10
Loading video...
By Amir Daftari
News Reporter
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
As artificial intelligence becomes more deeply embedded in business operations, companies are facing growing pressure to rethink how they deploy the technology and who controls the systems on which they increasingly rely.
McKinsey’s August 2026 State of AI survey of 1,719 respondents across 97 countries found that 60 percent expected their organizations to increase AI investment over the following year.
As businesses commit more resources to AI, questions around data security, ownership, infrastructure and dependence on external technology providers are becoming increasingly important.
Read More on Business
These concerns are bringing AI sovereignty into sharper focus.
...
Why AI Sovereignty Matters
AI sovereignty refers to an organization's ability to control its AI infrastructure, data, models and governance policies, rather than relying entirely on external providers.
This can involve hosting AI systems on company premises or in private and sovereign cloud environments, allowing businesses to retain greater oversight of how their technology operates and how sensitive information is handled.
One of the main drivers behind this approach is the risk of exposing confidential business information through public AI tools. Employees may inadvertently share customer proposals, internal reports or proprietary technology without understanding how that data is processed or stored.
IBM's Cost of a Data Breach Report 2026 found that one in four malicious breaches were AI-enabled, while more than 20 percent of organizations reported breaches targeting AI models or applications.
The global average cost of a data breach reached $4.99 million.
Private AI Infrastructure
For some businesses, addressing these risks means moving away from public AI tools toward systems that operate within their own secure environments. This can give organizations greater control over where their data is stored, who can access it and which AI models they use.
Andre Palaguine, CEO of Language Intelligence Corporation, argues that companies need to rethink their approach to AI as the technology becomes a fundamental part of their operations.
"Sovereignty starts with security, but it also requires control and governance. If I control the AI, if I can govern it, and if it operates in a secure environment, then I know that it belongs to the organization," Palaguine told Newsweek.
His company develops Fluent, a multilingual AI platform designed to operate on-premises or within private and sovereign cloud environments. The platform separates its application layer from individual large language models (LLMs), allowing organizations to work with different underlying models without rebuilding their entire technology infrastructure.
Palaguine says this approach also addresses the risk of exposing sensitive information through public AI systems.
"Today, any data can be confidential. It can involve employees, customers, vendors, proprietary technology, internal workflows, or a customer’s confidential information," he said. "If someone takes an RFP (Request for Proposal) containing sensitive information and puts it into a public AI system simply to understand it, they have introduced a risk that the organization may never have intended to create."
Greater Control
The shift toward greater control over AI infrastructure does not necessarily mean businesses are abandoning commercial providers.
Paul Henninger, Group Head of Consulting at KPMG UK and Switzerland, said organizations are becoming more deliberate about where their models, data and intellectual property are hosted.
"Rather than a wholesale shift away from commercial providers, the focus is increasingly on giving businesses greater control over where their models, data and intellectual property are hosted and governed," Henninger told Newsweek.
The financial implications of greater AI adoption are also forcing businesses to look beyond the price of individual tools and consider what it takes to make the technology work at scale.
"The technology itself is only part of the cost of scaling AI. Businesses can underestimate the investment needed around it, particularly in data, cybersecurity, governance, skills and the organizational changes required to embed AI effectively," Henninger said.
Henninger added that security and sovereignty are increasingly becoming purchasing and investment criteria, with "trust, control and resilience" becoming integral to business decisions rather than an afterthought.
Ownership Does Not Guarantee Control
But controlling where AI operates is not the same as understanding the technology itself.
Speaking to Newsweek, Fazl Barez, a senior researcher at the University of Oxford specializing in AI governance, said businesses should not assume that hosting AI privately gives them greater insight into how the underlying models work.
"We have seen enormous progress in what AI systems can do, but much less progress in understanding how they work and why they behave the way they do. You can have the model running entirely on your own servers and still have very little understanding of how it works or what it might do," Barez said.
Nor does moving AI into a private environment eliminate the risks of mistakes, manipulation or unexpected behavior. Barez questioned whether some providers were presenting local hosting as a more complete solution than it really is.
"I would also be skeptical of companies selling sovereignty as simply buying their infrastructure or hosting a model locally. That doesn't necessarily give you meaningful control over the technology. There is little value in having complete ownership of the infrastructure if you still cannot understand, evaluate or control the system running on it."
Ultimately, sovereignty cannot be measured simply by where a model is hosted or who owns the servers. Businesses also need the technical capabilities to scrutinize the technology, manage its risks and intervene when necessary.
Otherwise, they may have greater control over their infrastructure without gaining meaningful control over the AI itself.
Contact Newsweek editor on this story: Mike Hirst.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines