At Cisco Live 2025 in San Diego, Cisco unveiled an expanded portfolio of data center technologies, partnerships, and services aimed at powering the next wave of AI deployments—from hyperscale operators to emerging neocloud providers and enterprise IT. The launch builds on Cisco’s longstanding partnership with NVIDIA and responds to surging demand for AI-ready infrastructure that is secure, scalable, and operationally efficient. Cisco confirmed it has already exceeded $1 billion in AI infrastructure orders from hyperscalers in fiscal 2025, one quarter ahead of schedule.
Cisco and NVIDIA have worked closely for years to develop and validate joint solutions combining Cisco’s networking, compute, and security platforms with NVIDIA’s GPUs and AI software stack. In 2023, the companies announced their first collaborative reference designs for AI infrastructure, and earlier this year they co-developed the Cisco Secure AI Factory for NVIDIA, integrating Cisco UCS servers, Nexus switches, AI Defense, and Hypershield with NVIDIA NIM and NeMo microservices. At Cisco Live 2025, the companies showcased their first technical demonstration of Spectrum-X Ethernet networking using Cisco Silicon One G200-based switches and NVIDIA NICs, supporting both NX-OS and SONiC environments.
“The world is moving from chatbots intelligently answering our questions to agents conducting tasks and jobs fully autonomously. This is the agentic era of AI,” said Jeetu Patel, EVP and Chief Product Officer at Cisco. “As billions of AI agents begin working on our behalf, the demand for high-bandwidth, low-latency, and power-efficient networking will soar. Cisco is at the forefront, delivering advanced, secure networking technology that’s foundational to the AI-ready data centers of the future.”
Cisco’s latest infrastructure upgrades span AI workload optimization, fabric convergence, operational simplicity, and security integration. The new Unified Fabric Experience converges ACI and NX-OS VXLAN EVPN environments into a single fabric with unified control, data, and policy planes. The Nexus Dashboard, updated for July 2025, now provides a single-pane-of-glass view across LAN, SAN, IPFM, and AI/ML fabrics.
For real-time AI workload management, Cisco Intelligent Packet Flow dynamically steers traffic across congested network paths based on telemetry and GPU workload awareness. Customers gain end-to-end visibility across AI clusters, enabling proactive issue detection. Combined with Cisco’s AI Defense and Hypershield security platforms, enterprises can now deploy and safeguard AI workloads built on open models.
Other notable announcements:
- Expanded AI PODs: Modular, validated architectures optimized for training and fine-tuning workloads, now including support for the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU paired with Cisco UCS C845A M8 servers.
- 400G BiDi Optics: New bidirectional optics allow customers to upgrade to 400G while preserving existing multimode duplex fiber infrastructure, simplifying transitions without new cabling.
- NVIDIA Spectrum-X Integration: Demonstrated live at Cisco Live, Cisco Silicon One-powered switches successfully interfaced with NVIDIA NICs, supporting Spectrum-X AI Ethernet networking.
Cisco is also staking a leadership role in the emerging neocloud market—GPU-as-a-service and AI IaaS providers built from the ground up to meet AI demand. These providers are building next-gen data centers with Cisco’s networking, compute, observability, and security stack as a foundation.
Key partnerships announced:
- HUMAIN (Saudi Arabia): Cisco will supply Nexus switching, UCS servers, Hypershield, and Splunk to help HUMAIN build what it calls “the world’s most open and cost-efficient AI infrastructure.”
- G42 (UAE): Cisco and G42 announced a joint initiative to develop AI infrastructure serving both public and private sectors across the Middle East.
- Stargate UAE: Cisco has joined this regional AI compute consortium as a preferred technology partner for next-generation AI cluster deployments.
Other major Cisco product updates and milestones include:
- New Cisco 8000 Series Routers: Powered by Silicon One, these converged access and edge routers improve efficiency and enable dynamic traffic handling across high-density AI applications.
- Crosswork AI Agent Framework: A new multi-agent architecture allows service providers to deploy Cisco- and customer-built agents to automate network operations and accelerate decision-making.
- Satellite Network Integration: Cisco unveiled support for seamless integration between terrestrial and non-terrestrial networks, including satellite backhaul, enabling new monetization opportunities in maritime, IoT, and disaster recovery scenarios.
- Cisco surpassed $1B in AI infrastructure orders from hyperscalers in Q3 FY25
- Unified Nexus Dashboard delivers single-pane management for ACI, SAN, IPFM, and AI fabrics
- First Spectrum-X Ethernet integration demo with Cisco Silicon One + NVIDIA NICs
- New AI PODs support RTX PRO 6000 Blackwell GPUs on UCS C845A M8 servers
- Cisco partners with HUMAIN, G42, and Stargate UAE to build AI-native neoclouds
- 400G BiDi optics simplify transition from legacy fiber to high-speed networks
- Agile Services Networking now supports satellite-terrestrial convergence and AI agent orchestration
