The Essential Role of Megawatt-Class Computing in Shaping AI’s Future

Summary:

1. Jeremie Eliahou Ontiveros discusses the evolution of AI infrastructure at Data Center World 2025, highlighting rack-scale computing.
2. The industry is moving towards 1-MW racks, facing unprecedented complexity on the networking side.
3. Future trends will be shaped by power access challenges in the short term and economic considerations in the long term.

   —

   Title:
   Exploring the Future of AI Infrastructure at Data Center World 2025

   Introduction:
   At the recent Data Center World 2025 event, Jeremie Eliahou Ontiveros, a technology analyst from SemiAnalysis, shared valuable insights into the evolving landscape of AI infrastructure. The focus was on the emergence of rack-scale computing and the challenges and opportunities that lie ahead in the industry.

   Rack-Scale Computing:
   Eliahou highlighted the significance of rack-scale computing as a key architectural approach for AI workloads. The industry is witnessing a shift towards 1-MW racks, which brings about unprecedented complexity, particularly in terms of networking. Companies like Nvidia are already dealing with thousands of copper cables in their NVL72 racks, signaling the need for innovative solutions like co-packaged optics to overcome these challenges.

   Immediate Challenges and Long-Term Trends:
   In the short term, power access remains a critical bottleneck for providers, with the need to secure substantial capacity in strategic locations. Looking ahead to the projected $1 trillion data center capex milestone in 2030, Eliahou emphasized the role of economics in shaping the industry. Financial considerations will drive decisions around emerging technologies like fuel cells and behind-the-meter natural gas generation, as companies seek to justify significant capital expenditures with a focus on return on investment.

   Conclusion:
   As the data center industry continues to evolve, the insights shared by Jeremie Eliahou Ontiveros shed light on the challenges and opportunities that lie ahead. From the complexities of rack-scale computing to the economic factors driving future decisions, it is clear that the future of AI infrastructure will be shaped by a combination of technological advancements and financial considerations.

   Watch the full interview [here](insert link to video)

   By staying informed and adaptive to these trends, businesses can position themselves for success in a rapidly changing landscape of data center technology.