Cisco (AI Defense)

Cisco was founded in 1984 by Leonard Bosack and Sandy Lerner in San Jose, California, and grew to become the world's dominant networking infrastructure company before expanding into security, collaboration, and cloud software. The company trades on Nasdaq (ticker: CSCO) and reported approximately $55.8B in FY2025 revenue with approximately 86,200 employees. While not a pure-play AI security company, Cisco's security division — reinforced by its 2023 acquisition of Splunk for $28B — has become one of the industry's largest, and its move into AI security via the Cisco AI Defense product line represents a top-down enterprise platform strategy that smaller specialists cannot replicate.

Cisco's AI security strategy was formalized with the August 2024 announcement to acquire Robust Intelligence, a Cambridge/San Francisco-based AI model security startup with $44M in prior funding and pioneering work in AI firewalls and algorithmic red teaming. The acquisition closed in October 2024 and the Robust Intelligence team became the nucleus of Cisco Foundation AI — an internal research organization focused on open-source AI security models. In January 2025, Cisco launched Cisco AI Defense, an enterprise product for discovering, validating, and protecting AI applications; and in February 2026, it issued the largest expansion of AI Defense since launch, adding AI supply chain governance, MCP interaction controls, and runtime agent protection.

Cisco's technical differentiation in AI security is network-level enforcement — the ability to inspect, route, and block AI traffic at the infrastructure layer rather than at the application level. AI Defense is embedded in the Cisco Security Cloud and leverages Talos threat intelligence alongside the Robust Intelligence algorithmic red-teaming engine to continuously test AI models against hundreds of safety and security scenarios. Foundation AI released the first open-source reasoning model built for cybersecurity (Foundation-sec-8B-Reasoning) in January 2026, establishing Cisco as an active contributor to the open security AI ecosystem. The February 2026 Cisco AI Defense update added NVIDIA NeMo Guardrails integration, AI Bill of Materials (SBOM) for models, and multi-turn red teaming for agentic interactions.

Cisco competes in AI security as a platform incumbent against pure-play AI security vendors (HiddenLayer, Protect AI) and integrated platform rivals (Palo Alto Networks, Microsoft). Its structural advantage is distribution: Cisco's installed enterprise base, combined with Splunk's SIEM presence, means AI Defense can be positioned as an incremental add-on to existing security agreements rather than a new budget line item. The risk is that Cisco's AI Defense product — while technically sound — may be perceived as slower and less specialized compared to dedicated AI security startups. Foundation AI's open-source strategy is a deliberate counter: by building community credibility, Cisco establishes itself as the technically legitimate AI security choice for enterprises wary of startup lock-in.