CoreWeave experienced a nearly six percent increase in premarket trading on Wednesday following the announcement of a significant multi-year agreement with Perplexity, an emerging AI-driven search engine that has gained backing from notable figures like Jeff Bezos and Nvidia. This partnership designates CoreWeave as a crucial backend cloud partner for Perplexity AI, where it will manage next-generation inference tasks utilizing dedicated NVIDIA GB200 NVL72 clusters.
What does this partnership mean for CoreWeave and Perplexity? This collaboration positions CoreWeave as the foundational cloud partner for Perplexity's Sonar and Search API products as they scale their offerings. For AI applications running in production, factors such as performance, reliability, and an end-to-end AI cloud platform that streamlines compute operations are vital for success.
AI inference involves executing AI models in real time. This means utilizing trained models to make predictions or generate outputs based on new data inputs. The functionalities range from answering queries and making recommendations to classifying information and facilitating features like image recognition and language translation. Speed, stability of latency, and scalability of inference play critical roles in enhancing user experience for Perplexity’s products.
In pursuing this endeavor, CoreWeave is committed to helping Perplexity scale its inference workloads effectively. The decision to partner with CoreWeave was significantly influenced by their technical capabilities and a collaborative approach that supports AI-native companies in achieving their growth objectives. By leveraging CoreWeave’s infrastructure, Perplexity aims to improve efficiency and the quality of its models, thus providing enhanced AI-driven search and automation services across various sectors.
Perplexity has already started utilizing CoreWeave’s Kubernetes service for deploying workloads and is engaging W&B Models for training and fine-tuning within a broader multi-cloud strategy. As demand for computational power grows, specialized GPU cloud operators like CoreWeave are becoming invaluable partners for AI firms. CoreWeave stands out with impressive performance on MLPerf benchmarks and holds platinum rankings in evaluations that gauge performance and reliability.
Additionally, as part of the arrangement, CoreWeave will adopt Perplexity Enterprise Max internally, allowing its employees access to various web search capabilities, research tools, and advanced AI models through a single unified interface. This internal adoption stresses the commitment of both companies to utilizing advanced technology to bolster their operational capabilities.