Artificial intelligence is hot, will this heat burn investors?
Text | James Early
Pitchbook noted that venture capital investment in generative AI like ChatGPT has grown by 425% since 2020, from $500 million to more than $2.1 billion.
Of the roughly $350 billion in total assets in the global venture capital industry, about a quarter are expected to be invested in artificial intelligence. About 40% of companies are using AI or are exploring how to use it.
According to a query by Crunchbase, there are now 27,505 AI companies in the world, although not all of them are pure AI business companies.
AI: A crowded party. IMAGE CREDIT: CRUNCHBASE
ChatGPT is already the fastest-growing app ever, surpassing 100 million users in two months.
Whether you're a private or public market investor, AI seems like an obviously good investment (and maybe a good investor: an AI Powered ETF (AEIQ) that uses AI to choose to invest in was twice as profitable in January as the S&P 500).
But is that really the case?
Is biotechnology overrated?
To answer this question, biotechnology may give us some clues. Like artificial intelligence, biotech has a huge long-term bull market: the world's population is getting older, richer, and spending more wealth on medical technology – which is improving, by the way. Better health care is like floor heating: once you have it, you don't give it up.
Chart of health costs as a share of GDP in high- and lower-middle-income countries
In contrast, according to the World Bank, the United States spent 19.7% of GDP on health care in 2020 and 18.3% in 2021.
Since its inception in 2006, the S&P Biotech Stock Index has returned about three times as much as the S&P 500.
Biotechnology represents a breakthrough, or a series of breakthroughs, for mankind. This is the positive side of the general trend, and it is also the side that investors pay special attention to.
But the other side of the coin is that economics is a recursive social science. This means that if a great trend — and "great" means it adds value to our world — if too many people join in, it could turn out to be a bad investment. Most companies that participate in megatrends fail, and most investors perform mediocrely.
What usually happens is that the hype is growing, the standards getting lower, the projects getting worse and worse, and the money getting dumber.
This is a slogan that is particularly correct in the general trend.
Source: DESPAIR.COM
So, is AI as overrated as biotechnology?
Biotech is hot in 2021, thanks to low interest rates (which push up the present value of cash flows expected in the distant future), the dizzying "Everything Bubble" market, and arguably the "halo" of the coronavirus savior.
With enough demand, supply will emerge – and it will continue to emerge: according to IBISWorld, the number of biotech companies has grown at an annual rate of 7% since 2018, but the market size of the industry has grown by only 2.6% per year.
Not only that, but oversupply in the biotech sector is especially stupid due to a certain "ceiling" in terms of logistics regulation. As my friend Leon Tang, a biotech analyst, likes to point out, there are more than 3,000 biotech companies that want their products to be one of about 50 new drugs approved by the U.S. Food and Drug Administration each year.
This seems like a 1 in 60 chance, but it's slightly better than that, because not every biotech company is fighting for approval every year. However, it is still very sad that the market has not noticed these odds.
If there are 3,000 people attending an event that can only accommodate 50 people (or even hundreds), most people can only stand outside the door.
AI bubble: seems inevitable
Sometimes, the fundamentals of the economy are pushed down by greater fool theory. It's silly, but let's not hate it, because the flip side of silly theory — irrational avoidance — creates underpricing.
Fortunately, in the long run, the price of things returns to the value they bring to the economy.
Markets that do not return well to economic fundamentals (e.g. perpetually obsessed with speculation, insider trading, sudden or arbitrary policy changes, etc.) do not add value to their society – which should be the benefit of the market from a social point of view.
So from a rational point of view, a biotech bubble-like rise is bad, but its reversal is good, even if investors who follow the trend don't think so.
The biotech bubble comes and goes.
Data from YAHOO! FINANCE
Investors in the dot-com bubble felt the same way. Yes, some of the eventual big winners emerged from the ashes of the tech bubble (Amazon's stock price fell 95% at one point, but then rose 58,000%). But overall, the dot-com bubble is just a bigger version of the biotech bubble, one of the biggest investment bines the world has ever seen, and an empty joy for investors.
The dot-com bubble has kept investors' declines and gains almost flat before and after the bubble.
Data from YAHOO! FINANCE
For individual investors, this could be worse than empty joy. Brad Barber and Terrance Odean's academic evidence suggests that individual investors tend to buy when the market peaks, while Dalbar Research has found that individual investors tend to underperform market indices.
Can I still make money on AI?
The negative view — whether it's artificial intelligence, biotech, tech, or tulips — goes something like this: Maybe hot trends and the resulting bubbles aren't kind to investors, but new industries tend to be winner-take-all — there are only a few winners and plenty of losers here — so I just have to identify the ultimate winners and buy them.
