Schneider Electric CTO: "The Formula for AI Infrastructure Is Cost, Efficiency, and Power"

Jim Simonelli Defines AI Infrastructure as a "Factory"
"Must Be Built as an Integrated System, Not Just a Building"
"Grid Compatibility Is Essential for Sustainability"

The Chief Technology Officer (CTO) of a global energy management company has outlined three key pillars for artificial intelligence (AI) infrastructure: balancing cost, efficiency, and power. He emphasized that AI data centers, which require massive amounts of electricity, should not be viewed as mere buildings but rather as "AI factories"-that is, they must be constructed as integrated systems.

Jim Simonelli, Senior Vice President and CTO of Schneider Electric, presented this analysis during his talk on AI infrastructure at the "SK AI Summit 2025" held at COEX in Seoul on the afternoon of November 3.

Jim Simonelli, Senior Vice President and Chief Technology Officer (CTO) of Schneider Electric, is giving a presentation on "AI Infrastructure" at the "SK AI Summit 2025" held at COEX, Seoul, on the afternoon of the 3rd. Photo by Seongah Shim

CTO Simonelli defined AI data centers not as typical IT (information technology) facilities, but as "AI factories." He stated, "Today's AI factory should be designed as a system, not just a building. From the chip to the cooling system, we must optimize the flow of electricity, ensure that electrons reach computation without loss, and efficiently expel the heat generated in the process. This is our mission."

In this context, he proposed three principles necessary for designing and building AI factories: ▲Run Dense, ▲Run Hot, and ▲Run Steady. CTO Simonelli explained, "Building densely allows us to reduce space and resource consumption, running hot maximizes energy efficiency, and maintaining steady operations enables harmony with the power grid."

He also identified "token production" as the key factor determining the competitiveness of the AI industry. A token refers to the unit of data generated or processed by an AI model, representing the output of AI computation. CTO Simonelli stated, "The value of a data center is now determined not by the size of the building, but by how efficiently it produces tokens. For countries and companies alike, the efficiency of token production-output relative to input-will determine competitive advantage."

On the afternoon of the 3rd, a panel discussion on the theme of "AI Infrastructure" is ongoing at the SK AI Summit 2025 held at COEX in Seoul. Broadcast screen of SK AI Summit

Furthermore, he argued that these core conditions for AI infrastructure must ultimately achieve "grid harmony." He analyzed that data centers must evolve into grid-friendly structures to ensure sustainability.

CTO Simonelli emphasized, "Without considering harmony with the power grid, sustainability is impossible. The design must prevent the load fluctuations caused by AI training processes from burdening the grid." He added, "A single power failure could bring a data center with a capacity of several gigawatts (GW) to a halt. AI factories must evolve into systems that encompass not only efficiency and speed but also 'grid compatibility.'"

This content was produced with the assistance of AI translation services.