RP _ Rack & Power - Project (2026-03-11)

Why It Matters

Standardizing a high‑capacity, double‑width rack gives AI data centers a common, scalable platform, lowering costs and speeding deployment across the industry.

Key Takeaways

• Meta unveils double‑width OW rack spec for AI‑grade servers.
• Payload capacity jumps to 4,700 kg, triple previous limits.
• New toolless rails support up to 125 kg per IT gear.
• Base spec separates generic interfaces; Meta spec adds detailed dimensions.
• Community feedback sought before formal OCP contribution deadline April 24.

Summary

Meta’s engineering team, led by Daryl Daniel, presented the Open Rack Wide (OW) frame specification to the OCP community, positioning it as a standardized, double‑width rack solution for next‑generation AI workloads. The session outlined the project’s status, upcoming contribution timeline, and invited community input ahead of the EMA Global Summit abstract deadline on April 24.

The OW rack dramatically expands capacity: payload capability rises from 1,400 kg to a tested 4,700 kg, width doubles from 600 mm to 1,200 mm, and height adds roughly 100 mm. Structural analysis and validation testing support these figures, while toolless rails now handle up to 125 kg per IT component. The design retains compatibility with existing HPR bus‑bar mounting interfaces and introduces optional lift points, forklift access on all sides, and larger cable openings.

Daniel highlighted that the specification is split into a base spec—defining generic equipment and bus‑bar interfaces—and a Meta‑design spec that details geometry, payload limits, and accessory options such as removable cross‑braces and enhanced base platforms. He noted that the rack’s massive weight demands a reinforced base and that lift points enable crane handling for manufacturing or data‑center deployment.

If adopted, the OW rack could become the de‑facto standard for high‑density AI hardware, reducing custom engineering, easing integration across vendors, and accelerating data‑center scaling. Broad community feedback will shape the final OCP contribution, potentially influencing the next wave of AI‑focused infrastructure investments.

Original Description

Public call recording of RP _ Rack & Power project.