This crucial indicator likely signals that the excitement around artificial intelligence (AI) is starting to fade.
Since the start of 2023, no trend has propelled Wall Street’s major stock indexes to new heights more than the rise of artificial intelligence (AI).
The appeal of AI is the ability of software and systems to learn over time without human intervention. This allows AI-driven software and systems to perform tasks more efficiently and can evolve to learn new skills. With a target market spanning most sectors and industries, PwC analysts believe that AI will add a staggering $15.7 trillion in value to the global economy by 2030.
Dozens of companies have benefited from the AI revolution, but none are more emblematic of its success than the semiconductor giants. NVIDIA (NVDA 4.55%).
Image source: Getty Images.
Nvidia is leading the next leap in business innovation
Nvidia’s H100 graphics processing unit (GPU) quickly became the chip of choice for enterprises to run generative AI solutions and train large language models (LLMs). With demand overwhelming supply, Nvidia was able to dramatically increase the price of its H100 GPUs to $30,000-40,000 per chip, roughly two to three times what its main competitors were charging for their AI data center hardware.
The great thing about higher prices is that they have directly benefited Nvidia’s bottom line: Over the past five quarters ending April 28, 2024, the company’s adjusted gross margins increased by nearly 14 percentage points to 78.35%.
Nvidia isn’t shying away from investing in the future, either: The company’s next-generation Blackwell platform, due to hit the market next year, will accelerate computing power in six areas, including quantum computing and generative AI, and will be more energy efficient than the previous generation. Meanwhile, in June, CEO Jensen Huang briefly teased an all-new Rubin GPU architecture, which will run on a different processor (called Vera) and is due to debut in 2026.
The final piece of the puzzle that has helped Nvidia’s market cap increase by more than $2.8 trillion since the start of 2023 is its CUDA platform, the software platform that developers use to build LLMs, which works with the company’s flagship hardware to keep enterprise clients loyal to its solutions ecosystem.
While this appears to be a perfect expansion, Wall Street will get a first-hand look at just how fragile Nvidia and AI as a whole technology are on August 28th.
Nvidia’s very significant numbers could signal the AI bubble is about to burst
This coming Wednesday, August 28th, the Wall Street AI darling is scheduled to report its second-quarter results.
Over the past five quarters, Nvidia has crushed even the wildest analyst expectations. Strong enterprise demand for its AI-GPUs, exceptional pricing power, and little competition have allowed the company to build an order backlog that would be the envy of any tech company.
But the headline revenue and earnings figures don’t tell the whole story when it comes to the Aug. 28 release. Even if revenue and earnings beat analysts’ expectations, there’s another key number that could signal an end to the AI frenzy: Nvidia’s adjusted gross margin. Nvidia’s “adjusted” gross margin doesn’t include the impact of stock-based compensation, acquisition-related expenses and several other costs.
NVDA Gross Margin (Quarterly) data from YCharts.
Following Nvidia’s first-quarter earnings report, Huang and his team issued a second-quarter adjusted gross margin guidance of 75.5% (+/- 50 basis points), which represents a decline of 235-335 basis points from the first quarter.
The 285 basis point decline in the median adjusted gross margin estimate may not seem like a big deal considering Nvidia’s adjusted gross margins have expanded by about 1,370 basis points over the past five quarters, but it’s the reasons behind this forecast decline that are the real concern.
Nvidia’s computing dominance is unlikely to save the company from the inevitable crisis
While demand for Nvidia’s H100 GPUs is undoubtedly strong, it’s the company’s pricing power that has contributed the most: revenue growth has far outpaced increases in cost of goods sold, indicating that pricing power, bolstered by the ongoing shortage of AI-GPUs, is the company’s core driver.
The problem for Nvidia is that it’s not the only option. Advanced Micro Devices (AMD 2.16%) AMD is ramping up production of its MI300X AI-GPU, which is on average 50% to 75% cheaper than Nvidia’s H100. AMD is also not being hampered by issues with early-stage chip manufacturing suppliers, just like Nvidia.
Additionally, Nvidia’s top four net revenue customers are: Microsoft, Meta Platform (Nasdaq: META), Amazonand alphabet — and Intel are all developing AI-GPUs in-house for their own data centers. These in-house developed chips are meant to be complementary but ultimately cheaper and more accessible than Nvidia’s hardware. The companies account for about 40% of Nvidia’s revenue, and each has signaled they will rely less on Wall Street’s AI darlings going forward.
To make matters worse, just over two weeks ago reports emerged that Nvidia’s vaunted Blackwell chips would be delayed “at least three months” due to design flaws and supplier constraints — Nvidia’s inability to meet enterprise demand in a timely manner opens the door for AMD. Samsungand Huawei will take market share.
The biggest increase in NVIDIA’s gross margins comes from the critical shortage of AI-GPUs. But as new chips come to market and its major customers fill up valuable data center “real estate” with their chips, NVIDIA will inevitably find its pure pricing power weakening. The company’s 285 basis point average quarterly adjusted gross margin forecast decline is evidence that enthusiasm for AI is fading.
Image source: Getty Images.
When the AI bubble bursts, few companies will be hit harder than Nvidia.
Looking at the flip side of Nvidia’s Aug. 28 report, history presents another obstacle to the AI revolution.
Since the advent of the Internet 30 years ago, not a single large-market innovation, technology, or trendy trend has escaped an early stage bubble burst. Without exception, investors consistently overestimate the use cases and consumer/business adoption of a new technology or trend, ultimately leading to disappointment, fading euphoria, and a bursting of the bubble.
We have seen this play out in the internet, genome sequencing, business-to-business commerce and networking, housing, Chinese stocks, nanotechnology, 3D printing, cryptocurrency, cannabis, blockchain technology, virtual reality/augmented reality, the metaverse, and much more.
What’s more, we find that few of the companies building AI data centers have clear plans for how they will use the technology to increase revenue and profits: Meta Platforms, for example, has invested over $10 billion in Nvidia’s H100 GPUs but has no immediate plans to profit from its AI data center investment.
The fact that most companies don’t have a clear strategy when it comes to AI is a very clear indication that we are facing the next stage in a long series of bubbles.
This is not to say that artificial intelligence cannot be done. Finally (Keyword!) will significantly change the growth trajectory of American companies, but the technology will no doubt need time to mature.
If the AI bubble bursts as history suggests, no company will be hit harder than Nvidia, and the company’s adjusted gross margins next week should provide further evidence that this bubble is starting to show signs of popping.
