DeepInfra Surpasses $100M ARR, Triples Token Volume

Rapid growth reflects surging demand for the infrastructure required to run AI applications at production scale

PALO ALTO, Calif., Sept. 29, 2026 (GLOBE NEWSWIRE) -- DeepInfra, a purpose-built cloud platform for high-throughput AI inference, today announced it has surpassed $100 million in annual recurring revenue (ARR), reaching over $105 million annualized run-rate, representing nearly 14x year-over-year growth. The company now processes 22 trillion tokens per week, more than tripling its token volume since May.

“AI spending is moving decisively from the lab into production, and that shift is creating an entirely new set of demands,” said Nikola Borisov, co-founder and CEO of DeepInfra. “The companies building with AI today need inference infrastructure that can deliver the performance, reliability and economics to operate at enormous scale, and that is exactly what we’ve built DeepInfra to provide.”

The milestone comes just months after DeepInfra announced a $107 million Series B, which the company is using to expand its global infrastructure footprint, increase compute capacity and accelerate product development. Since then, DeepInfra has continued to expand its customer base and infrastructure to support increasingly demanding AI workloads, including its first data center space in Canada.

Powering the Next Generation of AI Applications

DeepInfra is built to meet the demands of production AI with a vertically integrated inference platform that combines dedicated AI compute, optimized software, and a fully managed developer experience. The platform supports 200+ open-source models, including multiple versions of leading models, giving enterprises the flexibility to choose the right model for each workload while maintaining stability in production without constantly re-tuning or migrating as new versions emerge.

Companies including LiveKit, the open-source platform for building real-time voice and video applications, Humans&, a frontier lab focused on long-horizon interactions with people and organizations, and OpenCode, an open source coding agent, use DeepInfra to support production AI workloads at scale.

"Conversational AI demands inference infrastructure that can keep pace with our real-time architecture," said Neil Dwyer who runs inference at LiveKit. "LiveKit inference has strict requirements to keep time-to-first-token below 500ms and tokens-per-second above 100, all with no downtime. DeepInfra has been instrumental in helping LiveKit scale to many billions of tokens per day, and we are looking forward to continuing our work together into the trillions."

With demand continuing to accelerate, DeepInfra is expanding its international infrastructure footprint, scaling its enterprise and custom DeepCluster offerings, and growing its commercial organization to support customers deploying AI at production scale. The company recently appointed Behzad Nouri to lead go-to-market, further strengthening its ability to serve a rapidly expanding global customer base. Prior to joining DeepInfra, Nouri spent more than a decade leading sales initiatives at Twilio.

To learn more, visit deepinfra.com.

About DeepInfra
Founded in 2022, DeepInfra is a purpose-built cloud inference platform for high-throughput AI, enabling companies to run open-source and proprietary AI and agentic models at scale. The company owns and operates its AI Compute infrastructure, supports 200+ open-source models, and processes over 22 trillion tokens per week — delivering the cost efficiency, low latency, and security that production AI demands. Its fully managed platform includes OpenAI-compatible APIs and enterprise-grade security compliance, including zero data retention and SOC 2 and ISO 27001 certification. For more information, visit deepinfra.com.

Media Contact:
Julianna Sheridan
press@deepinfra.com