AI Infrastructure Takes Center Stage at SOOCon25
On February 4, 2025, leading voices in AI infrastructure and open hardware gathered at State of Open Con 2025 (SOOCon25) at Doddington Forum 1, Bancroft, St Paul’s. Among them was Dan Scarbrough, Chief Commercial Officer at Stelia, who joined Amanda Brock, CEO of OpenUK, in a critical discussion on the future of AI infrastructure and data centres.
SOOCon25, or the State of Open Conference 2025, is the UK’s premier open technology conference, organized by OpenUK, a non-profit representing the UK’s open technology sector. It focuses on bringing together the global open-source community to explore advancements and challenges in areas like open-source software, open hardware, open data, open standards, and AI openness. The event is designed to unite developers, contributors, maintainers, enterprises, policymakers, and legal experts under one roof to collaborate on the future of open technology.
It’s the third annual iteration of the event, following SOOCon23 and SOOCon24, and is strategically scheduled right after FOSDEM in Brussels, making it convenient for international attendees to travel between the two. The conference featured two plenary sessions and seven content tracks, covering topics such as software security, open finance, mobile technology, and sustainability in open tech, with a particular emphasis in 2025 on AI, skills development, and public sector use of open technology. 
Key Takeaways: Scaling AI Infrastructure for the Future
The session highlighted the rapidly evolving enterprise AI landscape, where organisations are shifting from AI experimentation to execution at scale. The panel emphasized:
- Scalability Challenges: AI models are advancing faster than traditional infrastructure can support, demanding more efficient compute orchestration.
- Latency Bottlenecks: Data mobility and inference optimization are becoming mission-critical for real-time AI decision-making.
- Sustainability & Open Hardware: The role of open-source innovation in building energy-efficient, scalable AI data centres.
Stelia’s Perspective: AI Execution Requires Purpose-Built Infrastructure
As AI adoption accelerates, AI models alone are not enough—they require an infrastructure layer that ensures seamless inference at scale. Stelia’s Dan Scarbrough underscored this shift, noting:
“AI’s commercial value isn’t just in building powerful models—it’s in executing them at scale. Without optimized infrastructure, inference remains a bottleneck.”
The discussion reinforced that AI infrastructure is no longer an afterthought—it’s the foundation that enables real-time AI execution.
Looking Ahead: The AI Economy is Defined by Execution
SOOCon25 made it clear: the next phase of AI isn’t about experimentation—it’s about execution. Enterprises must prioritise scalable inference, optimised data mobility, and real-time AI decision-making to stay ahead.
For those navigating this shift, Stelia continues to provide the infrastructure layer powering AI-first enterprises.
Get in Touch
Want to learn more about how Stelia’s AI infrastructure solutions can accelerate your enterprise AI initiatives? Reach out to us at connect@stelia.io or visit our website to explore how we can help you scale AI execution with purpose-built infrastructure.
