Imbue, a San Francisco-based AI startup, has inked a deal with Dell Technologies Inc. to build a $150 million high performance cluster of servers to train its foundational AI models.
The company is pursuing an atypical strategy in the AI space of building its AI computing power by purchasing large amounts of hardware up front. OpenAI and Anthropic — the two leaders in building foundation models, AI models that can be used for a variety of tasks and purposes — have made major deals with cloud computing providers like Microsoft, Amazon and Google to train their models.
"Building a new generation of foundation models requires the very best IT infrastructure, and Dell Technologies has helped us deploy a custom cluster much more quickly than other providers could have," said Josh Albrecht, chief technology officer of Imbue, in a press release. "Dell has been an invaluable collaborator as we pursue our work to create AI systems with much stronger reasoning abilities."
Speaking with Bloomberg News, Albrecht said the deal with Dell will allow Imbue to retain its independence and not overly rely on one cloud provider.
The interconnected nature between AI startups and their computing partner was most evident with the recent leadership crisis at OpenAI, where an eventual peace deal with the ousted CEO and the board was negotiated by Microsoft CEO Satya Nadella.
Imbue is apparently already using the cluster, which is powered by Dell PowerEdge XE9680 servers with NVIDIA H100 Tensor Core GPUs. The servers are managed by a separate company, Berkeley-based Voltage Park.
The company says that it is trying to build AI with more reasoning capacity while developing AI agents, artificial intelligence programs that can act on their own to accomplish certain tasks.
Imbue was founded in 2021 by Albracht and CEO Kanjun Qiu, and is currently valued at over $1 billion. The new deal with Dell comes after Imbue had just announced a $200 million Series B fund backed partly by Nvidia, a major AI chipmaker.
