SAN FRANCISCO, April 28, 2022 – Today, Exafunction, a deep learning infrastructure company, announced a $25 million Series A investment, led by Greenoaks with participation from Founders Fund. Exafunction will use this financing to solve the most critical challenges facing deep learning, helping customers optimize their use of resources and improve performance, starting with the virtualization of GPUs.

With Exafunction, customers achieve 5-10x improvements in hardware utilization, getting more from their existing resources, speeding up their models, and cutting both costs and energy usage. Over the next year, Exafunction is focused on deepening and broadening its platform by helping existing customers squeeze even more performance out of their available hardware, and expanding the range of verticals that the company can serve. They are also working to speed up the deployment of models at scale, improving performance for companies with latency-sensitive applications.

“Deep learning is a generational technology, but reaping its benefits remains too costly, frustrating, and technically difficult,” said Varun Mohan, Co-Founder and CEO of Exafunction. “Before starting Exafunction, we saw first-hand the promise, but also the challenges of deploying these models in the real world. Now, we’re building an infrastructure layer to bring the potential of deep learning to everyone, while also helping advanced companies extend their compute capacity ceilings and remove impediments to progress.”

“Deep learning inference is the variable cost behind some of the most productive insights in the world,” said Sreyas Misra of Greenoaks. “But even the most sophisticated organizations struggle to run deep learning efficiently, wasting time, capital, and scarce GPU capacity in the process. This is an unusually technical problem that very few engineers can solve: Exafunction has several of them already, and we think they’ll hire many of the rest over time. We believe Varun and Douglas are building the reference infrastructure for serverless deep learning, and we are thrilled to partner with them as they pursue this mission.”

The funding follows a $3 million seed round that was raised last year, also led by Greenoaks and a number of angel investors, including Spencer Kimball (CEO Cockroach Labs), Neha Narkhede (ex-CTO Confluent), Sahir Azam (CPO Mongo), Carlos Delatorre (ex-CRO Mongo, CRO TripActions), Howie Liu (CEO Airtable), Richard Socher (CEO You.com, ex-chief scientist Salesforce) and others. As part of its plans to scale, Exafunction is also hiring a world-class engineering team with knowledge in machine learning and distributed systems.