There is a growing realisation across the industry that we’ve reached an artificial intelligence (AI) tipping point.
For the last couple of years, the conversation has been dominated by the question of what is possible with AI. Today, that question has shifted to, “now that we have it, how do we control it?” We’re also moving from a world where computation was plentiful to one where it’s becoming restricted again — not just by the availability of GPUs, but by the boundaries of geography, regulation and trust.
Many organisations are putting on the brakes, realising that a “cloud-only” AI strategy can introduce complexities regarding data sovereignty and intellectual property control. If you don’t understand your dependencies, your technology stack, where your data resides and how it is being used, you may be outsourcing your future.
This is why Red Hat, in collaboration with Deloitte and Intel, has introduced a blueprint for sovereign AI-ready private cloud.
Sovereignty: Resilience is permanent
Digital sovereignty isn’t a static compliance checkbox, it’s a strategic choice for independence. Recently, we’ve seen how shifts in the global landscape can make a fully cloud-focused strategy outdated. Laws will evolve and change over time, but your need for organisational resilience is permanent. This isn’t about reacting to change, but anticipating it. Resilience depends on understanding risks, reducing dependency and maintaining flexibility across environments.
To achieve this, AI must follow your infrastructure, not the other way around. Whether it’s in a local data centre or running at the edge, the platform must remain consistent. Relying heavily on “black box” external APIs can limit an organisation’s visibility and control over its AI innovation.
The open source safeguard
We believe AI without open source poses more risks than benefits to production systems. Because generative AI is largely built on trial and error rather than pure theory, where we could understand all the scenarios, we don’t always fully understand why a model behaves a certain way, especially if we don’t know how it was created. If we don’t build these systems in the open, we can’t effectively mitigate undesirable behaviours such as hallucinations, bias or security risks.
Our collaboration with Deloitte and Intel helps provide that transparency. As Red Hat has always done, we’re embracing the rapid innovation of the open-source community and making it enterprise-ready, governed and predictable, while working closely with global communities and academia.
A unified, governed stack
The sovereign AI-ready private cloud reference architecture is designed to solve the “where” and “how” of AI by bringing together Intel’s hardware-rooted trust with Red Hat OpenShift’s orchestration capabilities and Deloitte’s operational expertise.
- Smart infrastructure: Using Intel® Xeon® processors and Intel® TDX, we provide a programmable backbone where security protections are rooted in the silicon.
- Orchestration: OpenShift provides a consistent hybrid cloud foundation, so your AI workloads can run across diverse environments—private, hybrid, or edge—with enhanced portability.
- Model tuning and serving: By leveraging frameworks like OpenVINO and Red Hat AI are designed to help you tune and serve models — often those under 13B parameters — on CPU-based infrastructure, which can help improve your overall AI cost-efficiency.
Your AI, your rules
The shift we’re seeing is a return to balance. Organisations want the agility of the cloud with the control of the data centre. By building on a sovereign, open foundation, you’re helping your organisation remain the master of its own success.
The tools are ready. The blueprint is here. It’s time to take the driver’s seat.
The above commentary is contributed by Red Hat EMEA Chief Technology Officer Julio Guijarro
