The massive capital expenditure by hyperscalers on AI will likely create an oversupply of capacity. This will crash prices, creating a golden opportunity for a new generation of companies to build innovative applications on cheap AI, much like Amazon utilized the cheap bandwidth left after the dot-com bust.
Related Insights
Firms like OpenAI and Meta claim a compute shortage while also exploring selling compute capacity. This isn't a contradiction but a strategic evolution. They are buying all available supply to secure their own needs and then arbitraging the excess, effectively becoming smaller-scale cloud providers for AI.
The race to build power infrastructure for AI may lead to an oversupply if adoption follows a sigmoid curve. This excess capacity, much like the post-dot-com broadband glut, could become a positive externality that significantly lowers future energy prices for all consumers.
While an AI bubble seems negative, the overproduction of compute power creates a favorable environment for companies that consume it. As prices for compute drop, their cost of goods sold decreases, leading to higher gross margins and better business fundamentals.
A primary risk for major AI infrastructure investments is not just competition, but rapidly falling inference costs. As models become efficient enough to run on cheaper hardware, the economic justification for massive, multi-billion dollar investments in complex, high-end GPU clusters could be undermined, stranding capital.
Unlike the dot-com bubble's finite need for fiber optic cables, the demand for AI is infinite because it's about solving an endless stream of problems. This suggests the current infrastructure spending cycle is fundamentally different and more sustainable than previous tech booms.
Contrary to the idea that technology always gets cheaper, building on AI is less expensive now. The current phase is characterized by abundant venture capital and intense competition among AI tool providers, which subsidizes costs for developers. As the market consolidates, these costs will rise.
Unlike the speculative overcapacity of the dot-com bubble's 'dark fiber' (unused internet cables), the current AI buildout shows immediate utilization. New AI data centers reportedly run at 100% capacity upon coming online, suggesting that massive infrastructure spending is meeting real, not just anticipated, demand.
The current AI investment boom is focused on massive infrastructure build-outs. A counterintuitive threat to this trade is not that AI fails, but that it becomes more compute-efficient. This would reduce infrastructure demand, deflating the hardware bubble even as AI proves economically valuable.
A theory suggests Sam Altman's $1.4T in spending commitments may be a strategic move to trigger a massive overbuild of AI infrastructure. This would create a future "compute glut," driving down prices and ultimately benefiting OpenAI as a primary consumer of that capacity.
The economic principle that 'shortages create gluts' is playing out in AI. The current scarcity of specialized talent and chips creates massive profit incentives for new supply to enter the market, which will eventually lead to an overcorrection and a future glut, as seen historically in the chip industry.
