Back to news

Hyve Solutions and Astera Labs Collaborate to Bring AI Inference Capacity to Customers Faster

Globe Newswire•16/11/2026•09:00 ET
9

Key Highlights

  • ➤Hyve expands Orion integration across Apex rack-scale AI platform.
  • ➤Orion eight-GPU configuration delivered up to 89% higher tokens per second.
  • ➤Apex systems with Orion are targeted for first-quarter 2027 availability.
  • ➤Orion architecture supports scaling up to 160 accelerators in a pod.
  • ➤Hyve’s air-cooled Apex chassis supports higher-power accelerators.

Expert Statements

Pravin Patel, Vice president and general manager, AI Platform Solutions Group at Astera Labs

“Every GPU is waiting on something, and too often it is the fabric.”

Pravin Patel, Vice president and general manager, AI Platform Solutions Group at Astera Labs

“Thanks to Astera Labs’ advanced in-network compute and Hypercast™ features available in the Orion architecture, Orion is designed to let customers scale up to 160 accelerators in a pod with predictable, low-latency communication.”

Pravin Patel, Vice president and general manager, AI Platform Solutions Group at Astera Labs

“Hyve turns that into a deployable chassis, and customers can engage with Hyve today to get better cost per token.”

Dr. Dimitris Ziakas, Chief technology officer and vice president of Architecture at Hyve Solutions

“The rack is now the unit of AI, and rack-level systems are what Hyve builds.”

Dr. Dimitris Ziakas, Chief technology officer and vice president of Architecture at Hyve Solutions

“We design, integrate, validate, and deploy at rack scale. Pairing Hyve Apex with Astera Labs’ Orion gives customers scale-up AI that they can deploy once and evolve over time, swapping accelerators as their workloads change without rebuilding the system.”

News Highlights * Expanded Orion portfolio: Hyve Solutions to integrate and deploy the Orion Compute Board (Orion-CB), Fabric Tray (Orion-FT), and Link Board (Orion-LB), validated building blocks for open, rack-scale AI inference. * More tokens from the same CPUs and GPUs: Orion-CB's single-hop Scorpio fabric delivered up to 89% higher tokens per second in an 8-GPU configuration. [1] * Future-ready chassis: Hyve Solutions Apex MGX 6U AI server chassis is built to support higher-power accelerators as they become available. SAN JOSE, Calif. and FREMONT, Calif., Nov. 16, 2026 (GLOBE NEWSWIRE) -- Astera Labs, Inc. (Nasdaq: ALAB), a leader in semiconductor-based connectivity solutions for rack-scale AI infrastructure, and Hyve Solutions Corporation, a U.S.-based design and manufacturing partner for compute, storage, and networking systems in AI, cloud, and digital infrastructure, and a wholly owned subsidiary of TD SYNNEX Corporation (NYSE: SNX), today announced an expanded collaboration to help neoclouds and enterprises bring AI inference capacity online faster. Hyve will offer Astera Labs’ Orion Compute Board, Fabric Tray, and Link Board as part of the Hyve Solutions Apex platform at rack scale. Because the components are pre-validated within Astera Labs’ compute-agnostic reference architecture, customers can expect shorter integration and qualification cycles and the flexibility to choose the accelerators and CPUs that best fit their cost, lead time, and workload.

Get started

Create a free account to read the full story.

Sign Up

Already have an account? Log in