Three announcements came out of SUSECON in Prague on April 22, and I was following them from a distance while covering Adobe Summit in Las Vegas. Reading the press releases back-to-back, the thing that stands out is not any single product. It is the physical scale that the stack now claims to cover. Switch runs what it calls AI Factories, hyperscale data centers built for the heaviest AI and simulation workloads on the planet. SUSE is providing the governed execution layer underneath those workloads. That is not a partnership announcement written for a press release. It is a production decision that was already made.

I covered the first wave of SUSECON announcements, the sovereignty story of VMware exits and digital resilience, in a separate post at shashi.co. These three are a different kind of argument. They are about what SUSE is positioning as the destination stack, from the data center all the way down to a factory floor temperature sensor.

The Digital Twin Is Already in Production

The Switch announcement is the one most likely to get misread as a proof of concept. It is not. Switch has deployed NVIDIA Omniverse libraries alongside SUSE AI, SUSE Rancher Prime, and SUSE Linux Enterprise Server (SLES) on shared NVIDIA DGX systems to run real-time digital twins of its own data centers. These simulations continuously ingest operational data to model thermal dynamics, power usage, and infrastructure performance before any physical change is made.

The architecture problem this solves is real. Large-scale AI and simulation workloads historically required siloed infrastructure: one set of systems for graphics rendering, a separate set for AI processing. Running both on shared NVIDIA DGX hardware, governed by SUSE AI, collapses that split. Switch is using the platform to model its own operations and to run internal AI models that automate routine management tasks.

"A new class of enterprise applications now requires language models, simulation, and rendering to converge within a single system rather than across disconnected silos."

— Zia Syed, Chief Technology Officer, Switch

The air-gapped capability matters more than it might appear. Enterprise buyers in defense, critical infrastructure, and regulated manufacturing cannot run production AI workloads on systems with open internet connectivity. A platform that handles language models, simulation, and rendering within a closed environment removes a constraint that has kept those buyers on the sidelines.

MCP Is Now an Infrastructure Control Plane

The second announcement requires more careful reading. SUSE announced that it is integrating the Model Context Protocol (MCP) across its portfolio, partnering with Amazon Web Services, Fsas Technologies (a Fujitsu company), n8n, Revenium, and Stacklok to enable AI agents to autonomously manage Linux and Kubernetes environments. MCP is an open standard that gives AI agents a standardized way to communicate with underlying systems, including SUSE Rancher Prime and SUSE Multi-Linux Manager, across any distribution.

The constraint this addresses is structural. Enterprises adopting agentic AI have hit a practical ceiling: agents can reason and plan, but they lack a secure, standardized path into low-level infrastructure. Without that path, autonomous operations remain aspirational. SUSE's claim here is that it is the only vendor positioned to extend that MCP layer across any Kubernetes distribution and any Linux distribution simultaneously, not just its own.

Mikel Elorza Peña, IT Architect at Grupo Eroski, put the business case plainly in SUSE's announcement: the MCP integration allows the team to select the most appropriate language model for each specific workload rather than being locked to a single model across all tasks. That is not a trivial flexibility. Model selection is already a cost and performance variable that enterprise buyers are actively managing.

Revenium's participation adds a dimension the other partners do not cover. When AI agents provision clusters or deploy applications autonomously, they incur real infrastructure costs in milliseconds, faster than any manual approval process can track. Revenium provides what it describes as financial guardrails enforced at machine speed, catching agent-generated cost exposure before it accumulates. That is a governance layer, not a feature. It signals that SUSE is thinking about agentic AI in terms of enterprise risk, not just automation upside.

Stacklok's involvement is worth noting separately. Craig McLuckie, Stacklok's chief executive, is one of the co-creators of Kubernetes. Stacklok operates a registry of vetted MCP servers, and the SUSE Multi-Linux Manager server is already listed. That registry is an early-stage trust mechanism for enterprises that need confidence in the MCP tools their agents are invoking. The fact that SUSE is in it at launch matters more than the feature description does.

The Losant Acquisition Becomes a Product

SUSE acquired Losant in February 2026. The Day Two announcement converts that acquisition into a named product: SUSE Industrial Edge. The framing at SUSECON is that SUSE's edge portfolio previously covered what it calls the Near Edge (telco) and the Far Edge (hospital monitoring, marine engines, retail). Losant fills what SUSE describes as the Tiny Edge, the layer of constrained sensors and industrial devices where foundational operational data is generated.

The Industrial Edge platform is protocol-agnostic, meaning it can normalize data from Siemens equipment, Beckhoff controllers, HVAC systems, retail kiosks, OPC Unified Architecture (OPC UA) sources, and others into a single unified view. The no-code and low-code visual workflow engine is aimed at operations teams, not developers, which is the correct target audience for a platform selling into manufacturing, logistics, and facilities management.

SUSE is also joining the Linux Foundation's Margo steering committee, a standardization initiative for industrial edge interoperability, and committing to open-source the Losant technology. The open-sourcing announcement is consistent with SUSE's pattern. Open source the foundation, build support and services revenue on top. Enterprises that bought Losant's commercial platform before the acquisition will want clarity on what the open-source transition means for their existing agreements.

I wrote about the Losant acquisition when it was announced in February. The strategic logic was clear then: SUSE needed the Tiny Edge to make its edge portfolio credible end-to-end. What SUSECON adds is the integration story. SUSE Industrial Edge connects to SUSE AI, which means a sensor anomaly at a factory site can, in principle, trigger an agentic workflow that correlates with system logs, identifies a corrective action, and submits a patch request, all without a human in the loop. That is not vaporware. Each layer in that chain is shipping today.