The year is 2036. You’re sitting at your office desk—alongside 100 AI agents.
At least, that’s how Nvidia CEO Jensen Huang imagines work could be one day at Nvidia. Speaking at a Q&A session for media at the Nvidia GTC conference in San Jose, the CEO and cofounder said that in a decade, the company could expect to have about 75,000 workers—nearly double the 42,000 currently at the company—all working alongside millions of AI agents.
“In 10 years, we will hopefully have 75,000 employees, as small as possible, as big as necessary. They’re going to be super busy” Huang said to laughter. “Those 75,000 employees will be working with 7.5 million agents.”
That’s a 100-to-1 ratio of agents to humans. Huang’s comments reflect the rapidly growing AI adoption across industries. And companies are increasingly bullish on AI, encouraging employees to dive headfirst into the technology. Accenture CEO Julie Sweet said failure to adopt AI could actually cost workers a promotion. Other executives from companies like OpenTable and Salesforce see AI agents as the future of work.
Huang said those AI agents won’t exactly replace workers. Instead, they’ll be picking up the grunt work human employees don’t need to complete. “They’ll be working around the clock,” he said. “So hopefully our people don’t have to keep up with them.”
Expanding AI agent fleets
AI agents differ from what most people think of as AI, like the chatbot or LLM you turn to to search for a recipe, or to plan your next vacation. AI agents, instead, are software programs that autonomously achieve certain goals set for them by reasoning, planning, and taking actions, rather than simply responding to prompts.
Huang doesn’t foresee this technology solely being used by Nvidia. At the GTC conference, Huang also unveiled an open agent development platform, the Nvidia Agent Toolkit, to help enterprises build and run their own AI agents.
“Claude Code and OpenClaw have sparked the agent inflection point, extending AI beyond generation and reasoning into action,” Huang said in a press release. “Employees will be supercharged by teams of frontier, specialized, and custom-built agents they deploy and manage.” Nvidia notes companies like Adobe, Palantir, and Cisco are already working with Nvidia’s Agent Toolkit to enhance agentic capabilities across their platforms.
A November 2025 McKinsey survey found 62% of organizations were at least experimenting with AI agents (McKinsey itself has about 25,000 AI agents working alongside its 40,000 employees, according to CEO Bob Sternfels). But nearly two-thirds of surveyed companies hadn’t yet begun scaling AI.
AI agents have recently caught the attention of those both in and outside of Silicon Valley. Entrepreneur Matt Schlicht founded Moltbook, a platform where AI agents could speak with each other without human input, and the results were both captivating and terrifying as agents chatted about everything from productivity to the nature of their existence. Meta recently purchased the platform for an undisclosed sum. And Andrej Karpathy, one of the founding members of OpenAI, recently conducted a test with an AI agent, charging it with finding a more efficient way to train a small language model. The agent ran 700 experiments in two days, resulting in 20 optimizations.
Huang is optimistic about the power of AI, and believes agents are a critical building block on the road to solving some of humanity’s most complex questions. “We’re gonna solve some really incredible problems,” he said. “The things that we are thinking about today to solve, 10 years ago nobody would even imagine that [they’re] solvable”
“We’re thinking about drug discovery like it’s an engineering problem, people are talking about extending lives,” he continued. “We will all feel superhuman.”
