Podpower Episode AtlasIn this episode of Excess Returns, host Kai Wu welcomes financial historian, journalist, and investment strategist Edward Chancellor. Chancellor, renowned author of "Devil Take the Hindmost" and "The Price of Time," brings his deep expertise in financial speculation and capital cycles to analyze the current AI investment boom. He draws parallels to past technological manias, from the railway boom to the dot-com bubble, highlighting patterns of overinvestment, competitive fragmentation, and eventual shakeouts.
The discussion centers on whether AI's unprecedented capital expenditure, particularly in data centers and large language models, is sustainable. Chancellor expresses skepticism about the projected demand for AI, citing the inherent limitations of current large language models, such as hallucinations, and the historical tendency for investors to overestimate technological advancement and market size. He introduces the concept of "anti-bubbles" – undervalued sectors or companies that have been unfairly penalized due to the market's singular focus on a speculative boom, offering potential investment opportunities.
Chancellor also touches on China's capital cycle, the "Lindy effect" in the spirits industry, and the application of capital cycle theory to intangible assets like R&D and human capital. The episode concludes with a critical examination of the idea that bubbles can be productive for society, arguing that their long-term consequences, including capital misallocation and subsequent economic instability, often outweigh any perceived benefits. He shares his contrarian view on gold as a portfolio hedge in the current economic climate.
A theory suggesting that investment in a sector tends to surge following the advent of new technology, often leading to overcapacity, intense competition, and ultimately, diminished returns for investors. This pattern repeats across different industries and eras.
Why careUnderstanding capital cycles helps investors identify periods of overinvestment and potential underperformance, guiding them to avoid speculative bubbles and find value in overlooked sectors.
A game theory concept applied to investment booms, where individual companies are incentivized to invest heavily in a new technology, even if the collective outcome is overinvestment and suboptimal returns for the industry as a whole, due to fear of being left behind by competitors.
Why careThis explains why rational individual decisions can lead to irrational collective outcomes in investment cycles, contributing to bubbles and subsequent busts.
A term for when large language models (LLMs) generate incorrect or nonsensical information, which is an inherent limitation of their probabilistic inference mechanisms rather than a bug to be fixed.
Why careThe persistence of hallucinations limits the total addressable market for LLMs, as certain critical applications cannot tolerate such a degree of error, potentially dampening demand forecasts for AI.
Sectors or companies that are undervalued and overlooked by the market during a speculative boom in another area. These often represent sound businesses trading at cheap multiples, offering significant investment opportunities when the bubble bursts.
Why careIdentifying anti-bubbles provides a strategy for investors to find value and generate returns when the broader market is focused on speculative, overvalued assets.
A heuristic suggesting that the future life expectancy of some non-perishable things, like a technology or a company, is proportional to their current age. The longer something has survived, the longer it is likely to continue to survive.
Why careThis concept can be used to assess the long-term viability and resilience of established companies or industries, particularly those with enduring brands or products.
The idea that even if bubbles end in tears for investors, they can be beneficial for the economy in the long run by accelerating the development and adoption of genuinely transformative technologies.
Why careThis theory challenges traditional views on bubbles, suggesting a potential societal benefit despite individual investor losses, though Chancellor argues the long-term consequences of busts often outweigh these benefits.
"The general picture is a new technology, everyone gets very excited about it. A huge amount of investment, investors tend to anticipate the profits flowing more quickly than they can than they actually turn out to be the case."
"Markets are really, you know, they they have no trouble, you know, investing tons of money during these uh technology transitions, but they they they do have a trouble spotting the um the winners."
"If you've got a revolutionary new new technology and for reasons of inherent monopolistic environment, there is a case that the new technology can arrive without a huge overinvestment. But if the barriers to entry are relatively low, then you're likely historically to get overinvestment."
"The amount of hype in that, you know, oh, needless to say, by definition, all these technology manias have have large doses of hype, and in many cases, that hype over the long run is validated. There's never really never ever ever ever been as much hype as we see around AI."
"The lesson from China as far as I'm concerned is this is you don't want to when you're looking at markets, you don't want to look at valuation alone. You want to look at valuation and returns on capital and the capital cycle."
"I hold the view that having a decent allocation to gold is is a good portfolio position. So, and and that, you know, I mean, look, I may have this completely wrong, you know, in the last 10 years, actually having gold, you know, having a sort of equities gold portfolio, you know, has been obviously a lot better than having bonds."