Attention, Nvidia investors: History shows that this is the first trait that all innovations follow

About 30 years ago, the advent of the Internet and its general proliferation changed the arc of American business growth forever. Rather than relying on concrete operating models, e-commerce has opened international doors for businesses of all sizes.

Since this leap forward in the mid-1990s, there has been no shortage of touted innovations, game-changing technologies and popular trends, including genome decoding, business-to-business trading, US housing, nanotechnology, Chinese stocks, 3D printing, cryptocurrencies, blockchain technology , cannabis, augmented/virtual reality and the metaverse.

However, it’s the next big thing — the rise of artificial intelligence (AI) — that has Wall Street and investors seeing dollar signs.

Image source: Getty Images.

Nvidia is spearheading a $15.7 trillion addressable AI opportunity

Although estimates vary, as you’d expect with any hot innovation or technology, analysts at PwC estimate that artificial intelligence will add $15.7 trillion to the global economy by 2030 through a combination of productivity improvements and benefits from the consumption side. If PwC’s predictions are remotely in play, there would be several big winners from the AI ​​revolution.

No company has been a more direct beneficiary of the euphoria surrounding Wall Street’s next major innovation than the semiconductor mainstay. Nvidia (NVDA -1.66%)whose graphics processing units (GPUs) have quickly become the standard in AI-accelerated data centers.

According to analysts at TechInsights, Nvidia accounted for about 98% of GPUs shipped to data centers in 2022 and 2023. The H100 and its successor GPU architecture, Blackwell, will be counted on to train large language models to oversee solutions Generative AI, and make the split-second decisions needed by AI software and systems.

In addition to being the preferred supplier of hardware powering high-computing data centers, enterprise demand for Nvidia chips has overwhelmed supply. Nvidia has been able to raise the price of its chips to two to four times that of the competition, which has led to significantly higher adjusted gross margin over the past six quarters.

In many ways, Nvidia’s rise has been a role model for other innovators. However, history has a different story about what’s next for Wall Street’s AI darling.

Caveat emptor: Next-big innovations share this trait

Most upcoming trends on Wall Street promise a large addressable market and the ability of companies directly involved in these innovations, technologies or trends to benefit. Unfortunately, history has shown that the one trait that every trend in vogue since the mid-1990s shares is an early bubble-bursting development event.

It’s not uncommon for professional and everyday investors to become wide-eyed with excitement when large dollar figures are thrown around a hot trend. This is what often allows emotions to get in the way of better judgment.

Although some of the next big trends I mentioned earlier have gone on to become wildly successful, including the Internet and business-to-business commerce, every important innovation takes time to mature. Without fail, investors overestimate how quickly consumers and/or businesses will adopt a new innovation, technology or trend, leading to eventual disappointment and the bubble-burst event.

On the surface, it would appear that demand for AI hardware is incredibly strong. Orders for Nvidia’s H100 are pending, while the Blackwell chips, which are scheduled for initial shipment early next year, are believed to be sold by 2025.

More than that, customizable rack server and storage specialist Super Micro Computerwhose servers incorporate the Nvidia H100 GPU, posted a 110% year-over-year sales increase in the fiscal year ended June 30 and expects another 87% sales increase in fiscal 2025.

But dig beneath the surface-scratching headlines and you’ll find that year overwhelming most companies have no idea how they will generate a positive return on their AI investments anytime soon.

For example, Meta platforms (META 0.19%) is spending about $10.5 billion to acquire H100 GPUs from Nvidia with no near-term plans to monetize AI. This is a similar approach we’ve seen Meta take with its metaverse initiatives. This lack of a game plan demonstrates how early we are in the AI ​​development cycle and how easy it will be for lofty growth expectations to fall short.

Based on what history tells us about future innovations, it’s not a matter of ifbut When the AI ​​bubble is bursting.

Image source: Getty Images.

Historical precedent isn’t the only headwind against Nvidia

The problem for Nvidia shareholders is that there’s more to worry about than historical precedent.

For example, competition will come at Nvidia from all angles going forward. Even though the H100 and Blackwell GPU platforms will retain their computing advantages over external competitors, these competing chips are considerably cheaper and are ramping up production at a time when Nvidia can’t meet overwhelming enterprise demand. This is a recipe for Nvidia to lose market share and lose some of its otherworldly GPU pricing power.

And Nvidia doesn’t just have to worry about outside competitors. All four of our top net sales customers are developing AI-GPUs in-house for use in their data centers. This includes Meta Platforms, which develops the Meta Training and Inference Accelerator. Even if these top four customers, which together account for about 40% of Nvidia’s net sales, use their chips in a purely complementary way, it will reduce the chance of Nvidia hardware finding its way into AI-accelerated data centers.

We also saw the first sequential quarterly decline in adjusted gross margin in two years. The adjusted gross margin of 75.1% reported during the fiscal second quarter — down 320 basis points from the quarter ended April 28 — indicates that these competitive pressures are affecting the pricing power of Nvidia and mitigates the AI-GPU shortage that has driven prices. bigger first.

While it’s impossible to predict exactly when the AI ​​bubble will burst and when the euphoria will wear off, history has an unblemished record when it comes to the big future innovations of the past three decades. Caveat emptorNvidia investors.