The crypto market is volatile, mostly non-regulated, suffers from a limited track record, and there are wide disputes about whether the market possesses any intrinsic value. No matter one’s stance on the latter, there is arguably a broad consensus that these attributes cause the crypto market to stand apart from most other markets. Given this, crypto presents a poor model to be compared to other markets, except if it regards bubbles caused by immature but highly anticipated technologies. In this case, no market is arguably better suited than crypto to set side by side. In doing so, we see remarkable similarities between crypto and the recent rise of AI, including the seemingly one-way street of capital flowing to everything AI.
In recent years, the crypto market has experienced two such one-way streets before the capital pools were emptied, causing the bubbles to burst. These bubbles occurred in 2017 and 2021, when the value of Bitcoin, Ethereum and various other cryptocurrencies skyrocketed, only to subsequently plummet by as much as 90% the following year. Not only did prices rise, crypto projects and firms could raise money as easy as pie during the bubbles, as everyone wanted in on the action – then suddenly, nobody seemed to care anymore, leaving only a few interested in crypto.
At what point does the market’s collective imagination end?
Crypto and AI both have great prospects for the future as cornerstone technologies among an ever-growing crowd. Admitting we fairly share this outlook, we must acknowledge that both crypto and AI have yet to experience wide implementation across the world, having shown limited value creation at present. Worse yet, there’s no way of knowing at what point the technologies will enjoy adequate maturity to be applied widely, if ever. But it is roughly guaranteed that folks will overestimate the short-term significance of AI, as with most other new technologies, including previously crypto.
Up to technical maturity, the market judges technologies fairly blindly, based on collective imagination. As Benjamin Graham once said, "In a speculative market, what counts is imagination and not analysts”. Graham, often regarded as the father of value investing, possibly intended his words as a cautionary reminder for any market where guesswork about the future is a main pillar to assess it. As crypto and AI are not yet well-established technologies, the imagination of the future surpasses the tangible aspects of the present, so rather than relying on concrete data and figures, market participants allow their imagination to shape their perception of the industry's future impact on the world and, subsequently, its value. This is likely to truly deviate from how AI is to unfold later on, in the same way it has for crypto multiple times.
Besides, this judgment is mainly guided by retail investors. This group was largely absent in the AI rally earlier this year, as many retail investors had reduced their equity holdings upon surging interest rates, but it appears that they have returned in the past few months, particularly to AI-related stocks. We expect the retail flow to continue into AI, gradually turning it into a new darling of retail, as these investors enjoy markets with greater risk/reward, similar to crypto and meme stocks, in which retail investors are highly dominant. The noteworthy presence of this fear-of-missing-out crowd in a market led by imagination fuels a cocktail for promising bubbles.
We would be surprised if this cocktail does not involve at least one significant drawback of AI-related tradeable instruments, similar to the history of crypto, and not least the dot-com bubble. We expect the market to eventually reassess its projection of the near-term impact of AI, as the technology matures more slowly than expected and other challenges become evident, including regulatory uncertainty, which AI, like crypto, is likely to suffer from. And, except for a select few firms, the unfamiliarity with the extent to which individual firms are able to capture the value generated by AI adds to the uncertainty, potentially intensifying excess volatility.
“During a gold rush, sell shovels”
One of the select few firms whose ability to capture the value of AI is already clear as day is GPU manufacturer Nvidia. As the leading supplier of the computing required to train AI models (for example, those by OpenAI), Nvidia has achieved conspicuous revenue growth in the past year, mainly driven by the eagerness of companies to acquire GPUs so as not to fall between the cracks of the AI rage. In late May, Nvidia said it expects revenue of $11bn in the second quarter of 2023, far surpassing the average analyst forecast of $7.18bn. This AI hoarding has sent Nvidia past the $1tn valuation mark, alongside only five other firms, quadrupling its stock price since its 1-year low in October 2022. The $750bn expansion to Nvidia’s $250bn market capitalisation less than 12 months ago is a clear reminder that AI is a speculative market, by which the imagination of investors has persuaded them that something truly great lies ahead.
The story of Nvidia points to the adage that “during a gold rush, sell shovels”. This suggests that it’s more profitable to provide the required infrastructure to those pursuing a booming trend than to pursue the craze itself. In the crypto rush of 2021, Coinbase profited heavily by selling picks and shovels by facilitating crypto trading, similar to Nvidia now. In due time, the crypto bubble burst, shortening Coinbase’s revenue markedly and bringing an up to 90% drop in its stock price in a single year. Time will tell whether Nvidia is in an everlasting gold rush or will follow Coinbase’s “the higher you climb, the harder you fall” tale.