AI Investment Has Become 'Crowded'—Bill Gurley Warns of the Reality of a 'Thin-Margin Market'
While the AI investment boom could be the 'greatest wave,' renowned venture capitalist Bill Gurley describes the current market as 'crowded.' The background to this is a surge in valuations that cannot be explained by growth expectations alone, and a structure where capital is gathered before the winners have even been determined. His arguments provide hints for individual investors on how to distinguish between 'where the hype is' and 'where the reality is.'
1. 'AI investment has become crowded'—From an easy-to-win market to a thin-margin market
The essence of the 'crowding' that Gurley points out is that capital is flooding into promising deals, and the later an investor enters, the thinner their source of returns (equity) becomes. In the program, he cites examples where valuations jump in a short period, such as 'a seed round at 100 (pre-money) followed by an A-round at 300 (pre-money) with only a small amount of capital.'
What is important here is that because AI companies' revenue growth is so fast, VCs are also prone to a 'fear of missing out' mentality. The Apple Podcast program description also shows that this episode directly addresses 'why AI investment is becoming crowded.'
1-1. Rather than 'predicting the future,' 'accurately seeing the changes happening now'
Gurley describes investment judgment not as 'future prophecy,' but as 'seeing the present very clearly.' In your input text, the sentiment that
'one should not look at the future, but see the present clearly'
was also repeated. This is a perspective that is also effective for individual investors; for example, it implies that even with AI stocks, they should be evaluated based on the 'reality that will be implemented within five years.'
2. Not 'bubble or reality,' but 'the more real the wave, the more bubble it attracts'
To the question 'Is AI a bubble?', Gurley borrows the view of scholar Carlota Perez to explain that 'major technological waves invite speculation. Therefore, the more genuine the wave, the more it is accompanied by a bubble.'
Perez herself also discusses the historical pattern where the early stages of a technological revolution involve financialization, speculation, and bubbles, followed by a correction phase before moving to the next growth stage.
In other words, the higher the 'authenticity' of AI, the more natural it is for hype to be mixed in.
2-1. Therefore, a 'correction' is not a 'denial' but 'the next stage'
Gurley mentions the possibility that a phase (correction) will eventually come where prices and the cost of capital are re-evaluated. This is not about 'AI coming to an end,' but part of the transition from hype to optimization to the monetization of winners.
3. The biggest tinderbox is 'unit economics'—From market-share pricing to profitable pricing
His most practical warning is here.
Many AI companies today are going after the market with 'pricing to gain share' rather than 'pricing that reflects costs.' Therefore, when capital markets become exhausted (or funding is tightened), companies will be forced to shift from expanding deficits to prioritizing cash flow, just as Uber did.
As in your input text, Gurley even states that '(for some players) the scale of losses exceeds that of past Uber or Amazon levels,' emphasizing the reality that 'a tightening will come at some point.'
3-1. Investor checklist
What individual investors should look at in this phase, beyond flashy growth rates, is:
Room for price revisions (will customers churn if prices are raised?)
Gross margin structure (inference costs, infrastructure contracts, electricity, etc.)
Concentration of large customers and contract terms (weight of commitments). As the growth story shifts to a 'profitability story,' strong companies and fragile companies will be separated.
4. The next main battlefield is 'Open vs. Closed' + 'Inference Architecture'
Gurley cites the history of the dot-com era, when the industry shifted from 'expensive commercial servers' to 'open-source optimization,' to suggest that a similar wave of optimization may be coming for AI. Here, the points of contention are:
Open-source models vs. proprietary models
New designs for inference (specialized chips and alternative architectures)
And options like Google's TPU
(this also applies to your input text).
In short, the next phase will be decided not just by 'who has the most amazing model,' but by 'who can run it the cheapest and most reliably.'
5. What seems like an aside is actually essential: Gurley's 'career theory' also applies to investment decisions
This episode touches not only on AI investment but also on Gurley's book, 'Runnin' Down a Dream.' In summary, it explores the theme of 'how to build a career with few regrets,' emphasizing the importance of curiosity, learning, and mentors/peers.
To borrow words from the show, 'Would you want to learn about this even in your free time?' is the test for a 'suitable job.' This is similar to investing; rather than chasing superficial trends, betting on areas you are genuinely driven to understand is more likely to lead to strong, long-term decisions.
Conclusion
Bill Gurley's conclusion is simple.AI is a massive wave, but because it is massive, it is crowded, accompanied by speculation, and will eventually return to profitability.
The way for individual investors to fight is not just to 'ride the wave,' but to identify 'companies that can survive even when the reality of profitability arrives.' In the midst of the frenzy, a calm checklist becomes your weapon.
