Upstream Joins Forces with Cisco Cloud Control to Power Physical AI Intelligence for Agentic Operations

As part of Cisco’s marketplace expansion, the integration lets AI agents in Cisco Cloud Control query Upstream’s Model Context Protocol (MCP) servers to access real-time quality, operational observability, and cybersecurity risk context inferred by Upstream’s live digital twins across physical AI assets.

Upstream today announced its onboarding into the Cisco Cloud Control Marketplace as part of Cisco’s ecosystem expansion. By integrating Upstream’s live digital twin technology into Cisco’s unified AgenticOps platform via Model Context Protocol (MCP) servers, joint operational, engineering, and SOC teams, alongside autonomous AI agents, can now query real-time behavioral state, quality metrics, safety alerts, and physical AI cybersecurity risk directly within Cisco Cloud Control.

At the core of the integration is Upstream’s proprietary live digital twin technology. Unlike static log aggregators or point-in-time vulnerability scanners, Upstream maintains a continuous, stateful digital twin for every connected asset, synthesizing telematics, API streams, and behavioral data. This engine enables Upstream to infer real-time operational context, providing end-to-end observability, quality tracking, safety monitoring, and cybersecurity threat detection across connected vehicles, autonomous mobility, robotics, smart industrial equipment, and telematics feeds.

As enterprise teams adopt autonomous AI agents to manage infrastructure through natural-language interfaces, physical AI operational data streams have remained a critical blind spot. Through Upstream’s MCP integration, when an AI agent or operator inside Cisco Cloud Control flags a performance anomaly, quality issue, or potential cyber breach, it can directly query Upstream’s MCP servers to pull deep investigation context derived from these live Digital Twins. This comes to life in Cisco AI Canvas, the troubleshooting workspace within Cisco Cloud Control. Without switching tools, operators and agents can investigate and resolve IT issues leveraging both Upstream and Cisco context together.

“Teams shouldn’t have to leave their primary workspace to understand operational quality, performance glitches, or cybersecurity risk across their Physical AI ecosystem,” said Yoav Levy, Co-founder and CEO of Upstream. “Upstream uses live digital twins to continuously infer deep behavioral context, physical asset health, and cybersecurity intent. By putting that intelligence behind Cisco Cloud Control’s AI agents via MCP, that context is available inside the exact conversational interface operators use every day. Whether investigating a cyber attack, a safety signal, or a fleet-wide quality defect, it’s there the moment an investigation demands it.”

Unifying enterprise IT telemetry with physical AI statefulness is critical because operational quality, observability, and cybersecurity are deeply intertwined. Software bugs affecting component quality, telematics disruptions, or malicious cyber exploits often manifest initially as subtle behavioral anomalies. Querying Upstream’s MCP servers allows Cisco Cloud Control agents to leverage Upstream’s live digital twin intelligence so quality issues, operational defects, and high-risk lateral cyber movement surface early, allowing teams to defend and optimize physical AI infrastructure at machine speed.