DeepInfra: $107M Series B

DeepInfra raised a $107M Series B to scale its inference cloud, co-led by 500 Global and Georges Harik, with participation from A.Capital Ventures, Crescent Cove, Felicis, NVIDIA, Peak6, Samsung Next, Supermicro, and Upper90. 1

The company sits in AI infrastructure: high-throughput inference for open-source and agent-driven workloads. Its announcement says token volume has grown 25x since Series A, while the release says DeepInfra processes nearly five trillion tokens per week and that revenue has tripled since the beginning of 2026. 2

Why it matters for Granola-adjacent AI investing: agent products turn meetings, workflows, and enterprise actions into repeated model calls. DeepInfra frames those workloads as 50-100+ model calls per task, which makes inference cost, latency, and reliability a portfolio-level dependency rather than a back-office hosting choice. 1

Classification: infrastructure dependency. Follow-up priority: high. The diligence question is whether DeepInfra can convert token volume into durable gross margin as agent workloads scale.