified what they think is the 10 best stocks for investors to buy now… and Nvidia wasn’t one of them. The 10 stocks that made the cut could deliver monster returns in the years to come.
Think about when Nvidia made this list on April 15, 2005… if you had $1,000 invested at the time of our recommendation, you would have $769,685!*
Stock Advisor offers investors an easy-to-follow blueprint for success, including portfolio building guidance, regular analyst updates, and two new stock picks each month. The Stock Advisor has service more than quadrupled the return of the S&P 500 since 2002*.
View the 10 stocks »
*Stock Advisor returns as of August 26, 2024
Randi Zuckerberg, former chief market development officer and spokeswoman for Facebook and sister of Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool’s board of directors. Sean Williams holds positions at Meta Platforms. The Motley Fool holds positions at and recommends Advanced Micro Devices, Meta Platforms, Nvidia and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.
Nvidia Shares Have Peaked and 1 Understated Performance Metric Proves It was originally published by The Motley Fool
