Nasdaq-listed neocloud Nebius acquires Inferize, whose tech helps optimize GPU utilization and reduce AI request costs, sources say for $100M to $150M
The Israeli startup, which was founded by former Granulate executives and employs 17 people, built technology designed to reduce idle GPU capacity when AI demand changes.
Context & Ripple Effects
Nebius has been building beyond raw AI cloud capacity: it agreed to acquire inference-performance specialist Eigen AI in May and agent-search company Tavily in February. Inferize would add a complementary layer focused on keeping GPU capacity productive as request demand shifts.
Inferize also carries a direct lineage to Granulate, the infrastructure-optimization company acquired by Intel in 2022. That makes the reported deal a targeted purchase of optimization expertise rather than a broad expansion of Nebius’s cloud footprint.
First-order effects
- If the reported transaction closes, Nebius gains Inferize’s 17-person team and software designed to reduce idle GPU capacity and AI request costs.
- Inferize’s former Granulate executives would bring infrastructure-optimization expertise inside Nebius’s inference offering.
Second-order effects
- Combining Inferize’s utilization tools with Nebius’s planned Eigen AI acquisition gives Nebius more control over both chip performance and the cost of serving inference requests.
- Nebius’s large Microsoft computing commitment raises the commercial value of software that can keep costly GPU capacity occupied as AI demand fluctuates.
Third-order effects
- Nebius’s acquisitions of Eigen AI and Inferize point to neoclouds competing on software-managed inference efficiency, not only on access to GPU supply.
- The Granulate-to-Intel and Inferize-to-Nebius lineage suggests that small teams specializing in infrastructure efficiency are becoming strategic acquisition targets for AI-cloud operators.
The trend: AI cloud providers are moving to own the optimization software that turns scarce GPU capacity into lower-cost, production inference.