Artificial intelligence (AI) has dominated the narrative in the stock market this year. 

It's a big reason why stocks have rebounded even as interest rates have moved higher, and why the Nasdaq Composite is up 27% this year.

However, cracks are already starting to form in that narrative. ChatGPT, the AI chatbot that kicked off the frenzy around generative AI, saw traffic decline this summer, a sign that a significant percentage of its usage came from students looking for homework help. With school back in session, traffic appears to be rebounding, but the fact that its seasonality seems to be driven by the school year bodes poorly for its widespread adoption, something that so many, including top tech CEOs, have envisioned.

Now, there's another more concerning sign that the AI revolution may be overhyped.

The internet rumor mill is swirling with chatter that Microsoft (MSFT -0.12%) is cutting orders for Nvidia H100 AI chips as traffic to ChatGPT has slowed and interest in Microsoft 365 Copilot is weaker than expected.

The rumor is just that, and investors shouldn't overreact to it. Even if it's true, it's unclear by how much Microsoft is cutting its orders, but that scuttlebutt combined with the decline in traffic at ChatGPT shows that the hype around the AI boom could be overdone. Even if it's deserved -- and most tech CEOs seem to believe it is -- it will take time for businesses and consumers to adopt AI.

Right now, we're still in the phase where businesses are assembling their infrastructure to run AI processes such as large language models.

A digitally generated image of a face

Image source: Getty Images.

If you build it, will they come?

At this point, despite all the excitement around AI in the tech sector, there's only one major company that's seen a significant ramp in sales due to the emergence of generative AI.

That's Nvidia. The chip designer best known for its graphics processing units, has blown away Wall Street in its last two earnings reports. If you want evidence that there's demand for AI infrastructure, look no further than the chart below.

NVDA Revenue (Quarterly YoY Growth) Chart

NVDA Revenue (Quarterly YoY Growth) data by YCharts

During a time when much of the industry is struggling with a cyclical downturn, Nvidia has seen revenue explode, going from negative growth in the first quarter to a 101% jump, a surge that's virtually unprecedented outside of sectors like biotech.

Nvidia's second-quarter revenue is also well ahead of its peak during the crypto boom when demand for its GPUs also soared, and it expects even higher revenue in the third quarter.

Nvidia is in the enviable position of selling the "picks and shovels" for the AI gold rush. Big tech companies and others building out data centers and tech infrastructure need Nvidia's tools to run these high-powered AI programs, but just as long-gone prospectors know that purchasing a pick and a shovel is no guarantee of striking it rich, buying Nvidia chips to run large language models isn't a surefire path to profit either.

Microsoft, which invested $13 billion in ChatGPT-creator OpenAI, may be learning that now. CEO Satya Nadella helped fan the flames around generative AI, promising a new race in search when the company launched its ChatGPT-powered version of Bing, and said that the gross margin in search was going to drop forever as the playing field would be leveled and competitors would have to run the fastest AI models.

However, that prediction has so far missed the mark. Relatively little has changed in search since the launch of ChatGPT. Alphabet's Google still retains dominant market share and is generating huge profits. Although Alphabet released its own AI chatbot, Bard, generative AI doesn't seem to have changed the way most people use internet search.

What it means for AI investors

We'll learn soon enough if the rumors around Microsoft are true, but the broader lesson for investors may be to resist the hype around AI.

If generative AI really does take over the world, it won't happen in a bloodless coup overnight. Institutions are already pushing back at what many see as a threat to their livelihood. AI has become a sticking point in negotiations in the Hollywood writers strike, for example, and its eager embrace by students brings up serious questions about what the future of education should look like.

For investors, the interest in AI isn't unwarranted, and the drumbeat of excitement from tech CEOs shouldn't be ignored, but the mainstream adoption of these new tools is going to take time. Past tech revolutions didn't happen overnight. The internet, for example, was a much less-used, more primitive version of its current version in its early years. Even the iPhone and smartphones more generally took several years to displace conventional mobile phones.

AI investors should avoid the hype in the sector, and instead focus on diversifying, keeping valuations in mind, and being patient.

There could be plenty of big winners in the AI sector, but it will take years for those gains to unfold.