Case Study: Rain AI achieves tape-out of a low‑power AI accelerator in under a year with Synopsys Cloud and IP

Rain AI Tapes-Out Low Power AI Accelerator Chip with Synopsys Cloud and IP in Under a Year

Rain AI, a Silicon Valley startup founded in 2017, set out to build a novel, ultra‑efficient AI accelerator that balances speed, power, area and accuracy for on‑device inference and training. The project required full SoC architecture exploration and tight integration of digital and analog blocks with RISC‑V, multiple verification and signoff iterations, aggressive one‑year time‑to‑market goals, scalable EDA compute and IP, and no internal CAD/IT team to manage licenses and infrastructure.

Rain AI used the Synopsys Cloud SaaS design platform—including Platform Architect, Custom Compiler, IC Validator, StarRC, PrimeSim/PrimeTime and Synopsys IP—plus FlexEDA pay‑per‑use licensing and instant cloud compute. The cloud environment was provisioned in days, enabled extensive iterative runs, and delivered major acceleration: physical verification and extraction 3x faster, timing signoff 4x faster, ~30% improvement in engineering productivity and a 30% reduction in schedule, enabling Rain AI to tape‑out its low‑power AI accelerator in under a year.

Rain AI

Jean-Didier Allegrucci

VP of Engineering