In an essay for Foreign Affairs, Eric Schmidt, the ex-chief executive of Google, joins historian Niall Ferguson and researcher Lisa Klaassen in likening OpenAI and Anthropic to the East India Companies of the colonial era.
Those trading firms carried royal charters, yet raised private armies and governed regions on their own authority. The authors see the same half-leash today: the labs bargain with governments while remaining bound to them.
Scarcity moved from software to stuff
The case hangs on scaling laws. Progress now runs through power, minerals, silicon and money, so code has turned scarce again and with it geopolitical. Below a certain size, the authors contend, a player is not really competing; cobalt mined in the Congo and Kenyan labelers paid a few dollars an hour amount to what they call the plantation economics of the chatbot age.
A June episode serves as evidence. When the Commerce Department told Anthropic to bar foreign nationals from its two strongest models, the company could not check nationalities quickly and pulled Fable 5 and Mythos 5 everywhere. Roughly two weeks on, the government released Fable 5; access to Mythos 5 remains confined to selected US organizations.
Two strategies, opposite shapes
Each superpower trades what it has for what it wants. Washington offers costly closed technology and asks for political alignment. Beijing distributes open-weight models freely and gains footholds for its standards. On capability, the authors give the US about a six-month edge in chatbots and China the lead in robotics. Their bottlenecks mirror each other. Electricity is plentiful in China but advanced chips are not; the US has chips and wants minerals and a sturdier grid.
