TPCx-AI is an end to end AI benchmark standard developed by the TPC.

The benchmark measures the performance of an end-to-end machine learning or data science platform. The benchmark development has focused on emulating the behavior of representative industry AI solutions that are relevant in current production datacenters and cloud environments. This includes data management, preprocessing, training, scoring as well as serving phases. TPCx-AI also tries to address key challenges like scaling to larger datasets, representativeness to real world datasets and use cases as well as commercially available data science solutions that can be compared with a primary performance and price performance metric.


  • Industry Challenges
  • Motivation for TPCx-AI
  • Key features of TPCx-AI