IT Leaders Wrestle with AI, Sovereignty, Supply Chain and Deployment Flexing Conundrums

  • Martin Veitch, at Freelance Journalist

  • 13.10.2026 12:30 pm
  • #ArtificialIntelligence

Concerns over AI best practices, digital sovereignty compliance, access to hardware and need for operational flexibility are ubiquitous among IT leaders today, and they were certainly front and centre at last week’s Nutanix .NEXT on Tour London 2026 event.

In some ways, Nutanix is emblematic of these challenges-cum-opportunities. Having begun life as a pioneer in hyperconvergence (the concertina-ing of compute, storage, networking and data), the Silicon Valley company is now at the vanguard of the unified multi-cloud control plane market. It offers a sort of crossroads of options for companies seeking flexibility in where and how they run workloads. That means providing support and a familiar front-end over some variable and changing combination of public cloud, on-prem, service providers and containers across geographies. 

The following is an edited set of highlights from a session for media at .Next, hosted by Nutanix chief revenue officer Andrew Brinded and featuring two guests from highly regulated sectors. Nick Ovenden is CTO of UK & Ireland white-label insurance provider Markerstudy Group, and Edward O’Connor is CTO of Continent 8 Technologies, a service provider and datacentre operator that works with iGaming and online sports betting companies globally…

Compliance comes first

On rules, regulations and red tape, Markerstudy’s Ovenden said: “We have a substantial part of the market, so regulators are very interested in what we are doing.”

That sense of being closely watched also informs Markerstudy’s attitude towards agentic AI:

“We’re very much at the beginning. There’s very little touching live production data [and we are taking] a ‘softly, slowly’ approach. We have robust cybersecurity tooling but it’s the unknowns that tend to come for you. Nothing is ever 100 per cent secure.”

Markerstudy maintains an internal AI and innovation team and pursues business cases where there is strong return-on-investment potential. For example, Markerstudy uses Google’s CCAI contact centre AI platform to serve its customers. 

Ovenden said that the company is at pains to apply AI protections, however, and have “correct guardrails” to avoid errors.

O’Connor said that Continent 8 can help customers stay compliant by sharing the controls built into its own agile and compliant infrastructure: “We have strong ties into regulators, so wagers are recorded within the [specific US] state, for example,” he said.

Continent 8 Technologies and FuturePath AI recently launched Future8, an AI-native operating company built for regulated operators, formed as a joint venture between the two companies.

Continent 8 uses Nutanix across customers, many of which are modern, software-oriented and cloud-native, to offer flexibility in where and how workloads are managed. Deployment environments can be swapped out, depending on local rules or changing business needs. 

Continent 8 can deploy Nutanix clusters “almost like isolated islands …. [where] data will remain local and secure”, O’Connor said.

AI opportunities and controls

As organisations seek to benefit from agentic AI power without incurring undue risks, Continent 8 is an early adopter of Nutanix Agentic Gateway. Agentic Gateway supports the ability to burst out to frontier models while maintaining policies and easily turn instances off and on or adapt parameters.

O’Connor compared the Gateway to the way proxy servers became popular to control network traffic in the early days of the Web. “It’s the same principle 25 years later,” he added.

“For us, AI is less about replacing mundane BAU [business as usual] tasks and more about increasing capacity and scale and expanding our offers.” 

More broadly, O’Connor believes that partnering is a smart way for companies to move fast and control costs.

“Outsourcing to us is a much easier, faster way to be more efficient,” he said. “Buying your own infrastructure, standing it up and running it is inefficient. Our customers say they’re not infrastructure companies… that’s our job as a service provider platform.”

Nutanix’s Brinded agreed, noting that the company can invest in GPUs and apply large teams of software engineers across open weights and frontier models across over 20,000 customer datacentres. That knowledge and scale gets shared with partners and customers.

An example comes with digital sovereignty. Users can control resources using kill switches, role-based access control and chargeback mechanisms to avoid the risk of token overspend and Shadow AI, Brinded argued.

Ovenden said the company needed to maintain a “balancing act” on AI, where staff can innovate using AI tools but not have ungoverned access to vast resources. 

Freedom of movement

Nutanix’s Brinded said customers required freedom to move and change plans.

“In a world of uncertainty, the more optionality you have the better it is for digital sovereignty,” he added. “We can provide an airgapped environment with role-based access control for every person and if there are geopolitical events, customers can quickly move those environments off to another cloud or on premise.”

It’s not just geopolitics. Demand for AI workloads and shortages of GPUs and high-performance memory are forcing enterprises to box clever and flex capacity.

Ovenden said Markerstudy is seeing “huge constraints” on hyperscaler resource availability and “rationing what is available… [There are] capacity issues [with] too many people asking for the same things.”

Brinded said Nutanix expects such supply-chain challenges to persist for at least another 12 months, so the ability to have portable cross-platform licensing and management tools is more attractive than ever. That flexibility is also a benefit as companies continue to assess their options almost four years after Broadcom’s acquisition of VMware was followed by major licensing changes.

Last but not least, O’Connor also won favour with assembled journalists when he exercised caution on generative AI being widely used for content creation. There was a potential “backlash against AI content”, he said, suggesting that using AI for ideation “is wrong” but AI can better be used to refine draft ideas, he argued.  

“The individual’s voice is important,” he added. 

Related Blogs

If AI Does the Books, Who Takes the Blame?
  • 4 months 2 weeks ago 06:00 am
Where AI Offers Real Value
  • 6 years 8 months ago 08:00 am

Other Blogs