Startup Raises $80M to Simplify Access to AI Cloud Services

Former Google DeepMind scientists have launched Foundry, a new cloud platform aimed at making compute accessible for AI training, securing $80 million in seed and series A funding.

Sequoia Capital and Lightspeed Venture Partners co-led the investment.

Foundry will use the funds to expand and improve product offerings, as well as strike strategic partnerships.

Based in Palo Alto, California, Foundry has snapped up hardware including Nvidia’s H100 and A100s which users can access to train and run AI models.

Foundry users would be able to scale up or down their hardware needs on the fly, with the startup saying in a LinkedIn post that it wants to make accessing AI compute “as reliable and simple as flipping a light switch.”

Access to advanced AI hardware is currently dominated by large cloud providers like AWS, Microsoft Azure and Google Cloud. Foundry wants to disrupt the market.

“AI accelerator compute is arguably the most critical resource in civilization today, so the bottlenecks here reverberate broadly,” said Jared Quincy Davis, Foundry founder and CEO.

“While the much-discussed GPU shortage is part of the challenge we face, it's not the only issue. Arguably, the industry suffers vastly more from under-utilization than from under-supply. The work we are doing at Foundry is addressing this from multiple angles.”

Related:Nvidia Unveils Blackwell GPUs to Power Massive AI Models

Davis is formerly part of DeepMind’s deep learning team. Other Foundry staff were previously part of Stanford’s computer science Ph.D. program, which was led by Matei Zaharia who went on to co-found Databricks.

Their new venture also secured backing from Zaharia, as well as M12, Microsoft's Venture Fund, and angels including Google AI lead Jeff Dean and former Google CEO Eric Schmidt.