Astera Labs, Inc. (Nasdaq: ALAB) has announced Orion, a modular rack-scale reference design that enables customers to choose their own GPU and CPU. The production-grade connectivity solution is designed to help neocloud and enterprise customers deploy AI infrastructure faster and more efficiently. Orion addresses the challenge of committing to a single platform before knowing which accelerator will be available or affordable, reducing cost and risk.
It also helps ODMs and OEMs solve complex integration and validation challenges for new accelerator platforms. Built on open standards, Orion supports industry-leading inference accelerators and CPU ecosystems, allowing operators to deploy infrastructure once and scale inference workloads across the most cost-effective and readily available compute resources. At the core of Orion is a high-performance, standards-based, scale-up fabric that connects up to 160 accelerators per pod, delivering predictable, low-latency communication and expanded memory capacity.
Orion's optimized PCIe connectivity and composable infrastructure are designed to help accelerators extract more tokens-per-second across AI inference workloads. The complete modular rack-scale reference design includes system configurations for the compute tray, fabric tray, memory tray, link board, and compute board, all integrated with COSMOS Rack Manager and COSMOS Fabric Manager orchestration software. Pravin Patel, vice president and general manager of the AI Platform Solutions Group at Astera Labs, stated, "We're entering a new era of AI inference, where operators are judged on the tokens they deliver per dollar, not the hardware inside their racks.
Orion gives AI inference cloud providers a production-grade, compute-agnostic architecture so they can quickly bring inference capacity online, provide outcome-based pricing for their customers, and adapt as the industry and compute landscape evolves." Orion rack reference and system designs are on track for availability through Astera Labs' OEM and ODM partners beginning in Q1 2027. Astera Labs will showcase a powered Orion rack featuring interchangeable, pre-validated compute trays with different CPU and accelerator configurations at the 2026 Open Compute Project (OCP) Global Summit, taking place October 12-15 in San Jose, California.
Sumber: TechPowerUp News · Diringkas oleh HeadlinesBriefing