
NetBox Labs used its first user conference, NetBox Evolve at the Kennedy Space Center in Florida, to unveil a set of additions to its Infrastructure Intelligence Platform. The company introduced NetBox Designs, NetBox Analytics, NetBox Agents and the NetBox Platform MCP Server, all aimed at helping network and infrastructure teams design, document, validate and automate complex environments with more consistency. NetBox Labs also announced partnerships with Nvidia and Grafana Labs.
Key facts at a glance
- NetBox Designs is a template system for infrastructure design. Teams define a site, rack or IP plan once and then render it repeatedly down to component configuration, including IP addresses, prefixes and VLANs.
- NetBox Analytics provides ready-made dashboards for common infrastructure questions, such as lifecycle status, port capacity and power draw.
- NetBox Agents are pre-built agents for change request review, lifecycle and version compliance, data quality management and data import.
- The NetBox Platform MCP Server is a fully managed server that connects customer-built agents to NetBox data and context.
- NetBox Designs is in public preview. Agents are read-only by default, and proposed changes run on a branch and pass through architect review and NetBox Validation before they can be applied.
- Nvidia now integrates NetBox as a native source of truth for Nvidia Configuration Manager, the network orchestration component of Nvidia DSX OS. Grafana Labs and NetBox Labs released an official NetBox plugin in the Grafana catalog.
NetBox started in 2016 as an open-source tool for keeping a record of what is on a network. That record-of-truth role has become more important as networks have grown more distributed, more virtualized and more dependent on accurate context for automation. NetBox Labs is the commercial steward of the project and says the software is used by more than 10,000 organizations. The company was spun out of DNS platform provider NS1 after IBM acquired NS1 in 2023. In June, NetBox Labs launched the Infrastructure Intelligence Platform to cover the full infrastructure lifecycle rather than only documentation.
Design templates aim to remove repetitive manual work
AI data centers are filled with racks that are nearly identical in design. Retail chains often operate small, medium and large store formats, each with a repeatable IT footprint. In both cases, engineers do not want to lay out every rack, port and IP plan by hand. NetBox Designs targets that repetition. Kris Beevers, CEO of NetBox Labs, described designs as templatized Lego bricks for infrastructure. A team defines a site, rack or IP plan once and sets its parameters. NetBox then renders the plan as many times as needed, down to low-level components and their configuration. That can include assigning IP addresses, prefixes and virtual LANs across a fabric.
The goal is to speed design to the component level, reduce manual errors and compress time to market. The workflow does not stop at design. NetBox Asset Lifecycle carries planned objects through bill of materials, purchase order, shipment and installation. Teams can also work with Designs through agents. An engineer might ask an agent to swap an Arista switch for a Cisco switch in a rack. The agent uses Designs and the NetBox data set of infrastructure components to make the change. Beevers said he anticipates teams will work that way with infrastructure design.
Analytics turns inventory data into operational answers
NetBox Analytics adds a set of ready-made dashboards for common infrastructure questions. Rather than requiring every team to build its own queries and visualizations, the platform provides views into lifecycle status, capacity and operational risk. Beevers said teams can ask questions such as what is going into support soon or how much port capacity or power draw is left. Those questions sound simple, but they are difficult to answer consistently when infrastructure data is spread across spreadsheets, monitoring tools, configuration management databases and ticketing systems. By placing those questions on top of NetBox data, the platform aims to give teams a shared operational picture.
Agents bring controlled automation to infrastructure work
NetBox Agents are pre-built agents for change request review, lifecycle and version compliance, data quality management and data import. The agents run through an orchestration platform and can use NetBox context to inspect proposed changes, identify compliance gaps or clean up records. NetBox Labs also introduced the NetBox Platform MCP Server, a fully managed server that connects customer agents to NetBox. MCP stands for Model Context Protocol, an emerging way for AI models and agents to access external tools and data sources. Beevers said NetBox has always been a context layer for agents and infrastructure.
The company is positioning agents as a way to make infrastructure data more useful, not as an autonomous replacement for engineering judgment. Controls surround the agents. They are read-only by default. Changes start in the NetBox intent model, run on a branch and go through architect review. NetBox Validation tests each proposed change against compliance and operational rules and sends feedback to the agent. Beevers said infrastructure is an area where an agent doing the wrong thing can break the company. That risk explains the branch-based workflow and the emphasis on review before execution.
Partner integrations extend the platform
NetBox Labs also announced partnerships with Nvidia and Grafana Labs. Nvidia now integrates NetBox as a native source of truth for Nvidia Configuration Manager, the network orchestration component of Nvidia DSX OS. The integration is significant because AI data centers require tight coordination between compute, networking and storage. Grafana Labs and NetBox Labs released an official NetBox plugin in the Grafana catalog, giving teams another way to bring NetBox data into observability dashboards. NetBox Labs expects to add more partnerships in the next few months as it builds out the platform.
Enterprise sprawl raises the urgency for rationalization
Most of the world’s infrastructure is not AI data center infrastructure. It is enterprise infrastructure at banks, retailers, manufacturers and public sector organizations. Beevers described the state of many of those environments as sprawl. They hold 20 to 30 years of technology decisions and tool adoption, layered on top of each other. AI is now adding workload to those networks, while rising vulnerability rates add pressure. Teams need to drive a high rate of change, patch to stay secure and scale effectively. The same pressures that created the need for automation over the years are compounding faster. As a result, Beevers said, teams have to rationalize their stacks for managing all this stuff pretty fast.
NetBox Labs plans to keep adding to the platform. Infrastructure is complex enough that the company expects to partner with others, and it remains an open-source business at its core. Beevers said the mission has not changed: to make it easier to build and manage complex networks and infrastructure, full stop.
Source:Network World News
