This is not just faster inference. It’s a new foundation for AI at scale.

Did they generate their website with their own chips or on Nvidia hardware?