There’s some early work being done here by companies looking at making LLM ASICS like Taalas (HC1 gets 17k t/s for llama 8b - currently at 2.5kW which is closer to a single server, but this is their first chip).

There’s other options like photonic computing which might be able to reduce power significantly but are still in research as far as I can tell. Because so much money is invested in AI & traditional gpu inference is so power hungry, I would expect significant improvements in this space quickly.