The AI Bubble
Part I: Detecting A Mania
At present, collective social memory seems not merely imperfect, but downright amnesic.
By Aiden Singh, May 28, 2024
By now it’s a common occurrence: when new acquaintances learn of my profession, they frequently follow up by asking for stock tips or enquiring about my thoughts on specific holdings in their portfolios. My response is usually to refrain from providing advice because there’s too much context around any given trade to be explained concisely.
Still, these interactions tend to prove quite insightful for me, a sort of anecdotal measure of what’s going on in the world of retail trading. Indeed these conversations were key to helping me detect the presence of a speculative bubble in 2020-2021: nearly every query at the time came from someone holding a combination of cryptocurrencies, marijuana stocks, renewable energy and electric vehicle (EV) stocks, non-fungible tokens (NFTs), and meme stocks. Coupled with other indicators such as the intensity of financial product marketing to the general public, an increase in financial fraud centered around penny stocks, and the meteoric (and ultimately unsustainable) rise of figures such as Sam Bankman-Fried (SBF), it was clear that a speculative bubble was inflating.
Today, a similar phenomenon is underway.
Indeed some of the objects of today’s speculative mania are the same as those of the previous one. After a ‘crypto-winter’, digital tokens are experiencing a spring rejuvenation, aided by the Securities and Exchange Commission’s (SEC’s) decision to allow for the trading of spot bitcoin exchange-traded products (ETPs). Bitcoin, the most famous cryptocurrency, has roared back to its 2021 heights near $68,000 after plunging to $15,000 in 2022. And with the rise in bitcoin’s market price has come a new stream of individuals and companies seeking their fortunes in the imitation currency. For example, reports indicate that some students looking to get in on the action have been dropping out of university to launch crypto startups. And medical technology company Semler Scientific jumped over 30% in a day on news that it had adopted bitcoin as its primary treasury reserve asset.
Meme stocks, too, have experienced a revival of interest. On Sunday May 12th, an X account reportedly belonging to Keith Gill, who played a central role in setting off the 2020-2021 meme stock frenzy and is known on social media as ‘Roaring Kitty’, tweeted for the first time since 2021. That tweet sent shares in struggling enterprises GameStop and AMC into dizzying rallies, with the former surging over 200% in two days and the latter rocketing more than 100% over the same period. So beleaguered are these firms that, in an effort to shore up its balance sheet, GameStop has rushed to sell additional stock during this surge in its share price. Yet some speculators, undaunted by the 2022 crash, see fit to trade their holdings of cash and all the potential future streams of interest or revenue which that cash could lay claim to, in exchange for claims to these companies extremely tenuous streams of future profits in a gamble that they might be able to pass them on to some other speculator in short order.
But some of the speculative favorites during the last bubble appear not to be of such keen interest at the moment. And their fates offer important lessons for those so enthusiastically seeking quick fortunes in the market today.
Electrical vehicle (EV) companies had been among the most hyped stocks during the 2020-2021 bubble. Consider shares in EV start-up Fisker. At their peak in 2021, shares in the upstart car manufacturer traded above $25, giving the company a market capitalization over $12.5 billion, despite it never having generated a profit. The mis-pricing can be attributed to a mistaken market belief that mass adoption of EVs was imminent and companies like Fisker would soon become winners in a world without internal combustion engines. Today, the company is struggling to stay alive, its shares are trading at 7 cents, and the Wall Street Journal recently ran a piece detailing the company’s misfortunes.
Or consider shares in EV maker Rivian, which went public to great fanfare in November 2021. At the time, the market assigned the company a value of over $100 billion, with its stock trading at over $130. Today its shares trade at approximately $10 and the company has yet to produce a profit.
But, more than any other company, the darling of the EV craze was undoubtedly Tesla. In 2021, shares in the company hit a high of over $1,200. The market had assigned the company a valuation higher than the combined market capitalizations of Toyota, Mercedes Benz, Ford, General Motors (GM), and Volkswagen, which owns, among other brands, Audi, Lamborghini, Bentley, and Buggati.
At the time, one of Telsa’s most zealous proselytizers, Cathie Wood and her investment firm ARK Invest, talked up the possibility that we would all, in some not too distant future, be shuttled about by a network of autonomous Tesla taxis. Her firm set a 2025 price target for the stock of $3,000. And Woods dismissed the idea that the market may be in a bubble. Fast forward to May 2024 and fully autonomous vehicles remain an as-yet unsurmounted technological quandary for car manufacturers, networks of robo-taxis are yet to roam the streets, and Tesla’s shares trade at what would be roughly $530 absent a 3-for-1 stock split in 2022, less than half what they traded at during their 2021 peak.
Brief reflection at the time would have made clear that the valuation assigned to Tesla at its 2021 heights - greater than the combined valuations of those five rival manufacturers noted above - really only made sense if you believed Tesla was on the cusp of absolutely dominating the roads. This would have required EV adoption to become widespread in a short time-frame, that established automakers wouldn’t develop their own EVs to challenge Tesla, and that new EV manufacturers would not emerge to compete for market share. In the event, ubiquitous adoption of EVs has yet to occur - beset by a lack of charging infrastructure and concerns over range anxiety -, established players have rolled out their own EV models, and new entrants into the EV industry have ripped market share from Tesla, most notably in China.
