Neocloud provider CoreWeave and AI search vendor Perplexity have agreed to a multiyear deal to enhance Perplexity’s AI search and inference capabilities.
The agreement, financial terms of which were not disclosed, underscores the broad applicability of Heuristics and the ongoing shift from AI training to AI inference.
Vendors revealed on March 4 that Perplexity would move its next AI estimation charge on corewave Cloud. The partnership requires Nvidia’s GB200 NVL72 cluster to power the AI model of perplexity, Sonarand its search API ecosystem. Perplexity will also use the CKS (CoreWave Kubernetes Services) and W&B (Weight and Balance) models for model management and deployment. CKS and W&B are core components of CoreWeave’s AI Cloud Platform, with CKS being a managed service optimized for computationally intensive AI workloads, and the W&B model being a specialized “system of record” for managing the lifecycle of machine learning models.
Perplexity and CoreWave deal shows further AI market changes Inference, the process by which AI uses knowledge or data acquired during training. Recently, there have been deals in which vendors are participating only for speculation. For example, OpenAI recently committed to using 2 gigawatts of capacity on AWS’s Trenium3 and Trenium4 chips, following the expansion of its partnership with the cloud provider. Meta also plans to deploy millions of people Nvidia Blackwell and Rubin GPU To run high-volume estimating and agentic workloads.
over estimate
“Estimating is a constant, continuous workload,” said Nick Patience, an analyst at Futurum Group, adding that estimating is not nonstop. “Everyone across the ecosystem recognizes that Speculation is a huge opportunity on some scale.”
While it may seem like the emphasis on inference only benefits vendors — including AI labs, model makers, and hardware providers — there is a benefit for enterprise customers using AI in applications similar to those used by consumers, according to Sandy Venugopal, CIO of Coreview.
“For enterprise customers, they want to make sure that when they’re building AI features or products or capabilities for their platform, the inference that matters to their customers is there,” Venugopal said in an interview. “When someone comes in and uses AI on your product or platform, they want instant feedback. They want to see it in real time.”
The new emphasis on inference is also due to AI vendors such as Perplexity making AI-powered experiences accessible to a broader group of people and companies, leading to increased “real-world use,” said Mike Lyon, an analyst at Omdia, a division of Informa TechTarget. This growth provides an opportunity for vendors like CoreWeave to offer purpose-built AI computing.
Benefits and Challenges
He said CoreWave’s partnership with Perplexity is about diversifying the neocloud vendor’s customer base, as its revenue is heavily concentrated in contracts with Microsoft, OpenAI and Meta.
“Landing an AI applications company like Perplexity shows that the platform can attract a broad mix of customers with different workload profiles,” Lyon said. Perplexity, for its part, gets a chance to secure high-performance infrastructure that it doesn’t have to build itself, he said.
CoreWeave is also trying to make itself a more reliable estimation platform, Patience said.
“Winning a client like Perplexity is a big deal because Perplexity is a very demanding client,” he said. He said the API will run continuously in production, and if Perplexity is satisfied with CoreWave, it is less likely to switch to another provider.
Patience added, “Estimation is even more important for Perplexity because that’s basically its business.”
However, the challenge for CoreWeave is to prove itself in a market in which hyperscalers can compete with their own in-house custom chips.
“CoreView needs to continue to prove that a purpose-built AI cloud delivers better performance and economics than what hyperscalers offer natively,” Lyons said.
