Organization interprets Google's chip competitiveness: parameters are not as good as Nvidia's, but system-level engineering capabilities make cost efficiency roughly equal.

With the promotion of the latest self-developed TPU chip, tech giant Google is launching an offensive against AI chip leader NVIDIA. Recently, research firm SemiAnalysis released a report explaining how Google has become NVIDIA's most threatening competitor in the commercial chip field. SemiAnalysis' model data shows that from the perspective of TCO, Google's 7th generation TPU chip has a crushing cost advantage over NVIDIA: the more Google TPUs a customer purchases, the more capital expenditure on NVIDIA GPUs they can save. SemiAnalysis points out that, taking OpenAI as an example, since the release of the large model GPT-4o in May 2024, OpenAI has not yet completed any large-scale pre-training deployment for "next-generation cutting-edge models". Although OpenAI has not switched from NVIDIA GPUs to Google TPUs, under the competitive pressure brought by TPUs, OpenAI has already obtained around 30% discount on cluster quotes from NVIDIA.