Gimlet Labs partners with Cerebras to power ultrafast AI inference on Gimlet Cloud
CBRS•Gimlet Labs and Cerebras partnered to deliver AI inference at speeds of up to 3,000 tokens per second through Gimlet Cloud. The first Cerebras-powered datacenter is expected to come online later this year.
1. Cloud inference partnership
Gimlet Cloud will combine Cerebras wafer-scale compute with GPUs, using inference disaggregation to run different phases on the best-suited silicon. The collaboration expands joint customer work underway since last year, and an integrated solution is already serving tokens in private deployments.
