By Sergio Gaga Huerta, Chief Technology Officer, Cloudera
Imagine waking up one day to find your data, the bloodline of your business, customer trust, and the nation’s digital sovereignty no longer under your control. Simply because you clicked “accept” on a cloud provider’s terms and conditions three years ago without reading them. While you can access and play with it, you have no control over the ecosystem around it or how it evolves and changes.
This isn’t a dream. It’s the daily reality for thousands of organizations across the Middle East and beyond.
The Middle East Public Cloud market was valued at USD 38.5 billion in 2024 and is projected to reach USD 235.7 billion by 2033, representing a compound annual growth rate (CAGR) of 19.86% from 2025 to 2033. Major U.S. firms still dominate the main global providers in the Middle East.
Cloud adoption in the Middle East has lagged behind other regions. In some countries, high costs and uneven quality of international connectivity, along with regulatory uncertainty, especially around data storage locations, have dampened demand. Coupled with relatively small population sizes and economies that are heavily concentrated, these factors have also constrained the supply of cloud services from international providers.
The Promise and Reality of Multi-Cloud
To alleviate some of the painful issues, organisations, particularly small and medium-sized businesses, moved to multi-cloud, working with multiple suppliers to give them cloud flexibility, additional security, cost benefits, and ‘lift and shift’ migration abilities.
There was an assumption that companies could move all their workloads from the data center to the cloud “as-is,” and later rework them to take advantage of cloud benefits. But in reality, that follow-up step rarely happened.
Now, many enterprises are stuck with old on‑premise systems that were simply lifted into the cloud, sitting alongside newer, cloud‑first applications built by a younger generation of engineers.
In addition to that, data is spread across different silos, from big clusters in the old data centers to vendor‑locked CRM platforms holding customer information.
Choice and Sovereignty
At the heart of the issue lies something more fundamental than market share. It’s the erosion of choice. And with it, the erosion of sovereignty.
When choice disappears, so does control. Increasingly, businesses and governments are finding themselves locked into external platforms that shape the rules. Instead of driving innovation, companies end up just trying to keep up with someone else’s system. While regulators struggle to deal with black-box platforms they can’t fully oversee.
What we’re left with looks uncomfortably familiar: a digital world dominated by a few influential players, whose priorities don’t necessarily match what’s best for the rest of us. The cloud then doesn’t just become a service anymore; it becomes a trap.
The Way Out
So what do we do about it?
We now need to deploy systems where data actually stays in the hands of the people it belongs to, i.e., businesses, governments, and citizens. Which means building with flexibility, open standards, and open source, so customers stay in control.
But as AI becomes a core part of how organizations function, and with so much depending on cloud infrastructure, it’s clear we need to take a bigger step.
That’s where Private AI comes in. It lets companies run AI securely on their own terms, whether that’s in the public cloud, a private data center, a sovereign setup, an isolated bunker, or even on the factory floor with machines streaming sensor data.
The key is that the data remains safe, and control remains with the user.
National Security, Innovation, and Ethics
This matters far beyond the narrow concerns of infrastructure teams.
It matters for national security. Countries across the GCC are realizing that depending on foreign hyperscalers for AI model training and deployment is a sovereignty risk. What happens when geopolitical winds shift? What if a government’s access to its own citizen data is throttled because of foreign legal obligations?
It matters for innovation. Startups can’t afford egress fees that penalize exploration. Universities shouldn’t have to tailor their research to the commercial constraints of a single vendor. If AI is the steam engine of the 21st century, then we must ensure that its fuel, i.e., data and compute, is accessible and portable.
And it matters for ethics. One of the biggest fallacies in AI governance is the idea that ethics can be layered on top of closed systems. That we can regulate model bias, algorithmic fairness, and explainability, all while deploying the most sensitive models through third-party APIs running in black-box environments. It’s like trying to certify food safety by inspecting the menu.
True AI governance begins with infrastructure governance. That means knowing where models are hosted. It means logging and auditing every inference. It means ensuring models don’t cross borders or breach local data residency rules without consent. And all of that becomes impossible when you’re locked into a single provider or rely on pure managed services that act like black boxes.
The Future of Cloud: Choice, Not Destination
Today, enterprises face a stark choice. Continue down the path of increasing dependence, accepting the rising costs, and reduced autonomy that comes with it. Or step into a world where infrastructure is fluid, where AI can be sovereign, and where cloud is a capability owned, not rented.
It is not enough to simply choose a data centre provider and procure hardware, or negotiate an enterprise agreement with a Cloud provider. They need to be the stewards of absolute data control within the organization. From a budget and compliance perspective.
Regulation alone won’t fix it. We need architecture. We need code. We need defaults that empower, not entrap. That architecture now exists. And for the first time in a decade, there is a real alternative. Unbundling an empire and bringing cloud simplicity to every CPU and GPU on earth.
Because the right to compute – like the right to speak, to think, and to govern – must remain with the people who own the data.
