The debate over artificial intelligence jobs got its sharpest rebuttal yet from a hardware salesperson on Tuesday.
Nvidia CEO Jensen Huang published a rare stand-alone article on Tuesday laying out what he calls the “five-layer cake” of artificial intelligence infrastructure: base energy, then chips, then physical infrastructure, then models, then applications.
It positions AI as an electrification-scale industrial construct, rather than a software product or chatbot, which requires trillions of dollars in physical construction and legions of electricians, plumbers, pipefitters, steelworkers, and network technicians.
“These are technical, high-paying jobs, and demand exceeds supply. You don’t need a PhD in computer science to participate in this shift,” he said.
Huang’s argument for why such a massive expansion is needed begins with a fundamental shift in how computing works.
Traditional software retrieves stored instructions, while AI generates new output in real time, with each response creating new ones based on the context provided. It’s not about finding answers, it’s about reasoning as needed.
Because intelligence is generated in real time, the entire computing stack beneath it must be reinvented, which is why AI requires purpose-built infrastructure from the energy layer upwards, rather than running on existing data centers.
The time is clear. The article follows growing concerns about the impact of artificial intelligence on employment, from mass layoffs at Block Inc. to comments from Anthropic CEO Dario Amodei about job losses. Tech stocks have been selling off since the start of the year on these concerns.
Huang’s article, however, is a direct counter-narrative. He cited radiology as an example, arguing that artificial intelligence can help read scans, but the demand for radiologists continues to grow because productivity creates capability, and capability creates growth. “This is not a paradox,” he writes.
Huang sees energy as the foundation of the AI era.
“Intelligence generated in real time requires power generated in real time,” he wrote. “Energy is the first principle of AI infrastructure and the binding constraint on how much intelligence a system can produce.”
The impact of this framework extends beyond Nvidia’s supply chain. If energy is the constraint for AI, then anything that disrupts energy supplies, including the current wars in the Middle East, is more than just a macro headwind for markets. This is a direct bottleneck to the speed of artificial intelligence expansion.
Huang acknowledged that the expansion is still in its early stages. “We’ve invested hundreds of billions of dollars. Trillions of dollars of infrastructure still need to be built,” he said, adding that artificial intelligence factories were being built around the world “on an unprecedented scale.”
He also gave high recognition to open source models and used DeepSeek-R1 as an example to illustrate how to provide powerful inference models for free, “accelerate the adoption of the application layer and increase the demand for training, infrastructure, chips and energy underneath it.” Open source will not threaten Nvidia’s business. It feeds it.
