SAN FRANCISCO, June 23, 2026 (GLOBE NEWSWIRE) — Upbound, the company behind Crossplane, today released Modelplane, an open source control plane for AI inference fleets. Modelplane is designed to do for AI inference what Crossplane did for cloud infrastructure: provide a vendor-neutral, open source orchestration layer that unifies fragmented, heterogeneous environments under a single system of control.
AI inference is undergoing the same transformation that cloud infrastructure experienced a decade ago. Open-weight models have expanded who runs AI — neoclouds, regulated enterprises, and AI-native companies now operate their own GPU fleets across multiple clouds and on-premise environments. Cluster-level tooling has matured rapidly. But no open source system manages these fleets as a whole: placing model deployments across available capacity, autoscaling replicas based on load, provisioning and managing the underlying infrastructure, caching and distributing model weights, and routing inference through a unified gateway.
Modelplane fills that gap. Built on Crossplane, a CNCF Graduated project that serves as the foundation for internal platforms at thousands of organizations, including Apple, Nike, JPMC NASA Science Cloud, Elastic, Grafana, SAP, IBM, and Akamai, Modelplane inherits a mature infrastructure orchestration framework and extends it to the inference layer. It sits above existing serving engines, schedulers, and gateways composing them into a unified fleet-wide system.
“Open-weight models are changing who runs AI. Over the next few years, many more organizations will run inference on infrastructure they own and control, and they’ll all encounter the same problem: inference doesn’t stay on one cluster. Kubernetes became the standard control plane for compute. Crossplane extended that model to cloud infrastructure. AI inference needs the same layer. We’ve spent years watching operators build fleet-level inference platforms on top of Kubernetes and Crossplane. Modelplane is our contribution back to the community: an open control plane for AI inference, shaped by those real-world patterns and built in the open from day one.”
— Bassam Tabbara, CEO and Founder, Upbound
Modelplane is available today as an early open source developer release under the Apache 2.0 license. Upbound intends to develop the project in the open with the community as it evolves toward production readiness and plans to donate it to an open source foundation later this year. A control plane that sits above the ecosystem should belong to the ecosystem.
Supporting Industry Quote
“The cloud native ecosystem has always believed that neutral and open infrastructure wins at the orchestration layer. We saw it with Kubernetes for compute and Crossplane for cloud infrastructure. Modelplane applies that same principle to AI inference built on trusted CNCF projects. We are excited to see this space evolve and encourage upstream contribution over the long term. Organizations running inference at scale need an open control plane they can trust and Modelplane is a serious effort to build exactly that.”
— Chris Aniszczyk, CTO, Cloud Native Computing Foundation
What Modelplane Does
Modelplane is not another inference engine, scheduler, or gateway. It operates at the fleet level, above the cluster, and provides:
- Fleet-wide model scheduling: decides where models run across available and compatible GPU capacity, clouds, and regions
- Autoscaling: scales model replicas based on load and cost optimization targets
- Infrastructure lifecycle management: provisions and manages clusters, nodes, and accelerators
- Weight caching and distribution: efficiently propagates model weights across the fleet
- Unified inference gateway: routes requests with fallback to managed providers; enforces cost, compliance, and sovereignty policies
Built For Organizations That Run Their Own Inference
Modelplane is built for organizations that operate their own inference rather than rely on a managed API provider. It is designed to be neutral across models, infrastructure providers, accelerators, clouds, and serving stacks. Three segments define the target market:
- Neoclouds and AI factories — organizations building managed inference services on their own hardware who need a fleet control plane without building it from scratch
- Regulated and sovereign enterprises — organizations where inference must run inside infrastructure they control, under their own governance, due to sovereignty, security, or compliance requirements
- AI-native companies — organizations with significant inference spend that are moving to open-weight models and need control over cost, performance, and model differentiation
Availability
Modelplane is available today at https://github.com/modelplaneai/modelplane under the Apache 2.0 license. Documentation and getting-started guides are available at docs.modelplane.ai. Organizations interested in design partnership or commercial evaluation should contact Upbound at https://www.upbound.io/contact-us
Blog
About Upbound
Upbound is pioneering infrastructure platforms for the agentic AI era, serving Fortune 500 companies and platform engineers across more than 100 countries. The company empowers infrastructure and platform teams with intelligent control planes, based on Kubernetes and Crossplane, that provision, operate, and adapt so platforms are ready for both humans and AI agents. Upbound is the creator and primary maintainer of Crossplane, the popular open-source framework for building cloud-native control planes, with over 100 million downloads and adoption by more than 1,000 organizations in production worldwide. Founded in 2017 and headquartered in San Mateo, CA, a Series B startup backed by GV (formerly Google Ventures), Altimeter Capital, and Intel Capital, Upbound has raised $69M to date. For more information, visit http://www.upbound.io.
A photo accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/7b648025-790b-4908-ad37-9bd375d26ae2
