The Wafer Scale Engine chip is 56 times larger than Nvidia's Tesla V100 server-grade GPU and integrates 400,000 AI-optimized cores coupled with 18 Gigabytes of on-chip SRAM. This impressive computational power can help reduce processing times for AI and machine learning applications from months to mere minutes.
https://www.notebookcheck.net/Cerebras-breaks-trillion-transistor-count-with-chip-larger-than-an-iPad.431193.0.html