Cisco AI Unleashes Model Foundry and Expands Partner Tools to Power Agentic AI Infrastructure

Cisco AI Launches Model Foundry to Transform Partner Operations

Cisco unveiled Model Foundry at Partner Summit 2026, a new platform service within Cisco Cloud Control that grants partners access to Cisco's purpose-built AI models1

1

. The service lets partners discover, evaluate, fine-tuning, and deploy domain-specific models trained on four decades of networking, security, and operations expertise. Model Foundry addresses a critical gap: while general-purpose models struggle with specialized infrastructure tasks, Cisco's purpose-built models already understand network telemetry, CLI syntax, and troubleshooting topologies1

1

.

The platform provides OpenAI-compatible APIs, enabling seamless integration with existing AI agents and workflows. Partners receive tenant-scoped keys, per-model access controls, and complete usage visibility1

1

. For demanding use cases, fine-tuning capabilities allow partners to train models on customer-specific data including configs, log messages, alerts, and agent traces from thousands of devices.

Purpose-Built AI Models Outperform Generic Alternatives

Model Foundry launches with several specialized models designed for infrastructure management. Deep Network 1.4 handles network operations, root-cause analysis, and multi-vendor troubleshooting1

1

. Antares-3B focuses on agentic vulnerability localization, pinpointing source files most likely to contain specific vulnerabilities. Decision Model 1.0 serves as a fast decision layer for AI agents working on networks, selecting appropriate tools and determining when to escalate to human operators1

1

.

World Wide Technology tested Model Foundry with agentic troubleshooting harnesses. After fine-tuning the 31-billion-parameter Deep Network Model on agent and troubleshooting data, the model recommended exact correct fixes significantly more often than generic alternatives when tested against real Cisco devices on problems spanning BGP, OSPF, link failures, and router policy1

1

. The library also includes Foundation-Sec family, SecureBERT 2.0, and Cisco's open-weight Antares models.

Cisco Partners Gain New Revenue Opportunities Through Agentic AI

Cisco expanded its Cisco 360 Partner Program with new incentives recognizing advisors, developers, and distributors for their unique contributions2

2

. Ecosystem Deal Registration now rewards the influence that shapes deals, while dedicated paths for advisors and developers provide recognition for pre-sale roles. These updates create a unified value chain benefiting partners at every customer lifecycle step2

2

.

The agentic AI era creates opportunities across the customer journey: assessing AI readiness, modernizing infrastructure, integrating workflows, building skills, and delivering managed operations5

5

. A recent Cisco and Omdia study of 1,000 IT and network operations leaders found 51% already use agentic AI in production, with 84% expecting to reach an AI-led operating model within 12 months5

5

.

Cisco IQ Reaches 16,000 Customers Managing 20 Million Devices

Cisco announced general availability of Cisco IQ alongside Cisco Partner Support - Standard, allowing partners to own the complete customer experience under their own brand4

4

. Over 16,000 customers now run Cisco IQ across close to 20 million connected devices4

4

. The AI-powered platform determines which devices are affected by security advisories 96 times faster than historical methods, according to an IDC White Paper sponsored by Cisco4

4

.

Role-Based Access Control within Cisco IQ lets customers grant partners infrastructure visibility while maintaining full data control4

4

. Partners can integrate Cisco IQ's production APIs and the new Cisco IQ MCP Server into their own systems, weaving infrastructure intelligence into delivery workflows. The IDC study shows a 5.7 times faster ability to turn raw data into prioritized action plans, creating opportunities for 2.1x greater partner profitability4

4

.

Model Context Protocol Eliminates Custom Integration Complexity

Cisco introduced Model Context Protocol servers for its networking platforms, providing standardized bridges between AI agents and live network data3

3

. The technology eliminates the need for custom integrations and enables automated workflows for troubleshooting, reporting, and infrastructure management3

3

. Third-party applications can access live network telemetry through these standardized interfaces, accelerating AI agent deployment.

ServiceNow integration brings Cisco IQ intelligence into workflow automation. John Aisien, SVP of Central Product Management at ServiceNow, noted that bringing this intelligence into ServiceNow workflows turns insight into action, providing customers with more accurate asset data, faster issue response, and governed changes from request to closure4

4

. For managed service provider partners, this creates a scalable, repeatable way to run lifecycle management across every environment.

ThousandEyes Intelligence Embedded in Wi-Fi 7 Access Points

Cisco embedded ThousandEyes synthetic testing directly into Wi-Fi 6E and Wi-Fi 7 access points, enabling continuous background testing of application and network performance even when users aren't connected5

5

. The capability requires no additional deployment or configuration and is included with eligible Cisco Networking Essentials licenses starting in late fourth quarter 20265

5

.

ThousandEyes automatically discovers critical services, orchestrates active testing, and adapts as conditions change5

5

. Combined with infrastructure and experience telemetry, this intelligence helps teams and AI agents detect issues earlier, investigate causes, and validate recovery. The integration strengthens the foundation for autonomous operations by providing agents with better data to proactively fix issues before users experience them.

Quantum-Resilient Security Extends to Wireless Networks

Cisco introduced the industry's first Wi-Fi 7 access point with quantum-resilient capabilities, addressing concerns that quantum computers will crack current encryption algorithms3

3

. The technology combines Qualcomm-powered access points with client-side support from Intel wireless chipsets and Microsoft Windows software3

3

. This extends Cisco's quantum-readiness strategy across wireless networking, switches, and routers.

Vikas Butaney, Cisco's senior vice president of customer experience product management and marketing, emphasized that AI-powered threats have reduced the time between vulnerability detection and exploitation to hours3

3

. Cisco expanded its Live Protect security technology to additional switches, wireless controllers, and software-defined wide-area network platforms, deploying temporary Vulnerability Shields to protect infrastructure devices between vulnerability disclosure and permanent fixes3

3

.

AI-Ready Infrastructure Demands Platform-Centric Approach

Cisco Cloud Control provides teams and AI agents a unified environment to manage infrastructure and resolve issues2

2

. The platform integrates with Privileged Infrastructure Access through partnership with Teleport, bringing just-in-time, just-enough access to Cisco routers and switches2

2

. Network teams can keep critical work moving while privilege expires when work completes.

New AI-powered assessment playbooks within Cisco IQ help organizations evaluate quantum-resistant security readiness and identify devices running software approaching end of vulnerability support3

3

. Additional planned tools will assess software compatibility, prioritize bugs, and identify potential outage causes. The Cisco IQ AI Assistant gains access to device diagnostics and Technical Assistance Center knowledge, enabling more contextual troubleshooting with the ability to open, close, and escalate support cases directly3

3

.