The fates of Fisker, Rivian, and Tesla offer a cautionary lesson in misplaced market hype around the imminent adoption of some new technology. Speculative bubbles have historically been set off when some technological innovation or new financial product (whether truly world-changing or not of any actual value) is coupled with a widespread belief that riches are to be had by investing in the new technology zealously enough. Such was the case with electric vehicle stocks during the 2020-2021 bubble, tech stocks during the dot-com bubble, scrips in Alexander Hamilton’s new national bank in 1791, and numerous other speculative manias which have taken hold and fizzled out over the centuries.
Today the innovation, or supposed innovation, at the heart of a speculative fever is artificial intelligence (AI), with Wall Street and retail traders crowding into the stocks of companies which, however flimsily, portray themselves as promulgators or beneficiaries of the technology. And the weakness of the arguments in favor of AI stocks today are reminiscent of the infirmity of the arguments in favor of EV stocks yesterday.
The present AI hype appears to have been set off by the idea that large-language models (LLMs) of the sort that underpin ChatGPT, initially released in November 2022, are a major breakthrough heralding the imminent arrival of an artificial intelligence induced industrial revolution. Indeed the title of a May 2023 article in the Guardian exclaimed, “AI ‘could be as transformative as Industrial Revolution’”. A November article in Barrons proclaimed, “The rise of artificial intelligence heralds a new stage of the Industrial Revolution, one where machines think, learn, self-replicate, and can master many tasks that were once reserved for humans.” A recent story in the Wall Street Journal reads, “AI Is Driving ‘the Next Industrial Revolution.’ Wall Street Is Cashing In.” And at the time of writing, a Wikipedia page called “AI Boom” reads, “The AI boom, or AI spring, is an ongoing period of rapid progress in the field of artificial intelligence (AI). Prominent examples include protein folding prediction led by Google DeepMind and generative AI led by OpenAI.”
This hype around AI is not limited to the pages of newspapers or Wikipedia: tune into the business press from CNBC TV to Bloomberg Radio these days and you’ll hear breathless talk about artificial intelligence. Interview after interview with AI startup executives center on the supposedly groundbreaking revolutions their firms are on the verge of engendering. Executives of non-AI companies tell-tale of how the technology will soon revolutionize their own businesses, with C-suiters tripping over themselves to make mention of AI in their conference calls, lest shares in their firms be left out of the frenzy in all things AI. And Jamie Dimon, CEO of JP Morgan and one of Wall Street’s sacred cows, recently exclaimed about AI:
“This is not hype. This is real. When we had the internet bubble the first time around … that was hype. This is not hype. It’s real.”
But notably absent in all this discussion of an AI industrial revolution is something very basic: clearly elucidated terminology. Seldom is the phrase ‘AI’ ever even defined in all this rabble. What do these newsreaders and executives mean when they talk about AI? Are they talking about something that passes the test Alan Turing laid out in 1950 for determining whether a machine is artificially intelligent? Perhaps they’re referring to what computer scientists call ‘strong AI’, machines whose intelligence is indistinguishable from that of humans? Or do they mean something which mimics certain features of human intelligence such as the ability to process human language or mine data for patterns? If they mean the latter, is the technology they’re referring to truly revolutionary, worthy of prophecies of an industrial revolution, or merely making incremental improvements to technologies which already exist?
Add to this market hype around such a nebulously defined investment concept that Cathie Wood - whose firm has reportedly destroyed $14 billion in wealth over the last decade attempting to find the next Googles and Apples, only to stumble into Peloton - is a believer in the AI hype, and you’ve got plenty of signs of a speculative frenzy.
When investors concluded that the adoption of the internet would change the world, they euphorically bought anything remotely connected to this new technology. And they were right, the technology did change the world. Out of that period emerged companies such as Amazon which today are deeply embedded in the fabric of our lives and the global economy.
But the craze for all things internet also meant that frenzied speculators bid up stocks whose underlying companies were only remotely connected to this new technology or had no real hope of ever generating a profit. Among the most notorious of these was e-commerce company Pets.com which, despite a history of consistent losses, traded at a market capitalization north of $80 million when it went public in February 2000 only to shutter its virtual doors a few months later.
Today, it may be the case that we are on the verge of a societal revolution akin to the adoption of the internet: perhaps ChatGPT and generative AI are a heralds of a world transformed by intelligent machines. Or perhaps these technologies are just overhyped modest innovations which in retrospect will serve to show us just how far we truly are from the arrival of strong AI. Time will tell, though I’d bet on the latter.
Either way, euphoria has led speculators to uncritically bid up the prices of stocks whose underlying companies are even tangentially related to artificial intelligence, just as hype led speculators to uncritically bid up the prices of any stock remotely related to the internet during the dot-com craze. So whether we are on the cusp of I,Robot style humanoid robots, or just making some marginal improvements to existing technologies, one thing is clear: the market is in the midst of an AI bubble.