The rush for artificial intelligence investments is already sprouting wildly overvalued companies

Nvidia chief executive Jensen Huang, whose company’s market capitalisation is six times higher than Intel’s.

You gotta love a hype-induced bubble. Having been permanently ensconced in financial markets for more than three decades, I’ve seen a few and there’s another one forming right now.

It wasn’t that long ago – 2019, if memory serves – that Softbank pumped nearly $1bn into driverless delivery vehicle start-up Nuro.ai, which was set up by some ex-Google engineers and was already delivering groceries autonomously in Arizona.

It was the same year I was invited to invest (I didn’t) in Zoox, another autonomous driving hopeful.

The proliferation of suppliers was next. Audi, Tesla, Mercedes, Google and hundreds of others, including Samsung, Denso, Uber, Amazon, Honda and Microsoft, jumped on the driverless bandwagon – or were vacuumed into it.

At an April 2019 Detroit Economic Club event, Ford said it believed its fully driverless cars would be in commercial operation by 2021.

By October 2021, investment in driverless technology had surpassed $US44bn ($67bn).

And this was despite the discovery an autonomous truck company called Nikola had pushed a truck down a gentle hill, filmed it and claimed it was driving autonomously.

Then, in 2022, interest rates started to rise, and just as Nikola’s founder was charged with fraud, the autonomous bubble burst.

Almost as recently, we seen bubbles in ridesharing, the metaverse, blockchain, crypto, EVs, the internet of things (IoT), 3D printing, quantum computing, virtual reality, augmented reality and gamification. For each of these technologies – even though the technology remains – rising expectations and ultimately unrealistic enthusiasm was followed by disappointment and disillusionment.

Investing in technology is not easy because, often, the technology benefits the consumer but not the long-term investors.

Think about the invention of the car. In 1886 when Karl Benz first drove his Benz Patent-Motorwagen horseless carriage past a blacksmith, it would have been possible to predict it was transformational, but it would also have been impossible to predict which manufacturer of this disruptive technology would succeed. In the US alone, there have been 1665 car manufacturers that are now defunct.

Even when new technology changes the world, it can be difficult to pick a long-term winner.

Consider, for example, when television first appeared in the 1950s: More than 90 US manufacturers, including Admiral, General Electric, Magnavox, Philco, RCA, Silvertone and Westinghouse dominated the market.

In 1953 there was even a Radio-Electronics-Television Manufacturers Association, and it reported almost 7.3 million TVs were made in the US. By 1995, the last remaining US television company, Zenith, was sold to Korea’s LG.

Today it’s generative artificial intelligence (AI), and specifically the Large Language Model, that promises to change the world forever. The disruptive and transformational AI

technology that “threatens to reshape the world” is now also the latest investing fad.

is the latest fad’s poster child. At $US40,000 each its H100 chip, which solves the scalability issue for general language AI model creators, is being snapped up in the tens of thousands by the likes of Microsoft and Amazon, which are developing AI-dedicated data centres.

It is said the chip has given Nvidia a two-year headstart over its competitors. Meanwhile, the three times better performance of the chip compared to its predecessor permits smaller AI start-ups to reduce model learning times and speed up their own product launches.

The excitement is, predictably, palpable. Nvidia’s revenue is half that of its competitor Intel, but its market capitalisation is six times higher.

Meanwhile, Intel, whose chips are seen as ill-suited for AI’s needs, has announced it will introduce its Falcon Shores chip for AI computing, with 288 gigabytes of memory, in 2025.

Predicting these bubbles and subsequent deflations is easy. Predicting their magnitude and timing is more difficult.

Just ask Gartner, who produces the annual Hype Cycle of Emerging Technologies, which is widely accepted in the tech community. Back in 2014, Gartner predicted the hype around natural language question-answering systems, wearable user interfaces, including Apple’s iWatch, and autonomous vehicles, were all at or near their peak.

That year Gartner also predicted crypto, mobile health monitoring and big data were already past their respective peaks and descending into the trough of disillusionment.

Generally speaking (pun intended), there is merit in the idea of a cycle of hype that begins with R&D, is followed by start-ups and venture funding rounds, then early adopters, then mass media hype, a proliferation of suppliers and wide adoption.

Peak hype is then followed by negative press, supplier failure and consolidation, and the realisation that far fewer customers than the total addressable market actually adopted the technology despite the breathless headlines that the technology would change the world, reshape industries and usher in huge productivity gains and displace millions of jobs.

The language of “transformation” and “disruption” surrounding the development of AI is not unique. In fact, we have heard it many times before, and even very recently. While the transformational potential of AI may be underestimated (even amid the breathless headlines), the value to investors is likely to be over-estimated.

Nvidia’s share price is now 203 times historical core earnings, which is up from 50 times earnings in October last year, and 38 times sales. In a hype-inspired bubble, however, the multiples can and often do go a lot higher.

If not today, Nvidia’s share price will reach a point where buyers are guaranteed to lose money or at least are guaranteed to lock in pathetic returns. They may get the power of the tech right, but the investment very wrong.

AI-related equity prices will be the product of the same hype-cycle already experienced by investors in so many technologies that preceded AI.

There is always the new, new thing.

Roger Montgomery is founder and chief investment officer at Montgomery Investment Management