Keynote Analysis
summary
HPE announces a formal go‑to‑market with Nvidia, showing packaged systems, demos and a claim that customers will get Nvidia software like NIMs, NeMo and CUDA alongside HPE servers, storage, networking and GreenLake as HPE private AI cloud. Jensen takes roughly 25 minutes to lay out a model/data/compute stack and an imminent Blackwell chip; HPE stresses liquid cooling and shows eye‑popping 14k visuals. Claim: fix infrastructure first — customers face multi‑billion GPU backlogs and ~18% growth in AI server demand.
AI-written summary · about
