This week, Nvidia CEO Jensen Huang introduced the NVQLink, a system that connects quantum processors to GPU computing to significantly improve AI capabilities. The company also announced its collaboration with the Department of Energy to build seven AI super computers. Similarly, OpenAI also engages in building AI infrastructure under its Stargate initiative, expanding from Texas and New Mexico to Ohio, Wisconsin, and Michigan.
These expansions are typically covered as the next step in computing. A longer view, however, reveals a broader economic shift. We are witnessing the early stages of a global reindustrialization, centered not on the production of physical goods, but on the manufacturing and automation of intelligence. This transition is defined by the rise of a new type of infrastructure: the AI Factory.
From Mechanized to AI Factory
The First Industrial Revolution replaced muscle power with mechanical power, centered on the factory and the steam engine. This required massive manufacturing of machine tools and resources such as coal and water, the infrastructure for physical power. A century later, the Second Industrial Revolution, built on electricity and the assembly line, organized this power for unprecedented scale and efficiency.
The Computer Revolution of the mid to late 20th century was defined by inventions of the transistor, the microprocessor, and the integrated circuit. These inventions shrank computers from room-sized machines to personal devices on every desk. The economic model was about producing and selling physical hardware and the packaged software that ran on it. The personal computer changed the way people worked and lived, amplifying the capabilities of an individual user.
The current shift is different in nature. We are moving from an era of personal tools to one of intelligence production and automation at scale. The AI Factory’s raw materials are data and electrical power. Its machinery is vast arrays of specialized and general purpose algorithms. Its output is a stream of tokens, the fundamental units of understanding for AI, which can represent words, image fragments, or commands for a robot.
Automation and New Industrial Frontiers
The development and construction of these AI factories will potentially bring an industrial boom. This process drives a form of reindustrialization, not in traditional sectors, but in the foundational layers of the AI economy: the manufacturing of advanced computing hardware, the expansion of energy infrastructure, large-scale construction of data facilities, and the production of next-generation robotics. Companies like Nvidia, Meta, and others are developing world models, AI systems that simulate real-world physics to train and operate robots.
This pushes the AI revolution beyond the digital screen and back into the physical realm. Developing robots requires a resurgence of advanced manufacturing: producing precision actuators, sensors, and chips. Deploying them at scale will demand robust supply chains, quality control, and cross-industry collaboration, echoing the complex logistics of traditional industrial sectors.
The ripple effects of AI factory will reorganize labor across the economy. Industries like logistics, construction, manufacturing, assembling, and cooling systems will see massive growth. Automation through robotics and AI agents, however, is profoundly changing the kind of skills needed in the workforce. Underpinning it all will be an unprecedented need for energy generation and grid infrastructure, creating a significant expansion in the power sector.
AI computation is profoundly energy-intensive. OpenAI’s AI infrastructure and Tesla’s datacenter, for instance, are being built where power is abundant and relatively affordable. This dynamic makes access to scalable, sustainable energy essential for future AI development.
From Tools to Colleagues: The Economic Shift
The most significant change lies in the nature of AI’s output. Traditional software created tools like Excel and Photoshop, which assisted us in our work. The market for such tools is large, but finite.
In contrast, AI is turning into a digital colleague that claims to free people from work and tasks of any kind. AI coding assistant like Cursor generates and debugs code acts as a partner to an engineer. OpenAI’s new Atlas browser, which can search, synthesize, analyze and take actions all on one platform, is beginning to shape how we work and communicate.
OpenAI’s new Apps SDK, a toolkit that lets developers build fully interactive applications inside ChatGPT, positions the platform as a foundational infrastructure. This move is reminiscent of how the Windows operating system became the central platform for the PC era. By allowing anyone to build and use specialized apps within ChatGPT, it has the potential to become a primary gateway for digital services.
This dual nature defines the reindustrialization. The AI factory is itself a new core industry. Simultaneously, its product, AI, acts as a force multiplier for the global economy, both in digital and increasingly through robotics, in the physical world. We are witnessing the infrastructure to produce the commodity of the coming decades: intelligence.
