Pinned
The first-ever measured silicon numbers for @nvidia Vera Rubin NVL72 are in 😲
First measured performance shows 10x more tokens per megawatt than Blackwell.
No projections. Real results from live hardware.
CoreWeave
@CoreWeave
The Essential Cloud for AI
- Q2 2026 is in the books. $2.6B revenue, up 112% YoY $104B backlog $128M operating profit, margins expanding into Q3/Q4 Congratulations to our team, and thank you to our customers. Read the full Q2 2026 results $CRWV utm.io/ur9OB
- When everyone runs the same agent loop on the same frontier models, everyone converges on the same ceiling. We can help you start to close the loop at the agent level and then fine-tune the weights as well. That is the AI Loop: utm.io/ur64R
- Stop letting GPUs sit idle between training runs. In Episode 3 of Training Tuesdays, CoreWeave's Ninad Hogade and Deok Filho break down how to run training and inference side-by-side on a single cluster to maximize efficiency. 🗓️ Aug 18th at 8 AM PT / 11 AM ET Register to save
- The real ceiling on AI isn't ingenuity. It's cost. Make inference 10x cheaper and teams stop sampling data, they run all of it. That's not a small shift. That's a different class of application. And we're building CoreWeave for that shift. utm.io/up4Eb
