DeepSeek and the Price Collapse of Thinking

January gave us the whole post-labor thesis in one month, told three ways.

First: DeepSeek, a Chinese lab most Americans had never heard of, released R1 — a reasoning model competitive with the frontier — claiming a headline training cost under six million dollars, and open-weights at that. Markets panicked; Nvidia posted the largest single-day value loss in stock-market history before recovering its nerve. The panic was aimed at the wrong thing. Cheap reasoning isn't bad news for the AI economy — it's the point of the AI economy. The price of thinking fell, hard, and prices that fall like that never climb back.

Second: OpenAI shipped Operator, its first true agent — a model with its own browser that fills carts, books reservations, and completes forms. Clumsy? Yes. But remember what the first year of copilots looked like, and then remember what year three looked like.

Third: a $500 billion datacenter program called Stargate was announced from the White House. Whatever fraction ultimately gets built, the signal is unambiguous — the buildout of machine cognition is now national industrial policy.

The post-labor math

Hold the three together. The unit cost of intelligence is collapsing (DeepSeek), the delivery mechanism for that intelligence is becoming action rather than advice (Operator), and the supply side is being scaled like a wartime mobilization (Stargate). When something gets radically cheaper, becomes self-executing, and has trillions in committed supply — that thing is about to be everywhere, priced into everything, including wages.

Actionable Steps

Businesses: your five-year assumption should now be that reasoning approaching frontier quality trends toward commodity pricing — plan competitive advantage around data, distribution, trust, and taste instead. Workers: agents doing web-based tasks means the automation frontier now includes anything done through a browser; move your value toward judgment, relationships, and accountability. Policymakers: you cannot tariff a price collapse in cognition. Transition programs, not trade walls, are the lever you actually hold.

Cheap thought is the most important price change since cheap energy. This was the month it showed up on the ticker.