Gimlet Labs raises $300M Series B at $3B valuation
Gimlet aims to make model serving faster and more efficient by matching each phase of inference to suitable hardware.
Routing AI inference across mixed hardware
Gimlet Labs combines a distributed compiler with a cloud inference platform behind one API. Its software breaks complex model execution into smaller phases, such as prompt processing and token generation, then routes those phases across a mix of GPUs, CPUs and other accelerators suited to each workload.
This disaggregated approach is designed to reduce latency and improve power efficiency as models require more compute and memory bandwidth. Gimlet is expanding its cloud capacity and deployments for developers and enterprises running AI workloads.
Use of funds
Scale multi-silicon cloud capacity
Secure high-density data center pipelines
Expand cloud deployments
Grow engineering, operations and hardware teams