ABB Extends AI Data Center Power Grid to Direct Current
AI TIMES ·
✦ AI Summary
According to AI TIMES, ABB unveiled “Infinitus,” a portfolio of 800V direct current-based power infrastructure designed to reduce power loss…
According to AI TIMES, ABB unveiled “Infinitus,” a portfolio of 800V direct current-based power infrastructure designed to reduce power losses and space constraints in AI data centers on Sept. 21. ABB said the spread of next-generation AI chips could push power demand per rack above 1MW, and it proposed an integrated direct current architecture linking the power source to server racks as the solution. The portfolio is notable for bundling conversion, distribution, protection and cooling optimization around a solid-state transformer. It is designed to be applicable not only to cases that completely replace existing alternating current-centered facilities, but also to hybrid architectures that use both alternating current and direct current. ABB said direct current distribution can reduce repeated conversion steps, helping improve power efficiency and increase room for servers. The announcement shows that the focus of competition in AI infrastructure is shifting beyond semiconductor performance to how power can be supplied more densely and flexibly.
Perspective
The key point is that data center competitiveness no longer depends solely on compute hardware itself. If efforts to cut both power conversion losses and facility space become a reality, it could change the criteria for investment decisions aimed at securing more compute resources within the same infrastructure. In particular, the fact that direct current is being treated as an issue for the entire power architecture rather than just individual components suggests that future market comparisons are likely to shift from equipment performance to actual operating efficiency and deployment flexibility.
This perspective is BizCrush's own commentary and is not part of the reporting by AI TIMES.
This article was produced with the help of an automated content generation algorithm.
Source: AI TIMES
View originalThis article was summarized and organized by BizCrush based on the original article from AI TIMES. For exact quotations and full details, please refer to the original article.