This has been a year of milestones. Since mid-May, the iconic Dow Jones Industrial Average has notched its first close over 40,000, the broad-based S&P 500 climbed above the 5,500 level, and the growth-powered Nasdaq Composite surpassed the 18,000 mark. This full-fledged bull market comes courtesy of the euphoria surrounding the rise of artificial intelligence (AI).
Without getting overly complex, AI uses software and systems to undertake tasks that humans would normally oversee. What makes AI so special is the ability of software and systems to learn and evolve without human oversight. If AI software and systems become more proficient at their tasks over time or perhaps learn new skills, pretty much every sector and industry of the global economy would benefit.
Although dozens of companies have benefited from investors pouring their capital into the AI revolution, none stands out more than semiconductor titan Nvidia (NVDA 1.44%).
Image source: Getty Images.
Nvidia has enjoyed a historic operational ramp
Nvidia's calling card to success has been its AI-driven graphics processing units (GPUs). In particular, the company's H100 GPU has quickly become the go-to chip that allows for split-second decision-making in high-compute data centers.
According to the semiconductor-focused analysts at TechInsights, Nvidia was responsible for shipping 3.76 million of a possible 3.85 million AI GPUs in 2023. Businesses have lined up to purchase the H100 to train large language models and run generative AI solutions.
One of the key advantages for Nvidia is that enterprise demand for AI GPUs has handily outpaced supply. Based on the law of supply and demand, the price of a good or service is liable to increase if the supply of said good or service is insufficient to meet demand. Nvidia has been able to rapidly increase the price of its GPUs, which, in turn, sent its adjusted gross margin to more than 78% during the fiscal first quarter (ended April 28).
Despite its first-mover advantages, this market-leading company isn't sitting on its laurels. In the latter half of 2024, it's expected to roll out its next-generation AI GPU architecture, known as Blackwell. This new platform can expedite quantum computing, data processing, engineering simulations, and generative AI solutions, among other tasks.
Beyond Blackwell, CEO Jensen Huang teased its next-gen Rubin platform to the world in June. Rubin will have an all-new central processing unit (known as Vera) and is set to make its debut in 2026. In other words, Nvidia appears to have a path to sustained compute advantages over its peers.
Although Nvidia has absolutely crushed Wall Street's loftiest revenue and profit projections for more than a year, and its operational ramp has been nothing short of textbook, evidence is mounting that it may be in a bubble.
While history doesn't repeat to a T on Wall Street, it does have a tendency to rhyme. Based on three compelling charts, we may have already witnessed a top for Wall Street's hottest AI stock.
Next-big-thing innovations always endure an early-stage bubble
Over the last 30 years, Wall Street and investors have had no shortage of next-big-thing investment trends to chase. The advent of the internet, genome decoding, business-to-business commerce, China stocks, nanotechnology, 3D printing, blockchain technology/cryptocurrency, cannabis, and the metaverse have captivated the attention of new and seasoned investors.
But these trends share a similarity: early-stage bubbles.
Although it's impossible to predict when the music will stop, industry-leading businesses for all of the above trends eventually face-planted.
As you can see in the chart above, some of the buzziest investments over the last quarter of a century have nosedived after the bubble burst for their respective innovation, technology, or trend. In no particular order:
- The networking/e-commerce revolution saw Amazon and Cisco Systems shed in the neighborhood of 90% of their values during the dot-com bubble.
- 3D Systems and Stratasys, which led the 3D printing hype, have never come close to replicating their ascents from their early 2010s and currently sit 97% and 94% below their respective all-time highs.
- Canadian cannabis stock Canopy Growth has plummeted a whopping 99% since hitting its record-closing high.
- Cryptocurrency play Coinbase Global dumped nearly 90% from its peak close before rebounding.
- Social media maven Meta Platforms, which had been the face of the metaverse, plunged by about 80% before rallying to new highs.
The point I'm trying to make is that investors have overestimated the uptake of every new technology or trend for the last 30 years, and there's no reason to believe that AI is going to be any different. History (i.e., the above chart) strongly suggests that lofty growth forecasts for AI aren't going to be achieved, and a period of maturation will be needed for this potentially game-changing technology.
Nvidia's adjusted gross margin won't be sustainable
A second reason to suspect Wall Street's AI leader is in a bubble can be seen in the company's current and forecast adjusted gross margin.
As I noted earlier, Nvidia has undeniably benefited from AI GPU demand outpacing supply. The ability to raise its prices has sent its adjusted gross margin higher by 13.72 percentage points over the previous five quarters.
NVDA Gross Profit Margin (Quarterly) data by YCharts.
But as you can see in the chart above, an adjusted gross margin of around 60%, give or take a few percentage points in either direction, has been the norm over the trailing-10-year period. While this doesn't dismiss that Nvidia has steadily improved its adjusted gross margin over time, it does suggest that a 13.72-percentage-point increase in 15 months is unsustainable.
For instance, Nvidia guided to an adjusted gross margin of 75.5% (+/- 50 basis points) for the fiscal second quarter. Although this is still well above its historic average, a sequential decline of 235 to 335 basis points for the current quarter suggests that margin pressures are beginning to take shape.
During the second half of 2024, Intel will roll out its Gaudi 3 AI-accelerator chip on a broad-scale basis, and Advanced Micro Devices will beef up production of its MI300X AI GPU. Further, Nvidia's top four customers are all internally developing AI GPUs for their respective data centers.
Even if Nvidia's AI GPUs easily retain their compute advantages, overwhelming enterprise demand for GPUs may allow these competitors to thrive and/or reduce data center space for Nvidia's chips. In short, Nvidia's fiscal second-quarter margin guidance looks like a warning that its adjusted gross margin has peaked.
We've seen this before -- and it wasn't a pretty sight
The third damning chart for Nvidia has to do with its valuation.
To be fair, some traditional fundamental metrics suggest Nvidia could still be something of a bargain for long-term investors. For instance, Nvidia's shares ended the July 2 session at close to 34 times forward-year earnings. While this is a higher forward-year earnings multiple than the benchmark S&P 500, it's not unreasonable when you factor in that Wall Street is forecasting an annualized earnings growth rate of 46% for the company through 2028.
However, another valuation metric is undeniably concerning.
NVDA PS Ratio data by YCharts. PS Ratio = price-to-sales ratio.
As you can see in the chart above, Nvidia is valued at more than 38 times its trailing-12-month (TTM) price-to-sales (P/S) ratio. For market-leading businesses on Wall Street, this is one of the highest readings we've ever witnessed. In fact, the reading of 42 times TTM P/S that Nvidia reached in mid-June effectively matched the peak from Amazon and Cisco Systems prior to the bursting of the dot-com bubble.
History has consistently shown us that extended valuations can be powered by euphoria over the short run. But when examined over an extended timeline, market leaders with nosebleed P/S ratios are eventually pummeled. Although there's no concrete way to know ahead of time when a top has occurred, we've seen this scenario play out before, and the end result hasn't been pretty.
While Nvidia has the tools to remain a successful company over the very long term, these historic charts strongly suggest that the most direct AI beneficiary is in a bubble.
