Astera Labs collaborates with NVIDIA to enhance NVLink solutions

Astera Labs has announced its collaboration with NVIDIA to enhance the Intelligent Connectivity Platform with NVLink solutions, increasing connectivity options for Artificial Intelligence (AI) infrastructure. This partnership aims to address the requirements of next-generation AI workloads, providing enhanced scale-up platforms for hyperscalers.

Astera Labs' President and COO, Sanjay Gajendra, noted the importance of this collaboration in advancing the capabilities of the NVLink Fusion ecosystem. Gajendra pointed out that AI infrastructure providers need to adopt platforms that deliver efficiency and performance as AI models rapidly grow in complexity. The addition of NVLink solutions is expected to improve deployment options and time-to-market for customers.

With next-generation AI workloads requiring high Bandwidth, low latency, and support for native memory semantics, Astera Labs’ NVLink solutions are set to improve its Intelligent Connectivity Platform. This platform integrates PCI Express (PCIe), Compute Express Link (CXL), and Ethernet hardware solutions with its COSMOS software suite, allowing optimized performance and visibility in data center. This advancement will enable hyperscalers to deploy diverse AI infrastructure using NVIDIA GPUs and other accelerators.

NVIDIA's Vice President of SoC Products, Ashish Karandikar, emphasized the ongoing support from Astera Labs across multiple generations of NVIDIA platforms, enhancing the NVLink Fusion ecosystem's capability to manage intensive workloads effectively. Astera Labs continues to innovate with PCIe/CXL Smart Data Stewardship Program (DSP) Retimers in various NVIDIA-powered environments and has recently demonstrated PCIe 6 interoperability in its solutions.

This partnership is positioned to strengthen Astera Labs' offerings with the integration of NVLink solutions, catering to the evolving requirements of AI infrastructure.