Hyperscalers are deploying $725 billion in AI infrastructure this year while demand signals point in the opposite direction. A structural mismatch now defines the industry: massive capital commitments meet visible user rejection across multiple channels.
Gartner's latest consumer survey shows 50% of US consumers actively prefer brands that don't use generative AI. Wikipedia's community voted 44-2 to ban AI-generated content outright. Stack Overflow reports new-question volume down 78% year over year, a sharp indicator that developers view AI-assisted coding with skepticism or fatigue. Google's AI Overviews feature triggered a 58% collapse in top-page click-through rates, suggesting users find the output unhelpful or worse.
The pattern repeats: capacity expansion accelerates in exactly the market segments where adoption stalls hardest. Consumers reject AI outputs. Professional communities bar AI contributions. Search platforms see AI summaries tank user engagement. Yet infrastructure spending charges ahead at record pace.
This disconnect reflects fundamental misalignment between supply-side economics and actual market demand. Hyperscalers committed to massive capital expenditures based on AI's theoretical potential. Those investments must get deployed regardless of whether users want the resulting products. The result: "AI slop" floods distribution channels. Low-quality, generic, or simply unwanted AI-generated content proliferates because the infrastructure exists and must generate returns.
The economics look broken on closer inspection. Billions flow into compute and model training. Downstream, actual customers vote with their feet. Wikipedia excludes AI. Stack Overflow users abandon the platform. Search users ignore AI-generated answers. Consumer preferences shift toward AI-free alternatives.
The hyperscaler bet rested on the assumption that AI outputs would prove so valuable that adoption would follow inevitably. User behavior now contradicts that premise. Demand doesn't materialize at