Comment by petcat
20 hours ago
Data and GPU compute capacity.
Chinese labs do not release "open weight" models because they think it is good for humanity. They do it because they are severely GPU compute constrained due to the Nvidia and ASML embargos. All available GPUs have to go toward training which leaves very little capacity to host inference at scale. So they release the models and encourage EU and USA 3rd parties to host them.
China will inevitably produce their own AI chips and then they will lock down their models just the same as everyone else.
The real loser in all this is obviously Europe, which has nothing competitive once the Chinese models are gone.
> The real loser in all this is obviously Europe, which has nothing competitive once the Chinese models are gone.
Those models are open source / open weights. Europeans can continue with Chinese models as a base same like Cursor used Kimi K2.5 a base for their composer 2.0 model or SWE-2 from cognition is post-trained Kimi K3
Similar to how some labs have done free stealth model releases. Trillions of coding tokens ought to improve performance of the downstream model. However, improving the base model architecture remains the bottleneck since you will not reach RSI with just that (assuming you ever can). And compute, lots of compute…
That strategy only works for model generations that are released with open weights. If Kimi K4 is proprietary, then China has the newer weights and the EU does not.
It's also worth noting that more post-training does not change the model architecture. No amount of post-training on deepseek v4-flash will ever make it as efficient as deepseek v4.1-flash because 4.1-flash uses a fundamentally different architecture. Being stuck on an older base model means you don't keep up with these architectural improvements that unlock significant gains in speed / efficiency.
The new Mistral large is already competitive and it hasn't even finished training.
But how does that benefit the Chinese companies to have other firms host their models and gain profit from the inference, without having to pay the training costs? Training by itself is not profitable at all, only serving the newly trained model (inference) is.
Because Chinese firms and the CCP still want to be relevant at the forefront of frontier AI development. They want to have their own sovereign AI.
The alternative would be the EU strategy which is... nothing. They're content to just be techno-colonized by USA in China.