AurCore Technology introduced two AI-focused Ethernet switch platforms, the AES8000 Series 400G AI-Ready Switch and the AES9000 Series 800G Ultra-Performance Switch, aimed at AI and edge data center networks. The systems are built on Marvell switch silicon, and AurCore is working with PalC Networks around open networking software and integration.
The AES8000 Series 400G AI-Ready Switch is powered by the Marvell Prestera CX 8500 switch ASIC, which AurCore targets at edge and private data centers. AurCore’s AES9000 Series 800G Ultra-Performance Switch uses the Marvell Teralynx 10 Ethernet switch ASIC, positioning it for higher-performance environments including AI training and inference, general-purpose compute, and cloud data centers.
For data center operators building AI fabrics, the product split is a familiar one: a 400G class platform for private and edge deployments where cost, port mix, and operational simplicity tend to dominate, and an 800G class platform aimed at scaling out clusters where bandwidth pressure can get ugly fast. But switch silicon is only part of what determines whether a fabric is workable at scale; the network OS, telemetry, and operational model are where engineering teams feel lock-in and day-two pain.
AurCore is leaning into open networking, including support for SONiC, the open-source network operating system originally developed by Microsoft and used in hyperscale environments. The company’s partnership with PalC Networks centers on SONiC engineering and integration, with the goal of letting customers deploy open networking platforms and automate network infrastructure using open tools while maintaining control over the network architecture.
“The future of networking is Neo Cloud and next-gen AI Data Centers,” said MC Leo, CEO of AurCore Technology. CTO Bryan Wang added, “By combining advanced hardware with open networking software, AurCore enables organizations to accelerate AI training and inference while maintaining architectural freedom, intelligent telemetry and continuous performance optimization to yield ROI.”
Source: AurCore Technology
