investor daily logo

Dotcom bubble holds lessons amid AI frenzy: Fidelity

By Rhea Nath
4 minute read

The investment manager has addressed whether the artificial intelligence rally is demonstrating signs of a nascent bubble, noting parallels between the meteoric rise of Nvidia against that of Cisco in the late ’90s.

Addressing concerns that the current artificial intelligence (AI) rally could be reminiscent of the dotcom bubble of 2000, Fidelity International believes it is not yet time for investors to worry, although similarities might exist.

Particularly, the firm addressed similarities between the recent stock market performance of chipmaker Nvidia with that of Cisco Systems.

Last week, Nvidia reported earnings grew 22 per cent from the previous quarter to US$22.1 billion, rising 265 per cent from a year ago.

It shared optimism for the underlying technology and forecast revenue to hit $24 billion, plus or minus 2 per cent, for the first quarter of fiscal 2025.

Tom Stevenson, investment director at Fidelity International, noted that in the case of Nvidia and Cisco, “the share price trebled in a matter of a few months, then paused for breath for a few months more before turning left up the page as the fear of missing out sucked in the doubters”.

“Both companies’ shares increased sevenfold in less than two years. Nvidia remains close to its all-time high, and what happens next is one of the most important questions in investment right now,” he said.

Highlighting the benefit of Hindsight with Cisco, he explained: “We now know that in the two years after its share price peaked, it fell all the way back to where it started. Between March 2000 and September 2001, Cisco’s shares tumbled from US$78 to US$11, a loss of 86pc for anyone who had bought at the peak.”

Additionally, like the handful of internet-related stocks that dominated the market 25 years ago, there is narrow market leadership in the Magnificent Seven stocks, which account for around half of the S&P 500’s gains so far in 2024, he said.

“Then, as now, investors fixated on the power of the new. In the late 1990s it was the internet that was going to change everything. Today it is AI. In both periods, valuations of companies expected to benefit from the perceived paradigm shift diverged from those of the also-rans operating in the ‘old economy’,” Mr Stevenson said.

Elaborating that a bubble translates to a speculative frenzy, not just overvaluation, he explained that it comprises “a loss of reason” and “requires the abandonment of tried and tested fundamentals”.

In the case of Nvidia, he said the first indicator of a nascent bubble – the widespread adoption of a “new economy story”, looks to be in place.

“The potential for AI to change the world is unproven but you would be hard pushed to find many people who don’t think it matters. Most of us are enthralled and scared by AI in equal measure,” Mr Stevenson said.

Additionally, a generational divide has been observed, driving a wedge between the believers and others, with the believers needing to “invalidate the valuation framework into which old-style sceptics retreat”.

He added: “The easiest way to do this is to invent a new set of metrics that fit the bullish case better. In the dotcom bubble, we were told to dispense with outdated measures like price-to-earnings ratios and look instead at price-to-clicks or ‘eyeballs’.”

Other signs of trouble brewing, he explained, were when prices continue to rise even when news flow turns negative.

“When investors become selectively deaf, you can be sure a bubble is inflating. The recent fourth quarter results announcements from the Magnificent Seven were a mixed bag. Investors are choosing to see what they want to. Another potential red flag,” Mr Stevenson said.

“A final sign that we may be in a bubble is when investors, flush with their stock market gains, look for opportunities in other risky assets. Watch out for new asset classes to emerge, usually investments that offer no cash flows or asset backing but which rely on the ‘greater fool theory’ that someone else will pay an even higher price than you.”

He posited that, if an AI bubble does exist, it remains in its early stages.

“The valuation premium is well short of the level reached in 1999 or even in the early 1970s when the most popular shares were priced nearly twice as expensively as the rest of the pack,” he said.

“We are a long way off that today. Sentiment is not yet universally positive. When it becomes so, it will be time to worry. We’re not there yet.”