Let’s make China work for us, securely: Building India's domestic compute, keeping good terms with US

Chinese open-weight models now deliver near-frontier capability at a fraction of the cost, trailing Anthropic's frontier models by roughly eight months.

Moonshot AI's Kimi K3 beat Anthropic's Fable 5 on some coding-related benchmarks, showing rapid progress.

Using a proprietary model via a foreign API creates continuous dependence; for example, access could be removed by export controls.

Running a Chinese open-weight model on Indian GPUs in Indian data centres keeps the data pipeline entirely domestic, with no ongoing connection to the model creator.

There is no special backdoor in an open-weight model, which is just a static file of parameters, making it a more sovereign option than a foreign proprietary API.