Cisco Systems, Inc. (CSCO) is making a serious play for the AI infrastructure market. At Cisco Live EMEA on Tuesday, the networking giant unveiled a suite of innovations designed to handle the massive computational demands of modern AI data centers, headlined by its new Silicon One G300 chip.

The G300 Chip: Built for AI Scale

The star of the show is Cisco's Silicon One G300, a high-performance switching silicon with 102.4 Tbps switching capacity. That's a lot of throughput, and it's specifically engineered to power the backend networking for massive AI clusters where thousands of GPUs need to talk to each other constantly.

But raw speed isn't the only selling point. The G300 delivers 33% increased network utilization and cuts job completion time by 28%, which translates directly to more profitable AI data centers by squeezing more value out of expensive GPU hours. The chip's programmability also means it can adapt to new networking functions over time, giving customers long-term infrastructure value rather than obsolescence.

New Systems Tackle the Heat Problem

Cisco also introduced new systems—the N9000 and 8000 series—powered by the G300 to support AI workloads in both hyperscale and enterprise environments. These aren't your typical network switches. They're designed to handle the extreme power and thermal demands that come with AI processing.

The systems incorporate liquid cooling technology that enhances bandwidth density while improving energy efficiency by nearly 70%. Given that cooling costs can make or break the economics of an AI data center, that efficiency gain matters.

Cisco rounded out the hardware announcement with new optics including 1.6T OSFP and 800G Linear Pluggable Optics. These components are critical for scaling AI networks, reducing power consumption while maintaining high-performance connectivity.

Nexus One: Simplifying the Complex

Managing AI infrastructure is notoriously complicated, so Cisco advanced its Nexus One platform to provide unified management and AI networking capabilities. The platform brings together Cisco's silicon, systems, optics, and software into a single integrated solution.

Nexus One features AI Job Observability and integration with Splunk, offering deep insights into network performance and AI workload behavior. This unified fabric approach gives customers flexibility in how they deploy and adapt their networks, with built-in automation and customization options.

JPMorgan Sees Upside Potential in 2026

The Street is taking notice. JPMorgan hosted an investor meeting with Cisco following the company's fiscal first-quarter earnings release, and analyst Samik Chatterjee had some interesting observations.

Chatterjee emphasized that Cisco's fiscal 2026 guidance for hyperscaler AI orders, exceeding $4 billion, represents a "minimum-level target." The analyst also noted that the $2 billion AI opportunity pipeline, which includes NeoCloud and Enterprise AI, was primarily generated in just the last quarter. Notably, Sovereign AI opportunities aren't yet included in these estimates, suggesting potential upside.

The analyst did point out one challenge: to hit its target growth rate of 15% to 17% in security, Cisco will need a significant shift toward new products.

CSCO Price Action: Cisco Systems shares were up 0.25% at $87.00 during premarket trading on Tuesday. The stock is trading near its 52-week high of $87.24.