Astera Labs reported that hyperscale customers have begun testing its UALink interconnect technology to support large-scale AI model training. The company disclosed that over 20 hyperscale data centers have initiated proof-of-concept trials since mid-2025, representing a 150% increase in customer engagements compared with the prior six months. Management expects initial revenue from UALink deployments to begin in early 2026, projecting that UALink could contribute up to 25% of total product revenue by the end of that year. This growth underscores the importance of high-performance connectivity in AI infrastructure, with UALink’s low-latency, high-bandwidth design positioned to address bottlenecks in multi-GPU configurations.

Astera Labs will participate in two major investor conferences in Q1 2026: the 28th Annual Needham Growth Conference on January 14th in New York and the Morgan Stanley Technology, Media & Telecom Conference on March 3rd in San Francisco. The company’s presentations are scheduled for 10:15 a.m. Eastern Time and 1:05 p.m. Pacific Time, respectively. A live webcast of both sessions will be available on Astera Labs’ investor relations website. These events follow a year in which Astera Labs expanded its COSMOS software suite and secured three new custom connectivity contracts with leading cloud service providers, highlighting its strategy of integrating UALink, PCIe®, CXL®, NVLink and Ethernet into scalable AI infrastructure solutions.