AI Is the Talk Of Davos. Is That A Sell Signal?

Investors have already skimmed the froth off the hype, especially for smaller companies as it becomes apparent that AI is a technology already dominated by the biggest of the big. 

So, yes, Davos Man is late to the AI party - and selling to him/her/they would definitely generate profits for those who got in earlier. JL 

James MacIntosh reports in the Wall Street Journal:

The mood of the global elite meeting in Davos, Switzerland, is a useful indicator for investors—as long as they do the opposite. This year it was impossible to move in the snow-blanketed resort without having artificial intelligence pushed at you. (But) Davos Man (and increasingly Woman) is late to the party as shares in smaller AI companies that had been big beneficiaries of the first-half boom have collapsed. Investors can’t just bet on AI in general because there are too few listed AI stocks to allow proper diversification when there are likely to be huge winners and losers.

The mood of the global elite meeting in Davos, Switzerland, is a useful indicator for investors—as long as they do the opposite. 

When the elite are depressed, buy. When they’re positive, sell. When they’re focused on crypto, as in 2021, get out. This year it was impossible to move in the snow-blanketed resort without having artificial intelligence pushed at you.

Does this mean the AI excitement is overdone? Anything AI-related boomed last year following the launch of ChatGPT, so Davos Man (and increasingly Woman) is a little late to the party. Investors thinking about buying now have also missed some stunning gains, as AI stocks soared from January to June, and some of the biggest have carried on up, albeit at a much slower pace.

I’m inclined to think it is hyped, as I said in June; since then there has been more discrimination by investors, with some of the smaller AI stocks pummeled. But then I’m always cynical about the crowd.

 

To make a proper assessment, we need to consider both the reality of AI and what is already priced in.

“It is real what is happening in AI right now,” said Nicolai Tangen, head of Norway’s $1.5 trillion sovereign-wealth fund. “You speak to Sam Altman and you see what’s happening to their business, and you speak to 

Microsoft

, and so it is happening.” But in spite of this he has sold off the fund’s overweight position in Microsoft and the rest of the Magnificent Seven big tech stocks.

The reality distortion field in Davos is strong, but companies are increasingly excited about the prospects for productivity gains in their businesses from using AI.

BlackRock Chief Operating Officer Robert Goldstein is one of the most bullish. He said large language models—the technology behind ChatGPT—would have a phenomenally quick return on investment.

“I believe that the next year, 2024, is going to be a year when you start getting return on investments in a very tangible way from these technologies,” he said.

He said BlackRock staff will be given AI “co-pilots” to help them do their jobs better—rather than to replace their jobs. One example: First drafts of reports could be produced in minutes instead of days, before being reviewed by humans.

Daniel Pinto, president and chief operating officer of JPMorgan, expects a slower payoff, but at two to three years it’s still very fast.

“Technology’s probably the area that will get most of the benefit—20% more, 30% more output. But it will take time, it will take a lot of training to understand how they [programmers] can use it.”

These behind-the-scenes uses will be felt by employees much earlier than they are rolled out into consumer-facing applications that might help boost revenue, Pinto said. His view is echoed by Bejul Somaia, partner at the venture-capital firm Lightspeed, who thinks the last place AI will make itself felt is for customers of regulated industries such as finance and healthcare.

Pretty much every company is at least experimenting with AI. This explains the stock-market excitement about the providers of AI services to business, most obviously Microsoft, partner of the ChatGPT maker, OpenAI, and of the hardware used for AI, most obviously the chip maker 

Nvidia

.

Is the excitement already priced in? At the start of June, I argued that a minibubble had developed, with indiscriminate buying of any AI-related stock.

 

Since then investors have become more selective. Nvidia carried on up, albeit at a slower pace, as its earnings powered ahead of Wall Street’s already superbullish forecasts. But shares in smaller AI companies that had been big beneficiaries of the first-half boom have collapsed.

Big losers in the second half include 

SoundHound AI

, down 60% since June, 

C3.ai

, down a third, and the robotics company 

Symbotic

, which peaked a month later before dropping around 35% from the end of July.

This is good news for investors thinking of getting in now, as it suggests less-frothy pricing (if still not cheap: Symbotic trades at about 250 times forecast earnings for this year). 

Meanwhile, some other providers of AI services to companies have failed to gain traction, with 

International Business Machines

 stock lagging behind the S&P 500 last year and trading at less than 17 times forecast earnings.

 

Nvidia is no longer expensive compared with earnings, because earnings have risen so fast and are expected to soar again this year. Microsoft hasn’t yet had the earnings boost from AI that investors anticipate, and at 32 times forecast earnings is more expensive than any time from 2002 to 2020. 

None of this feels obviously wrong. There is clearly money to be made from selling picks and shovels to those playing with AI, and it might last a long time, helping Nvidia and Microsoft. Equally, there are plenty of ways of applying AI in spite of all the unpredictability and problems that large language models still have, such as inventing things and presenting them as facts.

But to make money from here I’m in the same position as I was in June: Investors can’t just bet on AI in general because there are too few listed AI stocks to allow proper diversification when there are likely to be huge winners and losers. Altman, CEO of OpenAI, seemed to be everywhere in Davos, but even its early lead in large language models is far from guaranteed to last.

The minibubble of the first half of the year has partially deflated; valuations aren’t extreme, but expectations are still high. Contrarians who want to bet against the “Davos Consensus” should steer clear of AI stocks. Gamblers can pick individual winners and losers. Everyone else should hope that AI does indeed boost productivity, helping earnings in the welter of big companies that are adopting it and expanding the economy.