Perplexity's new multi-year deal with CoreWeave is a significant move in the ongoing 'AI infrastructure land-grab', highlighting a strategic shift for fast-growing AI companies.
This partnership is fundamentally about Perplexity's multi-cloud strategy. Rather than relying solely on a single major provider like Microsoft Azure, Perplexity is intentionally diversifying its infrastructure. This approach allows them to avoid being locked into one vendor, secure computing capacity, and optimize for the best price and performance across different tasks like model training and real-time inference. It's a savvy way to manage both risk and cost as they scale.
For CoreWeave, this deal is a crucial vote of confidence. The company recently announced massive capital expenditures of $30–$35 billion for 2026 and a record backlog of $66.8 billion. Landing a high-profile, fast-growing customer like Perplexity helps prove to investors that there is strong demand to fill all the new capacity they are building. It validates their aggressive expansion and business model focused on specialized AI cloud services, often called the 'neocloud'.
The groundwork for this deal was laid over several months. First, CoreWeave's position was strengthened by Nvidia's direct support, including a $2 billion equity stake and financing guarantees. This backing assures customers that CoreWeave has priority access to the latest GPUs and financial stability. Second, Perplexity's own rapid growth, fueled by partnerships with Snap and Samsung, created an urgent need for more computing power. Their earlier $750 million deal with Azure had already signaled their intent to use multiple cloud providers, making a complementary deal with a neocloud specialist like CoreWeave a logical next step.
Ultimately, this partnership shows how the AI infrastructure market is evolving. It's no longer just about the giant hyperscalers. Specialized players like CoreWeave, backed by key chipmakers like Nvidia, are becoming essential partners for AI innovators like Perplexity, creating a more competitive and dynamic ecosystem.
- Neocloud: A new generation of cloud providers specializing in high-performance computing for specific workloads, particularly AI and machine learning, often offering better price-performance than general-purpose clouds.
- Hyperscaler: A large-scale cloud service provider that can offer a vast portfolio of services, such as Amazon Web Services (AWS), Microsoft Azure, and Google Cloud.
- Backlog: The total value of contracts signed with customers for future services that have not yet been delivered or recognized as revenue. It is a key indicator of future business.