NVIDIA CEO Jensen Huang has signaled a major positive outlook. On March 10th local time, Huang published a rare lengthy blog post titled "AI is a five layer cake," systematically outlining the development logic of the artificial intelligence (AI) industry. He believes the AI sector is undergoing a technological infrastructure build-out comparable to the Industrial Revolution. The AI industry is still in a very early stage of development; despite the industry having already invested hundreds of billions of dollars, the true potential of AI has not yet been fully realized. Trillions of dollars in continued investment will be needed in the future to complete the underlying infrastructure. Huang predicts that in the coming years, traditional software and application forms may disappear, with a new software paradigm, AI Agents, highly likely to become mainstream. Prior to this, McKinsey estimated that by 2030, cumulative global investment in data centers could reach $6.7 trillion to meet booming AI demand. This soaring capital expenditure forecast is one of the key factors driving current US economic growth.
Energy is the foundational principle of AI infrastructure. Huang pointed out that AI has become one of the most powerful forces shaping the world today. It is not merely a single smart application or model, but rather a critical infrastructure, akin to electricity and the internet, running on real hardware, energy, and economic foundations. It can absorb raw materials and transform them into intelligence at scale. In the future, every company will use AI, and every country will build AI infrastructure. In the article, Huang proposed a structural framework for the AI industry: a five-layer technology stack—Energy, Chips, Infrastructure, Models, and Applications. He emphasized that these five layers are strongly coupled. The most fundamental layer is Energy. Real-time generated intelligence requires real-time generated electricity. Every generated token is the result of electron movement, heat management, and the conversion of energy into computational power. There is no abstraction layer below this one. Energy is the first principle of AI infrastructure and the hard constraint determining how much intelligence a system can produce.
Above the energy layer are chips, Huang noted in the article. These processors are designed to convert energy into computational power efficiently and at scale. AI workloads require massive parallel computing power, high-bandwidth memory, and fast interconnect technologies. Advancements in the chip layer determine the speed of AI scaling and the affordability of intelligence. Built upon the chip layer is infrastructure, including land, power delivery, cooling systems, construction, networking, and systems that coordinate thousands of processors to work together as a single machine. These systems are "AI factories." They are designed not to store information, but to manufacture intelligence. Above the infrastructure layer are models. AI models can understand various types of information: language, biology, chemistry, physics, finance, medicine, and the physical world itself. Language models are just one category. Some of the most disruptive work currently is happening in areas like protein AI, chemistry AI, physical simulation, robotics, and autonomous systems.
Huang stated that the top layer, the application layer, is the core area where AI creates economic value, encompassing drug discovery platforms, industrial robots, legal assistants, self-driving cars, and more. The same underlying architecture can support different application outputs, and the space for innovation at the application layer remains vast. He predicts that in the coming years, traditional software and application forms may disappear, with a new software paradigm, AI Agents, highly likely to become mainstream. Every successful application will pull demand upward through every layer below it, from models and infrastructure to chips, all the way down to power plants, creating a powerful industrial pull-through effect.
Huang also pointed out that AI factories are being built because intelligence can now be generated in real-time. Chips are being redesigned because efficiency determines the speed of intelligence scaling. Energy has become critical because it limits the total amount of intelligence. Applications are accelerating because the models behind them have crossed a threshold and can finally be applied at scale. "Each layer reinforces the others," he wrote.
AI infrastructure construction is still in its early stages. "We have only invested a few hundred billion dollars so far, but need to build out infrastructure worth trillions of dollars in the future," Huang wrote. Globally, chip factories, server assembly plants, and AI data centers are being built at an accelerating pace. Huang described this trend as potentially becoming "one of the largest infrastructure build-outs in human history."
Addressing concerns about AI's impact on employment, Huang believes that AI will not eliminate jobs but instead create a large number of new employment opportunities, particularly in infrastructure and skilled technical trades. The workforce required to support AI infrastructure construction is immense. AI factories need electricians, plumbers, steelworkers, network technicians, installers, and operators—all high-skilled, high-wage positions that are currently in short supply. AI is helping to fill significant labor shortages globally in roles like truck drivers, nurses, and accountants, rather than creating unemployment. He emphasized, "Participating in this transformation does not necessarily require a Ph.D. in computer science."
Huang also specifically mentioned the role of open-source models in the AI ecosystem. A vast number of AI models globally are open-source, and businesses, research institutions, and nations rely on these models to participate in AI development. When open-source models reach advanced levels, they drive demand across the entire industry chain. He cited "DeepSeek-R1 as a prime example." After its public release, it spurred application development while also increasing demand for training compute, infrastructure, chips, and energy. In other words, a breakthrough in one model pulls demand downward through the entire industry chain.
In conclusion, Huang stressed that AI is not only transforming the software industry but will also impact energy, manufacturing, labor structures, and economic growth patterns. "AI is an industrial-scale transformation. It will change how we produce energy, how we build factories, how work is organized, and how economies grow," Huang said. He believes AI is still in its early stages. Vast amounts of infrastructure are yet to be built, and large numbers of skilled workers are still to be trained. "AI is becoming the infrastructure of the modern world."
Comments