ThinkOn: Sovereign cloud for the age of AI — built to last

Summary

ThinkOn built its platform on NetApp^® StorageGRID^® object storage software to give Canadian organizations a sovereignty-first alternative in a market dominated by foreign clouds. The payoff: a durable infrastructure business, a partnership where one plus one equals three, and 20 million Health Canada records ready for the age of AI.

There is a sovereignty crisis

Why does data sovereignty matter?

For years, the cloud made it easy to store data. AI is now making that data dramatically more valuable — and exposing a problem many organizations have spent years avoiding.

Where is all that data? Who controls it? Can it be trusted? Can it be moved where it needs to go without creating enormous costs or locking the organization into a single provider?

For Craig McLellan, CEO of ThinkOn in Canada, these aren’t theoretical questions. They are becoming some of the most important technology and business decisions organizations face.

Three companies control 85 percent of Canada's public cloud market — and none of them are Canadian.

“Every country in the world, in order to be the best country it can possibly be, has to have control over its data destiny,” McLellan says. “You can’t outsource the core without becoming a vassal state to the organizations you’ve outsourced it to.”

It’s a bold claim, and McLellan doesn’t back away from it. But it’s also, increasingly, a business case. As AI reshapes how organizations value, store, and move data, the question of who actually controls that data — and whether it can be trusted — has stopped being a compliance footnote and started being a boardroom conversation.

Be sovereign on a proven foundation

What makes StorageGRID the foundation for a sovereign cloud?

ThinkOn launched in 2013 on a simple premise: The world didn’t need another sales organization competing for customers. It needed reliable infrastructure delivered through partners who already had the relationships. That channel-first philosophy, paired with a second conviction — that data behaves differently than compute — has shaped every decision since.

“Data grows continuously,” McLellan said. That steady, compounding growth creates a predictable revenue stream that channel partners value, but only if the underlying platform is dependable. So rather than build proprietary technology to shave costs, ThinkOn has consistently chosen mature, proven components.

“If I can squeeze 10 percent out of something by using an open-source tool or my own software, that just doesn’t make sense,” he says. “We have to be very careful.”

That discipline is why ThinkOn built its object storage platform on NetApp’s StorageGRID — a decision McLellan describes less as a technology choice than a values match.

“There is nothing more important to me than durability and resilience,” he says. “StorageGRID does that.”

StorageGRID’s track record — more than 15 years in production — gave his team the confidence to build on it at scale.

The AI data rush

How is AI changing the value of stored data?

What’s changed the calculus for ThinkOn’s customers isn’t just sovereignty; it’s AI’s sudden, insatiable appetite for data. McLellan points to an inflection point his team didn’t fully anticipate: It became more cost-effective to retain data than to destroy it, and AI has stretched that retention horizon even further.

McLellan points to a ThinkOn project involving the digitization of 20 million applications for the distribution of pharmaceuticals and medical devices through Health Canada. The value isn’t simply having those records in digital form. It’s the possibility of discovering insights that were previously buried inside unstructured information.

“What I get excited about is all this unstructured data that’s coming online,” McLellan says. “Once we have it online, I think some of the models are going to find really amazing correlations.”

In other words, the next AI breakthrough may not come from creating more data. It may come from finally making sense of the data organizations already have.

Across hundreds of customer conversations, McLellan says the same three concerns surface.

First: “Oh, where the hell have I put this?” Or, the basic struggle to organize and curate data without limiting its future usefulness.

Second, cost predictability: “There's no quicker way to blow my business plan than to have unexpected costs.”

And third, supply chain integrity: the fear that data will become inaccessible, or worse, accessible to people who shouldn’t have it.

Trust, then truth

Why does trusted data start with cyber resilience?

Ask McLellan what ties sovereignty, AI readiness, and security together, and he returns to two words: trust and truth.

“If you have trust, then you can have truth,” he says. “And if you’ve got truth, you become the source for lots of things the customer may want to do in the future.”

It’s a philosophy that shows up operationally, not just rhetorically. ThinkOn began hardening its security posture in earnest in 2023, engaging Security Scorecard for continuous external monitoring and building an internal vulnerability scanner that runs across both its control plane and access plane.

“There is no level of diligence that, in my opinion, is too much,” he said. The guiding design principle is one he’s repeated for years: “If we don’t design for things to fail, then when we fail we’ll be redesigning whatever it is we did. So let’s design it to fail.”

That same care extends to sustainability. ThinkOn has signed on to Canada’s Net Zero Challenge, targeting full carbon neutrality by 2050, and now shares a simple efficiency metric — terabytes per watt — with clients evaluating their environmental footprint.

The future: A platform that scales

Perhaps ThinkOn’s sharpest insight is about flexibility itself. Customers today run hybrid environments — on-premises, private cloud, and hyperscale cloud simultaneously — and McLellan doesn't expect that to change.

“I don’t think there’ll be a full-court press to one of the three,” he says. What matters is that data, wherever it lives, “needs to be in a place where it can be managed and secured and leveraged.” Combined with NetApp’s channel reach and product maturity, that positions ThinkOn to sell services, sell product, or blend both — whatever a customer’s NVIDIA GPU clusters or neocloud strategy requires.

That flexibility is what makes the NetApp relationship durable rather than transactional.

“This is a scenario where one plus one equals three,” he says, “because no one has to feel a loss in the process of embracing it. And that is very unique.”

In a market increasingly defined by who can be trusted with the truth, ThinkOn is betting that trust, delivered consistently, is the only feature that scales.