Huang Renxun published a lengthy article proposing the AI "Five-Layer Cake" framework, stating that the scale of AI infrastructure will reach trillions of dollars.

According to CoinWorld, based on monitoring by 1M AI News, NVIDIA CEO Jensen Huang published a lengthy article titled “AI Is a 5-Layer Cake” on the company’s official blog on Tuesday. Starting from first principles, he breaks down the AI industry into five layers: energy, chips, infrastructure, models, and applications. He believes AI is not just a clever application or a single model but is “as fundamental as electricity and the internet.” Every successful AI application influences the entire industry chain downward, all the way to the power plants at the bottom that supply electricity. Huang pointed out that traditional software is “pre-made”—humans describe algorithms, and computers execute instructions—whereas AI breaks this pattern, achieving “real-time intelligent generation” for the first time. Because intelligence is generated in real time, the entire computing stack supporting it must be reinvented. He describes current investments as “only a few hundred billion dollars” but believes that future infrastructure development will still require trillions of dollars, “becoming one of the largest infrastructure projects in human history.” Regarding employment, Huang directly addressed recent concerns about AI replacing jobs. He emphasized that building AI factories requires many skilled trades such as electricians, plumbers, steelworkers, and network technicians, “all of which are highly technical, well-paid positions, and currently in extreme shortage. Participating in this transformation does not necessarily require a PhD in computer science.” Using radiology as an example, he explained that the demand for radiologists actually increases after AI-assisted image interpretation because “productivity creates capacity, and capacity drives growth.” He also praised the role of open-source models, stating that DeepSeek-R1 has accelerated application development by making powerful reasoning models widely available, while also increasing the demand for training compute power and infrastructure. For more details, read CoinWorld’s latest full translation: NVIDIA Jensen Huang’s Latest Article: The “Five-Layer Cake” of AI