Xsight Accelerate Program
Convert silicon performance into a solution-level advantage
Xsight Accelerate delivers programmable, line-rate solutions across data center, AI, cloud, and edge environments eliminating compromises between speed, power, and security while accelerating your time to market.
Built to enable
Wire-Rate Performance
Hardware acceleration and on-path CPU fabric delivers up to 800 Gbps or 12.8 Tbps. No compromise.
Programmable Freedom
Custom telemetry and new protocols without silicon re-spins. Your pipeline, your rules.
Extreme Efficiency
World-class processing at significantly lower TDP.
The program
The foundation for next-generation infrastructure
Xsight Accelerate significantly reduces time-to-market by providing a comprehensive suite of ready-to-deploy deliverables, including architectural blueprints, technical code, and integrated SONiC support for each solution.
This foundation allows organizations to stay agile in a shifting landscape, deploying flexible, high-density infrastructure that scales alongside the most demanding modern workloads.
Solution Catalog
Solutions launching across the program
Stay tuned as we dive deep into each solution through a series of strategic rollouts this year.
Top of Rack Switching
X2 delivers 12.8 Tbps in under 200W. Replace 2RU systems with 1RU switches and recover power for compute.
AI Networking
Open infrastructure that scales with your models, not against them.
Gateway + Network Packet Broker
Programmable packet brokering and gateway functions at line rate, without fixed-function constraints.
Broadband Infrastructure
Open, software-defined infrastructure for service providers modernizing connectivity networks.
Security Appliances
Wire-speed security for next-generation firewalls and security gateways.
Storage
Process storage workloads at 800 Gbps with line-rate encryption and NVMe-oF offload.
Host Virtualization
E1 DPU takes host virtualization further, offloading L4–L7 processing from the server, freeing host CPUs, and enabling more tenant workloads without the “slow path” overhead of traditional DPUs.
Smart NIC
64 Arm cores in the data path. 800 Gbps throughput. Standard Linux, DPDK, and SONiC.
Continuous Innovation
We are continuously expanding the program with new, innovative solutions.
Why Xsight Accelerate
Built different. Measured in results.
Performance claims need proof. Here's what the architecture delivers at the silicon level and what that means for your infrastructure.
800G Stateful Processing
The E1 DPU delivers 800 Gbps, redefining what efficiency means for DPU-class silicon.
12.8T Switching Capacity
X2 Universal Ethernet Switch hits 12.8 Tbps high-radix switching while staying under 200W.
Zero-Loss AI & Storage Fabric
Integrated PFC, ECN, and deep shared buffers guarantee zero packet loss for latency-sensitive AI training and NVMe storage traffic.
Run-to-Completion Architecture
Full control over the packet pipeline. Custom headers and new protocols without waiting for a silicon re-spin cycle.
1RU Smart Switch Consolidation
L4–L7 services integrated into a single 1RU platform, eliminating middleboxes and reducing rack footprint dramatically
Open SONiC-Native Architecture
Full integration with SONiC and standard NOS stacks. No vendor lock-in, seamless migration from legacy infrastructure.
Ready to start your journey?
Let's discuss what Xsight Accelerate can do for your infrastructure.