NVIDIA has defined artificial intelligence (AI) as a core infrastructure that supports society as a whole, not just a simple technology, and has proposed a new structure for the AI ecosystem.

On the 11th, NVIDIA likened the AI stack, which spans energy, chips, infrastructure, models, and applications, to a "five-tier cake," and announced that the global AI infrastructure transformation is in full swing.

According to Nvidia, AI is evolving from a collection of individual applications or single models to a vital infrastructure, like power and networks.

The explanation is that, due to the nature of AI, which generates intelligence in real time, the entire computing structure that supports it is being fundamentally redesigned.

At the very bottom of the AI stack is energy.

Real-time inference and generation processes consume enormous amounts of power, and the capacity to supply energy will soon determine the limits of how much intelligence AI can produce.

Above that sits an AI-specific chip that supports massively parallel computing and high-bandwidth memory.

Advances in chip technology are considered a key factor in determining the speed and cost structure of AI expansion.

Tens of thousands of processors are placed on a single chip. An infrastructure that operates as a system is built.

This facility, called the "AI Factory," is more like a factory that produces intelligence than a space for storing data.

A large-scale infrastructure combining land, power, cooling, networks, and system software is essential.

On the upper floor, there is an AI model that has learned various industrial knowledge.

Beyond language models, it expands into specialized fields such as life sciences, chemistry, physics, finance, and robotics, and is becoming the foundation for driving real-world industrial innovation.

At the top are applications that create economic value, such as new drug development, autonomous driving, industrial automation, and legal and medical support.

Nvidia diagnosed that the construction of this AI infrastructure is still in its early stages.

While hundreds of billions of dollars have been invested to date, additional investment in the trillions of dollars is expected in the future.

Accordingly, construction of semiconductor factories, AI data centers, and computing facilities is expanding globally.

In this process, the demand for skilled workers is also rapidly increasing.

AI factories require specialized personnel in electrical, cooling, networking, and facilities, leading to the creation of new, high-paying jobs.

Nvidia explained that the spread of AI is structured to create new demand by increasing productivity rather than replacing existing jobs.

Nvidia predicted that AI is now emerging as a strategic asset beyond companies and into countries, and that future competition for AI infrastructure will reshape the global economy and industrial structure.

The analysis is that the AI transition is just beginning, and its speed and scope will determine the shape of future society.