When most people hear about artificial intelligence, they think about software. Models, algorithms, chatbots, and code. But the real story unfolding behind the AI boom is far more physical.
AI runs on hardware. A lot of it.
The massive growth of AI is driving an equally massive expansion of data centers and the infrastructure that powers them. Inside those facilities are thousands of GPUs, servers, cooling systems, networking equipment, and electrical components. Every one of those systems requires specialized manufacturing somewhere in the supply chain.
This is where the story becomes interesting for manufacturers.
One recent example illustrates the trend well. Earlier this month, NVIDIA announced a $2 billion investment in Coherent, a company that produces advanced optical and laser technologies used in high-speed data communication systems. The goal of the partnership is to expand manufacturing capacity and accelerate the development of optical technologies needed to support next-generation AI data centers.
Why optics?
As AI systems scale, the amount of data moving between processors becomes enormous. Traditional electrical connections struggle to keep up with the bandwidth requirements. Optical technologies—using lasers and fiber connections—allow data to move much faster and more efficiently between chips and across data center infrastructure.
In simple terms, AI systems are becoming so powerful that the industry needs an entirely new generation of physical networking hardware to keep them running.
That hardware doesn’t appear out of thin air. It requires complex manufacturing across multiple tiers of suppliers.
Optical systems require precision housings, specialty materials, heat management systems, electrical components, and highly engineered assemblies. The finished laser modules and optical networking equipment sit at the top of a long manufacturing supply chain.
Many manufacturers already produce components similar to what this ecosystem requires, but they often don’t realize it.
A machining company that produces precision aluminum components for aerospace may already have the tolerances needed for optical system housings. A metal fabricator that produces electrical enclosures may be able to support power infrastructure inside data centers. Plastics manufacturers may already have the capability to produce cable management systems, cooling components, or connector housings.
The opportunity often exists two or three layers down the supply chain.
Most manufacturers will never supply directly to a company like NVIDIA, just as most automotive suppliers never sell directly to the car brand on the hood. Instead, they support the companies that support the companies building the final system.
That’s where the AI infrastructure boom becomes a manufacturing opportunity.
Over the next several years, companies like Microsoft, Amazon, Google, and Meta are expected to invest hundreds of billions of dollars expanding AI infrastructure. These facilities require enormous amounts of electrical equipment, cooling systems, structural components, and networking hardware.
The technology may be cutting edge, but the supply chain behind it still relies heavily on advanced manufacturing.
For manufacturers, the key question isn’t whether AI will impact the industrial economy. It already is.
The real question is whether manufacturers can recognize where their existing capabilities fit into this rapidly growing supply chain—and position themselves to capture the opportunity before someone else does.