We're Probably in an AI Bubble, But the Assets We're Building Might Be Too Valuable to Waste

We start by defining what separates bubble psychology from genuine boom fundamentals, looking at valuation multiples, winner concentration, and whether AI's real capabilities justify the extreme enthusiasm relative to earnings. We explore the physical constraints shaping the cycle: GPU chokepoints controlled by a handful of vendors, exponential growth in energy demands that strain power grids, and whether compute scarcity is actually inflating valuations. We examine the enterprise reality check—ninety-five percent of organizations getting zero AI returns—and consider when this gap between spending and outcomes forces a market correction. Finally, we explore the silver lining: the data centers, trained models, and specialized talent being built today could become invaluable infrastructure for a more mature, practical AI era—much like fiber became the internet's backbone after the dot-com crash.

• https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf
• https://www.reuters.com/markets/europe/is-todays-ai-boom-bigger-than-dotcom-bubble-2025-07-22
• https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/ai-power-expanding-data-center-capacity-to-meet-growing-demand
•