Kyndryl Canada head on sovereign strategy in post-Mythos era

Canadian organisations are speeding up cloud and network modernisation projects in direct response to security concerns raised by frontier AI models, according to Brian Medeiros, President of Kyndryl Canada.

The business leader said that clients who had previously set multi-year modernisation timelines are now asking to move considerably faster, citing the emergence of frontier models such as Mythos as a catalyst. 

Medeiros said the most immediate impact is on patching and vulnerability management, as organisations try to close gaps before they can be exploited.

"Patching in itself is a challenge. It takes a lot of moving parts within an organisation to be able to address the changes and the downtime that may be required," said Medeiros. "We've seen an acceleration in clients accelerating their SD-WAN transformations. For example, we work with Palo Alto [Networks], where we embed security right into those network transformations. That acceleration is coming for this exact reason."

He added, as organisations scale AI, they're placing greater emphasis on modernisation and governance while shifting from experimentation to execution. Public sector clients tend to place greater weight on keeping data within a province or the country's borders, partly reflecting restrictions on where staff supporting that data can be located, while private sector clients have historically shown more appetite for storing data outside Canada. 

With Canadian clients in both the public and private sectors, the questions still focus on: 'Who controls my data and AI models? How are they governed? How do I reduce risk while maintaining flexibility and access to innovation?'

Medeiros said other clients are speeding up migration of applications away from legacy, on-premises data centres to hyperscale cloud providers, believing that doing so sooner reduces the risk that frontier models will expose weaknesses in their existing estate. Additional protection comes from technical debt repair, such as patching, identity, segmentation and resilience.

This urgency is compounding a problem Kyndryl had already identified in its most recent annual Readiness Report, published in October 2025: that ageing technology foundations are holding Canadian organisations back. The report found that 67 per cent of Canadian leaders surveyed believed that foundational technology issues were an impediment to innovation. On the global stage, that figure was 57 per cent.

"Canada, relative to other markets, has been a bit lagging in that area," said Medeiros. "They've fallen behind on the technical aspects, and that  concern was raised by leadership as they're trying to adopt AI and scale AI, on the fact that that's going to hold them back a little bit."

Medeiros said data sovereignty has become an increasingly prominent concern for Canadian clients since the release of the federal government's national AI strategy, but that the conversation has moved on from a preference for Canadian-based data centres that was more prominent a year ago. He said organisations increasingly focus on how data is managed and protected, rather than solely on where it physically sits.

Barriers to unlocking AI

Medeiros added that business leaders should establish what outcome the organisation is trying to reach in the first place, and design the use of AI around that, which requires people who understand the business beyond their own immediate area.

From Medeiros, organisations that combine modernisation of their technology estate with AI adoption, rather than treating the two as sequential, tend to pull ahead, as do those that give staff space to build familiarity with the technology directly, rather than relying on top-down instruction.

"AI isn't something where you can just read a book and understand the technology and how it applies to you. It's very hands-on, and organisations that are building that culture are getting ahead."