Astera Labs Expands Connectivity Portfolio with Custom Solutions

Astera Labs Inc., an active player in semiconductor-based connectivity solutions for rack-scale AI infrastructure, announced plans to deliver custom connectivity solutions, a new offering designed to address the increasing complexity and diversity of next-gen AI infrastructure featuring multiple types of computing resources to optimize performance, energy efficiency, and cost for a wide range of workloads.Building on the May 2025 Link opens in a new tabannouncement of joining the Nvidia NVLink Fusion Ecosystem,, the company is now focused on collaborating with hyperscaler partners to develop custom solutions designed to support NVLink connectivity.

As AI infrastructure grows increasingly complex, hyperscalers need custom connectivity solutions tailored to their unique architectural requirements and workload-specific performance goals. Astera Labs aims to address this demand with custom connectivity solutions-an additive and complementary extension to the company’s industry-standards-based Intelligent Connectivity Platform. Custom solutions enable infrastructure providers to design heterogeneous systems that integrate diverse accelerators into unified, high-performance rack-scale platforms.

“Hyperscalers are accelerating the deployment of next-gen AI infrastructure and require proven connectivity partners who can consistently deliver at cloud scale,” said Sanjay Gajendra, president and COO, Astera Labs. “Our collaboration with Nvidia enables greater flexibility for XPU connectivity through NVLink Fusion enabled solutions. With NVLink Fusion support, our custom connectivity solutions will be engineered to sustain multiple terabytes per second of low-latency data throughput and are expected to be additive to our overall footprint in next-gen heterogenous AI infrastructure.”

The new custom connectivity solutions will leverage Astera Labs’ proven expertise in delivering purpose-built connectivity solutions and COSMOS software-defined architecture for the world’s leading hyperscalers and AI platform providers. Newly acquired capabilities, including photonic chiplet technology, are expected to enhance the company’s vision to provide a total offering for hyperscalers deploying standard, custom, and hybrid AI infrastructure.

“NVLink Fusion, combined with Astera Labs’ purpose-built connectivity solutions, gives customers access to the industry’s best scale-up interconnect with NVLink in a proven rack-scale architecture,” said Ashish Karandikar, VP, Engineering, Nvidia. “With Astera Labs joining our growing NVLink Fusion ecosystem, innovators have more choice in bringing semi-custom AI systems to market faster – with higher bandwidth, lower latency performance.”