Artificial-intelligence stocks have taken off this year, and the comparison seemingly every expert wants to make is to the late 1990s dot-com bubble. Every time I hear the connection, I figure the person making it is either 1) young or 2) forgetful.
A.I. is a relatively new industry and while some technical indicators make the dot-com analogy valid, the best comparison for what is happening now — and what comes next — is to look at its forerunners, specifically to the PC boom of the 1980s.
To see why, let’s look at where the dot-com era falls flat in a side-by-side comparison. The dot-com bubble started building in the mid-1990s, when everything about the internet was heating up and taking over mainstream consciousness.
The growth of the internet — and the plain-to-see idea that it would impact virtually every aspect of daily life — generated buzz that inflated the shares of practically every company with a website, even if they weren’t “internet companies” at all. With everyone clamoring for internet investments, new companies with little or no track record or business plan easily raised money through IPOs. The get-rich-quick internet helium made everything float, fueling a broad-based rally; internet stocks flew highest, uncoupled from reality.
Those high-growth, no-dividend new companies got additional jet fuel from the Taxpayer Relief Act of 1997, which had dividends taxed at the same rates as regular income, creating a powerful incentive for investors to favor low/no-payout stocks over those with significant dividends but bigger tax bills.
The A.I. rally, in contrast, so far hasn’t been broad; most companies whose stocks are benefitting from the evolving technologies aren’t true “artificial intelligence companies.”
Even Nvidia
NVDA,
the undisputed stock-market winner of A.I.’s rise, is more a maker of GPUs (graphics processing units) that benefits from the new technology than it is “an artificial intelligence company.”
Likewise, the top holdings of the WisdomTree Artificial Intelligence and Innovation Fund
WTAI,
are classic tech stocks and semiconductor makers that are likely beneficiaries of the A.I. revolution, but they’re not a pure play on the future of artificial intelligence.
This is where the comparison needs to shift to the 1980s, when the computer industry moved from mainframes to PC, an evolution that started years earlier but that the mainstream public only woke up to when it first saw how lives could change if a computer was brought into the home.
The market saw the PC as liftoff to an evolution that would “change everything.” (It did, too, as it was the first step leading ultimately to the move toward both the internet and smart phones which, in turn, led to the changes we’re seeing now.)
In the mid- and late 1980s, Wall Street zeroed in on Wang Laboratories, Prime Computer, Compaq and others that were poised to be the big winners. Microsoft was largely overlooked because the public was fascinated with the hardware and not the operating systems; Apple at that time was held back by a legacy of having started by making computer kits sold out of a garage.
Where the similarity to the current situation with A.I. is most striking: Investors recognized that the PC was going to change the world, but had no idea who would win big with it.
“You can see the revolution coming and assume somebody is going to make money off of A.I, and for a long time to come,” said Ben Inker, co-head of the asset-allocation team at GMO, in a recent interview. “Meanwhile, most of the companies who look even a little bit like they could be those big winners are already trading at levels where they are priced as if they’ve won when they haven’t made any money yet.”
That’s where the 1980s comparison holds too, as investors chased Wang Labs and other PC hopefuls through astronomical prices and into catastrophic losses.
“ Focus on baskets of stocks, rather than shooting for the moon with one or two individual names. ”
Barry Ritholtz of Ritholtz Wealth Management said in an interview for my Money Life with Chuck Jaffe podcast that he expects the current A.I. evolution to play out like the PC situation because the ‘80s/’90s were “notorious for having lots of losers but a handful of home runs.”
In those conditions, Ritholtz said, investors should focus on baskets of stocks rather than shooting for the moon with one or two individual names. He added: “It’s the closest thing to venture investing in the public markets. You buy a basket of [100 stocks], 60 are going to go out of business, 20 are going to do nothing but a dozen are going to do well and a handful will just explode and make up for all of the losers. “A.I. is probably going to be something like that.”
The good news is that where technology cycles used to take 15- to 20 years to play out, they are faster now, just as the evolutionary cycle of technology has sped up.
But for investors looking at the A.I. revolution as a springboard for their portfolio, they should pay a visit to the graveyard of those early computer companies, and think about just how hard it was to identify not just which companies but which technologies would win out.
Read: And thus the ChatGPT backlash has begun
Inker noted that the early PC investors would have been better off going for the disk-drive makers rather than the computer manufacturers, at least until the point when data storage became a commodity business and the prices cratered. He expects A.I. to go through several phases, most of which won’t be obvious until they pass.
In the 1980s, for example, behemoths IBM and Xerox spent a fortune trying to stave off the competition and keep pace with the changes. The spending that today’s tech giants such as Apple
AAPL,
Alphabet
GOOG,
and Microsoft
MSFT,
are doing when it comes to A.I. may help them weather the changes ahead, but it’s much too early to assume that current titans can’t be taken down in a long evolutionary cycle like the one we see brewing now.
More: George Soros: ‘I am instinctively opposed to AI, but I don’t know how it can be stopped.’
Also read: It’s all systems go for stock market bulls
Read the full article here
