CoreWeave

For decades, quant teams relied on on-prem clusters for control and security. But as AI-driven workloads scale to thousands of GPUs and run for weeks at a time, the old model is hitting a breaking point. In our latest whitepaper, we dive into why the definition of security has fundamentally shifted. It’s no longer just about isolation, it’s about: 🔹 Reliable Execution: Ensuring performance consistency as you scale 🔹 Provable Behavior: Verifying security with workload-level telemetry 🔹 Deep Observability: Gaining the visibility needed to scale without sacrificing the precision quants demand Read the full analysis here: https://hubs.la/Q049D8f90

Come see Harshdeep Banwait session at HumanX next week! You’ll walk away with clear principles to evaluate platforms, pressure-test your architecture, and move from pilot to production with confidence in the agentic era. Check out our event schedule and book a meeting here: https://hubs.la/Q049BYgm0

As the AI landscape evolves at breakneck speed, the underlying AI platform has never been more critical. If you're joining us for HumanX, you won't want to miss CoreWeave Co-founder & CTO Peter Salanki for two deep dives into the platform powering the pioneers of AI. Mark your calendars for Wednesday, April 8th: 🔹 10:50 AM | The Infrastructure Behind the AI Boom Peter joins Naveen Rao (CEO of Unconventional AI) for a fireside chat on the Main Stage. They’ll be breaking down the complexities of scaling and what it takes to build a world-class AI cloud. 🔹 12:45 PM | Technical Q&A with Peter Looking for a more interactive experience? Head over to the Builders Theatre for a deep-dive Q&A. Bring your questions on orchestration, performance, and scaling production-ready AI. Get the full details and see our HumanX schedule here: https://lnkd.in/e-46iMWf

CoreWeave’s MLPerf 6.0 results are in, and the message is clear: our innovation velocity remains unmatched. By being early to scale the NVIDIA Blackwell platform, we’ve achieved 2x throughput performance on GB300 NVL72 for reasoning models like DeepSeek-R1. Leadership is about more than just having the latest GPUs – it’s about a commitment to continuous improvement. Whether you're scaling on H100s or moving to the next frontier of Blackwell, we are constantly refining our full-stack optimization to ensure your AI platform is never the bottleneck. Excellence is a habit we’ve demonstrated benchmark after benchmark. As the AI landscape evolves, our goal remains the same: deliver the highest standard of performance and reliability for our customers. Read the full MLCommons MLPerf 6.0 breakdown: https://hubs.la/Q049sr2w0

CoreWeave's latest financing facility supports the continued growth of our AI cloud platform. It reflects strong demand for AI infrastructure as a durable, investment-grade asset class and growing confidence in our model and execution. We are building the foundation for the next generation of AI. https://lnkd.in/euE7fiJe

We are thrilled to announce that CoreWeave is a platinum sponsor at HumanX in San Francisco, April 6-9! We’re looking forward to seeing everyone at the Moscone Center and showcasing how we’re scaling the next generation of AI workloads as the #1 AI Cloud. #HumanX #CoreWeave

You’re invited! Join our watch party on April 9th for Who’s Ready for Anything with Meta’s Sy Choudhury. Details below ⬇️

Today in SF! The AI Infra Meetup with SkyPilot and CoreWeave is happening tonight. Join engineers from Meta Facebook AI, SkyPilot, and CoreWeave for tech talks on scaling training and batch jobs across K8s, Slurm, and cloud, plus plenty of time to mix and connect. The event is at capacity, but the waitlist is open! Join now: https://lnkd.in/gHXaMbnM

The next chapter for AI infrastructure is here: llm-d is entering the Cloud Native Computing Foundation (CNCF) Sandbox. As production inference becomes a standard cloud-native capability, CoreWeave is proud to continue its role as a founding contributor alongside Red Hat, IBM, Google and NVIDIA. llm-d matters because production inference places new demands on infrastructure: it needs more intelligent orchestration, economics aligned to application value, and the flexibility to run across public cloud, private data centers, and the edge. By entering the CNCF Sandbox, llm-d is moving onto neutral ground, helping make production inference more accessible, portable, and efficient for the broader industry. Read the full deep-dive from Urvashi Chowdhary, VP of Product Management, AI Services at CoreWeave: https://hubs.la/Q0488r_n0