Nvidia engineers building agents is now the preferred work at the chipmaker, according to chief executive Jensen Huang, who says the shift away from writing code has been welcomed by his technical staff.
‘These agentic systems are new skills, and now we have a lot of software engineers building agents,’ Huang said in an interview published by the company on Wednesday. ‘If you ask me, every one of my software engineers prefers to be building agents than to be writing Python code.’
The message is the latest in a string of public interventions by Huang on the question of what AI actually does to the people who work with it. He has delivered versions of the argument at Carnegie Mellon University‘s commencement and on the Computex stage, according to The Next Web, sharpening the case over several months.
From Python to Pipelines: What Nvidia Engineers Building Agents Now Do
Huang described a concrete change in day-to-day work. Because of AI, Nvidia’s engineers are doing less coding, which he characterised as ‘like typing.’ In its place, they are working on agents, benchmarks, and guardrails.
‘You’re taking all the mundane work, and you’re trying to get this agent to do it,’ he said. ‘That requires imagination, that requires creativity, a lot of technology.’
The framing positions agent development as a more intellectually demanding role than writing code line by line. Huang’s view is that AI has not simplified his engineers’ jobs; it has redirected their effort toward higher-order tasks.
Huang’s Broader Case Against AI-Driven Job Losses
The Wednesday interview sits within a broader argument Huang has been making about AI and employment. He has consistently pushed back against the view that AI will eliminate white-collar roles in large numbers, a concern voiced by other technology leaders including Anthropic chief executive Dario Amodei and Amazon chief Andy Jassy.
‘The amount of work that we have to do to bring AI into the world is really quite incredible,’ Huang said. ‘So it’s creating a whole bunch of jobs. And, my software engineers love this.’
He expanded on the theme at the Milken Institute Global Conference on 5 May 2026, in a session titled ‘Leading in the Age of AI.’ There, Huang discussed agentic AI alongside what he described as a 1,000x increase in compute demand and AI’s potential as a driver of US re-industrialisation.
‘The first thing that AI is doing right now is creating an enormous number of jobs,’ he said in a separate May television interview. ‘AI creates jobs. AI is the United States’s best opportunity to re-industrialize ourselves.’
Huang cofounded the chipmaker in 1993 and has spent recent months making the case that agentic AI represents a structural upgrade to knowledge work rather than a path to redundancy.
AI agents break a task down into smaller steps, each designed to accomplish a specific sub-goal that contributes to a larger outcome. For Nvidia’s engineers, that means writing less procedural code and spending more time designing the systems that orchestrate those steps, alongside the benchmarks that test them and the guardrails that keep them within bounds.
Whether that description holds for the wider technology workforce, where layoffs have accompanied AI investment at several large companies, remains the live question. Huang’s answer, offered consistently since at least the start of 2026, is that the demand created by building and deploying AI far exceeds what automation removes. The Milken session in May, where he also addressed AI safety, showed how central that message has become to his public role. The next test of that argument arrives each quarter, when Nvidia reports the headcount behind its own surging revenue.