The artificial intelligence investment boom has split market observers into two sharply opposing camps — and the divide is growing. Major equity indices continue to reach record highs, driven largely by a concentrated group of AI-exposed megacaps, while capital expenditure on AI infrastructure is climbing to levels that rival national-scale projects.
The Bear Case: Déjà Vu from the Dot-Com Era
Sceptics, including valuation expert Aswath Damodaran and AI critic Gary Marcus, argue that expectations have significantly outrun the technology’s demonstrated capabilities. They point to a crowded landscape of roughly 300 AI unicorns, stretched valuations, and the structural risk that many AI application companies are built on the same underlying models — creating redundancy rather than differentiation. The parallel to the dot-com era is hard to ignore: during that cycle, an estimated 50–60% of venture-backed internet companies ultimately failed or disappeared after the crash.
The Bull Case: Infrastructure, Not Just Hype
On the other side, strategists such as Josh Brown and Tom Lee contend that investors are still underestimating AI’s long-term economic impact. Their core argument is that the current wave is fundamentally infrastructure-led — more akin to the buildout of electricity grids or telecommunications networks than a software fad. Infrastructure investments of this kind tend to carry long-term pricing power and generate stickier usage patterns, making the risk profile meaningfully different from the dot-com era.
The Real Question Is Magnitude, Not Survival
The more nuanced framing emerging from the debate is not whether AI will endure — most analysts on both sides accept that it will — but how large and how fast its economic returns will materialise. That outcome hinges on how well demand growth, infrastructure capacity, and capital returns align over the coming years. Both bulls and bears may ultimately be correct about different segments of the market: transformative for infrastructure plays, brutal for the long tail of application-layer startups.
Source: Techreport. This summary was produced with AI assistance and edited for context; please refer to the original for full details.