Two stories dropped within 24 hours of each other and nobody put them together. The Trump administration cleared Anthropic's Mythos 5 for use by over 100 companies and government agencies, a controlled domestic rollout that smells like strategic rationing. Simultaneously, Asian AI startups are launching Mythos-comparable models specifically designed to sidestep export ban risk. The geopolitical logic is almost too clean.
When Export Controls Create the Competition They Fear
The history of technology export controls is not encouraging for the controlling party. Restricting access to advanced semiconductors accelerated China's domestic chip ambitions. Restricting access to frontier AI models is producing the same dynamic, faster. The Asian labs launching now aren't just filling a gap: they're building a parallel ecosystem with zero dependency on US licensing, US compute, or US regulatory goodwill. Fast Company's reporting on congressional AI regulation efforts captures the paradox: the more aggressively the US tries to contain frontier AI, the more it incentivizes the rest of the world to build outside the blast radius.
NIH Grants, Political Screening, and the Pattern
There's a structural echo here worth naming. Nature's investigation into NIH grant screening documents how political criteria are now stalling scientific funding. Whether it's biomedical research or AI capabilities, the US government is simultaneously trying to lead and control the same fields, and the friction is creating openings for everyone else. The AI cold war isn't coming. It's the current operating environment. The question is whether controlled domestic deployment of Mythos constitutes a strategy or just a slower form of ceding ground. As Stanford's Fred Turner has noted, the California-to-Texas ideological move in tech has always been about reframing control as freedom. Mythos rationing is that move applied to geopolitics.