🤖 AI Summary
Data center vendors and hyperscalers are accelerating a move away from the long-standing 19‑inch EIA-310 rack toward 21‑inch Open Rack designs—adding roughly 51 mm of internal width—to better support modern AI infrastructure. Omdia’s research predicts the 21‑inch format (driven by Open Compute Project principles and modular DC‑MHS designs) could make up more than 70% of rack shipments by 2030, with Dell potentially shifting up to 90% of its future systems. Dell, HPE and large cloud providers are standardizing on rack‑scale modularity so compute and storage can be composed and swapped as modules rather than discrete servers, while suppliers like Wiwynn are expanding capacity to satisfy robust AI order pipelines.
Technically, the extra width gives engineers room for larger fans, improved airflow paths, denser cabling, on‑rack power distribution and easier integration of liquid cooling—critical for high‑power, heat‑dense ML accelerators. That physical headroom simplifies thermal and electrical designs, enables higher power delivery per U, and future‑proofs racks for increasing processor density. Market implications include rack shipments outpacing server shipments for the first time in a decade and an expected rack revenue market of about $4.1B by 2030, signaling a step change in how data centers are planned, cooled and scaled for AI workloads.
Loading comments...
login to comment
loading comments...
no comments yet