NVIDIA's CEO has talked about the 'agentic AI' inflection point at the Morgan Stanley conference, and he has called out OpenClaw as the "most important" software release of our times.
NVIDIA's CEO Says that Agentic AI Has Brought Uses 1,000x Higher Tokens, Bringing In Immense Compute Demand
Jensen has talked about AI being a "5-layer cake", and one of the more interesting layers that yields the most returns to hyperscalers and frontier labs is the applications layer. OpenClaw and AI agents are examples of how AI, when placed in a hyper-personalized environment, yields results that replicate human workloads. NVIDIA's CEO was asked about how he sees enterprise demand for AI evolving, and when discussing multiple industry inflection points, he mentioned OpenClaw, calling it a piece of software that surpasses Linux in adoption.
Probably the biggest phenomenon that’s happening, and if you’re paying attention to it, I’m sure you are, OpenClaw is probably the single most important release of software, you know, probably ever. If you look at OpenClaw and the adoption of it, you know, Linux took, right, some 30 years to reach this level. OpenClaw in, what is it, 3 weeks, has now surpassed Linux. It is now the single most downloaded open source software in history, and it took 3 weeks.
- NVIDIA's CEO Jensen Huang
Jensen is fascinated by how, through a series of prompts, agents like OpenClaw can execute tasks that would traditionally require domain-level expertise and significant time. The reason OpenClaw has gained immense popularity isn't that its implementation is complex; rather, it has shown the world that AI has use cases that can directly impact consumers' lives, making redundant tasks much easier. And for enterprises like NVIDIA, agents like OpenClaw have created demand that the industry wouldn't have anticipated at all.
NVIDIA's CEO notes that with agents, token consumption has risen by a whopping 1,000 times, creating a 'compute vacuum' where, no matter how large hardware deployments become to address the token demand, they will remain constrained as long as agentic AI continues to infiltrate human workloads. Agents can perform bulk web searches, image generation, complex analysis, and other workloads, which have dramatically increased token utilization rates, which means that for NVIDIA and others, there's a whole lot of compute demand popping up.
When we talk about the specific compute architectures assigned to the "5-layer" cake we mentioned above, Hopper and Blackwell focused specifically on training workloads. However, with Vera Rubin, agentic AI constraints would be addressed by focusing on long-context workloads through ramped-up onboard memory elements and platforms like ICMS. Given how massive the compute-token imbalance currently is, the demand for Rubin should be gigantic.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