This is certainly doable – if you really buy right, it will make you very rich. However, this sentence is followed by a twist, and a twist of turns.
You might guess the first "twist": you don't have much to win.
Few things have changed the world as much as cars, especially in the United States. But in U.S. history, more than 3,000 automakers have come and gone. Of course, some of the losers were acquired by the winners, but in general, if you bought any particular car manufacturer in the meantime, you are likely to lose money.
Many people know the story of the American gold rush of the mid-19th century: California miners made an average of $10-15 a day (a generous number if we count everyone who tried). But the ones who really make money are those who sell nuggets to those who want to get rich
Although some scholars have expressed doubts about the extent of the Dutch tulip mania in the 17th century, whoever paid the price of a mansion for a tulip bulb may regret it.
One might say that this is why the bubble was created in the first place, or at least exacerbated it. After all, instead of pickaxes and shovels, the modern-day gold rush fuels the poor, overpriced versions of the "winners" they want to buy.
But please don't let me discourage you. If you're really good at identifying future winners, you definitely should. This is a huge comparative advantage (evidence suggests that only about 10% of individual investors are actually good stock pickers).
What will be the ratio of losers to winners in AI?
If you're not sure, but still want to enter the market – this is what we call a "turning point" above – then we need to discuss histograms.
If you've ever bought a pack of Pokemon cards (or baseball, soccer—or run any venture capital fund), you'll understand that most of the cards in a pack are useless. In fact, many packages are useless or nearly useless. So anyone who wants to invest by buying a Pokemon pack — while it might be a silly thing to do, we're just using it for theoretical purposes — needs to consider two histogram questions:
1 How many packs do I need to purchase before I buy one or more "winners"?
2 Does the Winner Card Rollup feature I get offset the total defeat of my Loser Card? That is, will my winner card pay for my loser card?
It's a bit unfair to equate Pokemon card pack investing with venture capital, because buying Pokemon card packs randomly doesn't require any skill, and venture capital at least tries to use tricks to distinguish between good and bad businesses.
In contrast, index investing is closer to Pokemon card pack investing, although most index investing is not designed to make big money by itself.
I'm usually skeptical of VC return numbers, and a look at a few VC firms shows that their "average" annual returns range from the stock market to 57% – but VCs are more like a Pokemon bag effect: Correlation Ventures found that 65% of VC deals lose money, and only 4% have a return on investment of 10 times or more. In fact, Institutional Investor magazine notes that VC funds need to make more than 500 investments to catch a super winner.
Venture capital: More deals attract more winners.
Image source: INSTITUTIONALINVESTOR.COM
What are the benefits of making so many transactions? Just look at the white square in the middle: it's higher than among VC funds that have made 500 investments (though note that these are figures before management fees, which may charge a 2% to 20% management fee, which can be adjusted up or down depending on how popular the VC and the particular VC fund are at the time). In addition, the spread of earnings is much tighter than I expected.
How to win the AI war?
We have already discussed:
In a "gold rush", real money is made from people who want to make money in the gold rush.
The value of the tech bubble is supported by the people who participate in it.
The biotech bubble happened because people put more money into biotech stocks, not because biotechnology added more value to the world.
In our society's transition to motorized vehicles, most companies (and their investors) that try to participate in this transition have failed.
However, everything is not as negative as it seems. What the author really wants to say is:
Investors often assume that hot new things that drive human progress automatically become a great investment, but too many investors think this way that these hot new things may end up with disappointing endings.
Superwinners in megatrends do create wealth for a few and a handful of investors, but there are usually hundreds or thousands of losers behind the superwinners.
Diversification by investing in funds is possible, but there is also a high risk that the fund's returns will be diluted by the same losers.
When the market realizes that companies in most popular sectors will fail rather than succeed, the sector's share price will fall.
Of course, funds can earn good returns by investing in popular projects. Individual investors can too, which often happens. If you're an investor in AI, then I hope you can make money from it too.
My goal is not to dissuade you from investing in AI. There is no doubt that artificial intelligence will change the world. Precedence Research predicts that its market size will grow from $120 billion today to $1.6 trillion by 2030. The field of AI will undoubtedly produce numerous millionaires and billionaires.
But if history is any guide, the truth is that it leaves behind many failed companies and investors who suffer losses along the way. ■
This article was translated from https://www.forbes.com/sites/investor-hub/2023/03/04/daydream-believers-read-this-before-you-invest-in-ai
Forbes China exclusive manuscript, please do not reprint without permission
Image source: Pinterest
Follow the Forbes WeChat public account
Never miss the news
▽
The WeChat public account has recently been revised
⭐ Star "Forbes"
Get the latest push as soon as possible
Long press the picture to scan the code to download the Forbes Chinese APP