The artificial intelligence boom is creating a new challenge for the global technology industry: power availability. As companies build increasingly large AI data centers packed with GPUs and high-bandwidth memory, electricity demand is rising rapidly, turning power infrastructure into one of the biggest constraints on AI expansion.
A recent analysis from Morgan Stanley warns that power constraints could delay AI data-center deployments, creating ripple effects across the semiconductor supply chain. While NVIDIA and Broadcom may be relatively protected because of their positions in AI computing, companies supplying memory, optical components and other infrastructure could face greater pressure if new data centers cannot secure sufficient electricity.
The problem is being driven by the enormous power requirements of modern AI workloads. Traditional data centers were designed around much lower computing densities, while AI clusters combine thousands of high-performance GPUs with networking, cooling and storage infrastructure. This means new facilities require not only more electricity but also advanced power-delivery and cooling systems.
The situation is pushing the industry toward technologies such as 800V DC power architectures, which can deliver electricity more efficiently to high-density AI racks. Microchip Technology and Navitas Semiconductor recently introduced an 800V-to-6V reference platform specifically for next-generation AI data centers, highlighting how seriously the industry is treating the power bottleneck.
For the AI industry, this creates a new equation: having the latest GPU is not enough if a data center cannot obtain enough electricity to operate it. Delays in grid connections, limited generation capacity and the growing energy requirements of AI infrastructure could therefore become major factors determining how quickly the next generation of AI systems can scale.
Comments