NVIDIA CEO Jensen Huang at Dell Technologies World: ‘Demand Is Going Parabolic, Utterly Parabolic’
NVIDIA Blog · NVIDIA Writers · 2026-05-18
NVIDIA and Dell announced at Dell Technologies World 2026 that the Vera Rubin NVL72 delivers 10x lower cost-per-token for agentic AI inference and that the Vera CPU runs agent workloads 50% faster than x86, as enterprise AI investment shifts from cloud pilots to on-premises production.
Appears in
Extraction
Topics: nvidia-hardwareagentic-aienterprise-aivera-rubinon-premises-ai
Claims
- NVIDIA Vera Rubin NVL72 delivers up to 10x lower cost-per-token than NVIDIA Blackwell for large-scale agentic AI inference.
- NVIDIA Vera CPU, with 1.2 TB/s memory bandwidth, completes agentic workloads 50% faster than x86 processors.
- Worldwide AI infrastructure spending could reach $3-4 trillion by 2030, with token consumption projected to grow 3,400%.
- 67% of AI workloads now run outside the cloud, including on-premises, at the edge, or in colocation facilities.
- NVIDIA Confidential Computing enables enterprises to deploy frontier AI models on-premises without exposing model weights or sensitive data.
Key quotes
We've now arrived at the era of useful AI, which is the reason why demand is going parabolic, utterly parabolic.
The rate of change has gone parabolic, and it's not slowing down.
I think we're on the verge of maybe being able to end disease as we know it. Something like that was completely unimaginable 20 years ago, but today we can imagine it.