AI Infrastructure, Orchestration & On-Prem AI | Stockholm MLOps Event #28 | February 27 2026

Insights from the Stockholm MLOps community

Stockholm MLOps #28 focused on a topic becoming increasingly important across enterprise AI:

On-prem MLOps and operational AI infrastructure.

Across the event, the discussions centered around:

  • Open Source AI stacks

  • orchestration

  • AI factories

  • infrastructure ownership

  • operational scalability

  • and secure AI deployment in regulated environments

The event reflected a broader shift happening across the industry:

AI is increasingly becoming an infrastructure and operational challenge — not just a model challenge.

Summary — Key Signals from the Event

  • Infrastructure ownership is becoming strategic

  • Open Source AI stacks are becoming foundational building blocks

  • AI orchestration is emerging as a core operational layer

  • Governance and trust are moving closer to infrastructure

  • Operational maturity increasingly determines AI success

Speakers & Companies Featured at Stockholm MLOps #28

Andreas Bergqvist — AI Specialist Sales EMEA, Red Hat

Andreas Bergqvist presented Red Hat’s approach to Open Source AI infrastructure and operational AI systems.

The session focused on:

  • modular AI building blocks

  • orchestration

  • trust and security

  • and operational AI deployment from day one

The talk strongly reflected how Open Source infrastructure is increasingly becoming part of enterprise AI strategy.

Leif Nordlund — AI Specialist, Lenovo

Leif Nordlund focused on practical infrastructure considerations for building AI systems on-prem.

The session covered:

  • GPU infrastructure

  • scaling from compact systems to large GPU environments

  • operational infrastructure planning

  • and lessons learned from AI infrastructure deployments over the past decade

The discussion highlighted how infrastructure planning increasingly shapes AI strategy itself.

Simon Janeck — Head of R&D and AI, Aixia

Simon Janeck focused on operational AI delivery and AI factory concepts.

The session reflected the growing need for:

  • deployment consistency

  • operational support

  • orchestration

  • infrastructure management

  • and production-ready AI systems

The “AI Factory” framing highlighted the shift from isolated PoCs toward operational AI environments.

Johan Lennartson — AI Strategist, Sogeti

Johan Lennartson shared lessons from supporting a large government agency on its AI journey.

The discussion focused on:

  • secure AI deployment in the public sector

  • governance

  • hardware constraints

  • innovation management

  • and long-term operational value creation

The session highlighted how regulated environments require very different operational AI decisions compared to startup-style deployments.

Key Operational Signals from the Event

1. Infrastructure Ownership Is Becoming Strategic

Several sessions reflected a growing move toward infrastructure ownership and operational control.

Organizations increasingly want:

  • governance

  • portability

  • operational flexibility

  • and infrastructure independence

Infrastructure decisions increasingly shape AI capability itself.

2. Open Source AI Is Becoming Operational Infrastructure

The event strongly reflected growing interest in Open Source AI stacks and modular infrastructure.

The focus increasingly shifts toward:

  • interoperability

  • flexibility

  • orchestration

  • and avoiding long-term lock-in

Open Source is increasingly viewed as operational infrastructure rather than experimentation tooling.

3. AI Orchestration Is Becoming a Core Engineering Layer

As organizations move beyond isolated AI experiments, orchestration increasingly becomes necessary for:

  • deployment consistency

  • lifecycle management

  • operational scaling

  • workload coordination

  • and infrastructure governance

AI systems increasingly resemble platform engineering systems.

4. AI Factories Require Operational Discipline

The “AI Factory” framing reflected a broader shift toward repeatable operational AI systems.

Scaling AI successfully increasingly requires:

  • deployment standards

  • observability

  • operational support

  • governance

  • and infrastructure planning

Operational maturity increasingly becomes the differentiator between experimentation and production.

5. Governance and Security Are Moving Into Infrastructure

Security and trust repeatedly appeared as infrastructure concerns rather than downstream policy discussions.

This increasingly affects:

  • deployment architecture

  • orchestration

  • runtime operations

  • access control

  • and operational governance

Governance increasingly becomes part of the runtime environment itself.

6. Hardware Constraints Continue to Shape AI Strategy

The event repeatedly highlighted how infrastructure limitations continue to shape operational AI decisions.

GPU allocation, hardware availability, and infrastructure planning increasingly affect:

  • scaling strategies

  • deployment decisions

  • and operational priorities

Infrastructure availability may increasingly become more important than model capability alone.

Tensions Emerging from the Event

Cloud vs On-Prem

Organizations increasingly want cloud flexibility while maintaining operational control.

Sovereignty vs Convenience

Managed AI services simplify operations, while on-prem infrastructure provides greater ownership and governance.

Portability vs Optimization

Modular systems improve flexibility, while optimized deployments often depend on tightly integrated environments.

Innovation Speed vs Governance

Governance requirements can slow experimentation unless integrated directly into operational workflows.

Scaling Ambition vs Operational Readiness

Many organizations want enterprise-scale AI before operational processes are mature enough to support it.

Why These Signals Matter

Stockholm MLOps #28 reflected a broader industry shift:

AI is increasingly becoming infrastructure.

The next phase of MLOps increasingly revolves around:

  • operational scalability

  • orchestration

  • governance

  • runtime reliability

  • infrastructure ownership

  • and long-term operational management

The event highlighted the growing convergence between:

  • AI engineering

  • infrastructure strategy

  • platform engineering

  • and enterprise operations

Related Insights

- Sovereign AI in Production | Stockholm MLOps #31

- Optimizing Inference | Stockholm MLOps #29
- AI Infrastructure & Operational Scale | Stockholm MLOps #27

Join the Community

👉 Explore all events: https://www.meetup.com/stockholm-mlops-community/

👉 Explore this event: On-Prem MLOps with Lenovo & Red Hat #28

Event Details

  • Location: AI Sweden, Stockholm

  • Date: February 26, 2026

  • Companies represented: Red Hat, Lenovo, Aixia, Sogeti

  • Topics: On-Prem AI, AI infrastructure, orchestration, Open Source AI, operational MLOps

  • Event: Stockholm MLOps #28

Previous
Previous

Inference Optimization in Production | Stockholm MLOps #29 | March 12, 2026

Next
Next

Operationalizing AI at Scale | Stockholm MLOps Event #27 | February 12 2026