CoreWeave, a specialized, AI-focused cloud provider offering high performance compute services, has landed a whopping $2.3 billion in debt financing from several private equity firms.
CoreWeave, a specialist cloud provider offering high performance computing services to meet growing corporate demand for generative AI workloads, announced Thursday that it has received a $2.3 billion debt financing package from several asset management firms.
The key to CoreWeave’s focus on the AI market is in its hardware. The company sells primarily GPU-based virtual machines, which are particularly well-suited for AI workloads. According to Gartner vice president and analyst Arun Chandrasekaran, CoreWeave’s advertised low cost is a function of its ties to Nvidia, with which, CoreWeave has said, it has a preferred supplier arrangement, enabling it to pass on savings.
“They’re not general purpose for all workloads, and you’re not going to run web applications [on CoreWeave,]” Chandrasekaran said. “But they’re really focused on these machine learning workloads.”
Financial support for generative AI compute
CoreWeave’s new financial support is an interesting development in the generative AI marketplace, he noted, given that there are several different layers to that market, which all face different conditions for growth. At the infrastructure level where CoreWeave and competitors like Paperspace, Lambda Labs and a few others operate, their specialization is enough to set them apart and provide a different value proposition than the major hyperscalers.
“A lot of the attention is on the [large language] model layer these days, but I have a feeling that the model space is going to be hard to differentiate in the long run,” said Chandrasekaran. “The question is what is the competitive mode for these companies [and] it’s very fragmented and unclear who the winners and losers will be.”
The demand for generative AI cloud services, both for training and operational purposes, is likely to continue to grow for the foreseeable future, he added.
“As you can imagine, these generative AI models require an enormous amount of infrastructure to train and also to run,” Chandrasekaran said. “There are many different ways in which you could train models today, but many are getting trained in cloud infrastructure.”
The equity firms behind the debt package — which include Magnetar Capital and Blackstone — said that CoreWeave is well-positioned to become a key part of the rapidly growing marketplace for generative AI, thanks to its focus on providing a low-cost, AI-specialized cloud computing alternative to hyperscalers like Amazon and Microsoft.
“CoreWeave is well equipped to meet the world’s increasing need for high-performance compute and serve as a value-added provider to each of its customers,” said Magnetar managing partner and chief investment officer David Snyderman, in a statement.
SOURCE: Network World