The Overlooked Chips Powering the AI Boom

Why It Matters

The surge in AI‑driven electricity demand makes power semiconductors a critical bottleneck; securing supply and advancing efficiency will directly affect AI scalability and data‑center operating costs.

Key Takeaways

• Power chip market to hit $100 B by 2029, 7% CAGR
• AI data centers lose 10‑15% power as heat, driving efficiency race
• Infineon leads with ~45% share; market remains highly fragmented
• Silicon carbide and gallium nitride enable higher power density for AI
• U.S. seeks federal support to offset China’s growing power‑chip push

Pulse Analysis

The explosion of generative AI has turned power semiconductors into a strategic commodity. While CPUs and GPUs dominate headlines, the electricity that fuels these processors must be stepped down, conditioned, and delivered efficiently—a task handled by power chips. Omdia’s forecast of $100 billion in revenue by 2029 reflects a 7% compound annual growth rate, outpacing many other semiconductor segments. This growth is not merely a function of higher AI workloads; it also stems from the need to curtail the 10‑15% energy loss that currently manifests as heat in high‑voltage paths to GPUs, a loss that translates into higher operating expenses for data‑center operators.

Technological innovation is at the heart of the sector’s evolution. Traditional silicon power devices are giving way to wide‑bandgap materials such as silicon carbide (SiC) and gallium nitride (GaN), which can handle higher voltages and frequencies while dissipating less heat. Researchers are also exploring gallium oxide and diamond for even greater power density, a metric that determines how many watts can be packed into a given chip footprint. These advances promise to shrink the physical size of power modules, reduce cooling requirements, and ultimately lower the total cost of ownership for AI infrastructure, as well as for electric vehicles, robotics, and other high‑performance applications.

Geopolitics adds another layer of urgency. China already accounts for roughly 40% of global power‑chip sales and is actively cultivating a domestic supply chain to reduce reliance on foreign vendors. In response, U.S. manufacturers such as Onsemi and Texas Instruments are lobbying for increased federal research funding and incentives to boost domestic production capacity. The risk of a supply crunch—echoing the COVID‑era shortage of low‑cost logic chips—could become a critical bottleneck for AI rollouts if power semiconductors are unavailable. Proactive policy measures and diversified manufacturing footprints are therefore essential to safeguard the AI ecosystem’s growth trajectory.