Cisco Systems

Cisco Systems, Inc. (NASDAQ: CSCO) is headquartered in San Jose, California, and generated approximately $54 billion in revenue in FY2024. Founded in 1984, Cisco built the foundational internet routing and switching infrastructure and still commands approximately 40% of enterprise networking revenue globally — the largest single vendor by installed base. In the AI era, Cisco's primary role is serving enterprise and telco customers deploying private AI infrastructure, rather than the hyperscaler core where Arista and NVIDIA InfiniBand dominate. Cisco's Nexus 9000 series switches — built on a combination of Broadcom Tomahawk and Cisco's own Silicon One ASICs — provide the data center fabric for Fortune 500 enterprise AI deployments, healthcare AI, and government clouds that require on-premises GPU clusters. Silicon One is Cisco's proprietary switch ASIC program, with dies fabricated at TSMC. Unlike Arista's merchant-silicon strategy, Cisco designs its own routing and switching chips to maintain differentiation on features like telemetry granularity, security group tags, and distributed buffering. The Silicon One G200 (2024) targets 25.6 Tbps switching capacity, competing with Broadcom's Tomahawk in hyperscale deployments and the Jericho series in service provider routing. Cisco's optical networking division — inherited from the $2.6B acquisition of Acacia Communications in 2021 — provides the coherent optical transceivers and line-card technology used in AI data center inter-cluster connectivity. As GPU clusters grow to hundreds of thousands of nodes, inter-data center bandwidth requirements grow proportionally, and Cisco's silicon photonics and dense wavelength division multiplexing (DWDM) technology serves that connectivity layer. Cisco is also a major provider to SoftBank, NTT, and other Japanese and Asian telcos building AI infrastructure. Its presence in enterprise, telco, and government sectors makes it a complementary player to Arista's hyperscaler focus, collectively covering the full spectrum of AI networking demand.