CoreWeave Announces Agreement to Power Perplexity’s AI Inference Workloads

Key Takeaways

• Multi-year deal powers Perplexity inference on CoreWeave Cloud
• Dedicated NVIDIA GB200 NVL72 clusters ensure low latency
• Perplexity Enterprise Max adds AI search and research tools
• CoreWeave leverages Kubernetes and W&B for model lifecycle
• Partnership highlights growth of AI-native multi-cloud strategies

Summary

CoreWeave and Perplexity have signed a multi‑year strategic partnership to run Perplexity’s inference workloads on CoreWeave Cloud, using dedicated NVIDIA GB200 NVL72 clusters. The deal includes deployment of Perplexity Enterprise Max, which adds advanced search, research, and data‑visualization capabilities for Perplexity employees. Perplexity is already leveraging CoreWeave’s Kubernetes service and Weights & Biases Models to move models from experimentation to production. The collaboration underscores CoreWeave’s role as a specialized AI cloud provider capable of supporting high‑demand, low‑latency production environments.

Pulse Analysis

The rapid expansion of generative‑AI services has turned inference performance into a competitive differentiator. CoreWeave, a cloud provider built exclusively for AI workloads, leverages purpose‑designed hardware and a software stack that minimizes latency while keeping costs predictable. By offering dedicated NVIDIA GB200 NVL72 clusters and a managed Kubernetes environment, the company can deliver the consistent throughput that production‑grade models demand. This infrastructure focus distinguishes CoreWeave from general‑purpose hyperscalers, positioning it as a go‑to platform for firms that cannot afford the variability of commodity cloud resources.

Perplexity, known for its answer engine, Comet Browser, and multimodal orchestration platform, processes more than 1.5 billion queries each month. To sustain that volume, the firm requires an inference layer that can scale instantly and maintain sub‑second response times. The new agreement moves Perplexity’s Sonar and Search APIs onto CoreWeave’s dedicated clusters, while also deploying the internally developed Perplexity Enterprise Max suite. Integration with Weights & Biases Models further streamlines the path from experimentation to production, giving engineers tighter control over model versioning, monitoring, and resource allocation.

The partnership signals a broader shift toward specialized AI clouds as enterprises adopt multi‑cloud strategies to avoid vendor lock‑in and optimize performance. CoreWeave’s recent Platinum rankings in MLPerf and SemiAnalysis benchmarks reinforce its credibility, potentially attracting additional AI‑native startups seeking reliable inference at scale. For investors, the deal adds a recurring revenue stream for CoreWeave and validates Perplexity’s growth trajectory, which could translate into higher valuation multiples for both publicly listed entities. As AI workloads become more mission‑critical, collaborations that combine hardware excellence with end‑to‑end tooling are likely to become the industry norm.