NetBox was not originally designed to manage AI infrastructure. When the project was first released as an open source tool in 2016, the goal was straightforward: Give network teams a reliable record of what was on the network. The project spread well beyond that original scope. It is now embedded across more than 10,000 organizations, from enterprise networks to AI data centers running some of the most demanding infrastructure builds in the industry.
As the technology marks its 10-year anniversary, NetBox Labs is expanding the platform further with the launch of its Infrastructure Intelligence Platform. Key capabilities in the platform include:
- NetBox Data Exchange (NDX): A database of infrastructure component metadata covering device logical characteristics, lifecycle dates, environmental data, and observability profiles across tens of thousands of device types.
- NetBox Asset Lifecycle: A procurement pipeline connecting network design to physical deployment through bills of materials, purchase orders, shipment tracking, and spares management against planned DCIM objects.
- NetBox Validation: Pre-change compliance and safety verification against regulatory frameworks and organizational policy, with self-correction capabilities for AI agents,
- NetBox Labs Platform MCP Server and Agent Skills: An enterprise-grade hosted MCP server that exposes the full platform to any MCP-compatible agent, alongside an open source library of agent skills.
“It is not at all uncommon for me to hop on with the VP of infrastructure in some large enterprise and hear, ‘I have 14 network observability tools right now. Help,'” Kris Beevers, CEO of NetBox Labs, told Network World. “Tool sprawl is the problem at the moment that is holding these teams back. So, we’re giving them a cohesive platform that they can consolidate against, that works well together, which is really driving velocity for us.”
A decade of NetBox
Beevers didn’t create the open-source NetBox project. He first encountered it while he was leading DNS platform provider NS1. “We kept finding NetBox at such a volume and frequency that it caused us to say, what is this thing? Why is it everywhere?” Beevers said.
NS1 hired the core open source contributors around 2020 and began building commercial tooling around the project. NetBox Cloud launched in 2021. Then IBM acquired NS1 in 2023, and NetBox Labs was spun out as an independent company in 2023, led by Beevers.
NetBox Labs has continued to expand both the open-source project as well as a growing set of commercial offerings. In the last several years, in particular, there has been a focus on helping to enable AI infrastructure buildout as well as providing AI-enabled tooling.
NetBox is deployed at more than 10,000 organizations across enterprise networks, AI data centers and OT environments. The project has more than 20,000 GitHub stars across 350-plus releases, with contributions from nearly 400 developers.
That portfolio of products, combined with an accelerating infrastructure market, created the need for a unifying platform story. The Infrastructure Intelligence Platform is the answer to that.
“The rate of change of infrastructure is super fast, and the time pressure around infrastructure is super high,” Beevers said.
The Infrastructure Intelligence Platform: From design to deployment
NetBox Labs is positioning the platform to span the full infrastructure lifecycle, from initial procurement through decommission. That scope extends well beyond NetBox’s traditional role as a network inventory system, now reaching into procurement teams and buying centers.
“The infrastructure intelligence platform now covers that whole array of jobs and spans that whole lifecycle — from as early as design and procurement, racking and stacking, and plugging in, through to configuration and lifecycle management of the infrastructure across every dimension of operations for our customers,” Beevers said.
NDX provides the device metadata foundation that feeds the lifecycle workflow. Environmental and lifecycle data, covering heat output, end-of-support dates and physical specifications, flows into Asset Lifecycle, where it drives procurement planning, bill of materials generation and hardware reconciliation against the network design.
Maintaining that data thread through to production is where Validation comes in. Validation addresses two questions: whether the intended configuration meets compliance requirements and policy, and whether a planned change is safe before it reaches production. As agents take on more infrastructure work, the pre-change guardrail becomes more critical.
Validation pairs with the existing NetBox Assurance product. Where Validation operates on design intent, Assurance checks whether operational infrastructure matches that intent.
“So, is my intent correct? And is the infrastructure operating in line with my intent? Those are the two halves of this,” Beevers said. “Validation and assurance go really well together.”
Agentic infrastructure and the road ahead
AI infrastructure is leading to network buildouts, but there is also a need to enable agentic AI for Netbox as well. The platform is designed to support two concurrent access paths: full agent addressability and improved interfaces for human operators.
“The general philosophy that we’re pursuing right now is what we call a two front doors philosophy,” Beevers said. “We need everything in our platform to be accessible and operable by agents, because increasingly they’re going to do more of the work. But at the same time, we’re investing a lot, in up-leveled, more strategic experiences for the humans operating this stuff, so that they can drive the strategy and the shape of the work that’s being done by the agents.”