AI inference startup Wafer secures $40 million in Series A funding, with a team of just 8 employees, while hitting $8 million in annual recurring revenue (ARR).

Beating AI News Flash: Wafer, an AI inference startup with an 8-person team, has raised $40 million in Series A funding, valuing the company at over $200 million. The startup previously turned down acquisition offers from multiple cloud providers and inference service firms. Wafer’s inference business reached $8 million in annual recurring revenue (ARR) in just about three months. Unlike chipmakers, Wafer focuses on helping models run faster and cheaper on existing hardware. It uses AI agents to automatically adjust models, inference engines, kernels, caches, quantization, and scheduling, then optimizes for different hardware like NVIDIA and AMD. Historically, much of this work required manual tuning by inference performance engineers, a task Wafer aims to automate with AI. In internal tests, Wafer found GLM 5.2 running on AMD MI355X delivers ~80% of the throughput of NVIDIA B200 at less than half the cost. Current clients include Vercel and Inworld AI. Vercel has also independently published tests showing Wafer’s throughput for GLM 5.2 is roughly twice that of other serverless providers. Post-funding, Wafer is expanding its team immediately. It has recently opened four roles: engineering, growth, CEO office, and go-to-market (GTM), all requiring 5 days of in-office work weekly in San Francisco. Technical positions offer a $250,000 base salary plus equity.