Hey friends,
I keep getting pulled into the same conversation with leadership teams. They want to talk stack. Copilot or Claude. Build our own or buy. Azure or AWS. Open source or proprietary.
Every time, I push the conversation back one step. Because none of those debates make sense until you answer a much simpler question first.
Where do you want your data to live?
Most teams have never sat down and answered it properly. They have a feeling, usually some version of "we'd prefer it stayed in Australia," but they have never tested whether that feeling matches their actual reality. So every AI decision after that gets made on top of a foundation nobody has checked.
There Is No Right Answer Here
Before I lay out the three positions, I want to be clear about something. None of them is the correct one. This is not a values judgement. It is a question of fit.
What you do, who you serve, what obligations you sit under, what your IP is worth, what your customers expect. Those things determine where you land. A health-tech founder and a marketing agency should not be making the same call here. And the moment a leadership team accepts that, the rest of the AI conversation gets a lot easier.
Position One: You Have No Residency Requirement
The first position is the simplest. Your business does not have a regulatory obligation to keep data in Australia. You do not hold the kind of IP that demands it. Your customers do not expect it.
If that is you, own it. Say it out loud. "We are okay with our data living overseas, provided the vendor handles privacy and security properly."
That single sentence opens up your horizons. The vast majority of the best AI tooling is built in the US. Your stack choices multiply. Your speed of adoption increases. You still have to do the diligence on how each vendor stores, protects, and governs your data, but the decision tree is much shorter.
A lot of businesses sit in this category and don't realise it. They have absorbed a vague nervousness about data leaving the country that doesn't actually apply to them. If that's you, name it and move on.
Position Two: Data Must Stay In Australia, But Cloud Is Fine
The second position is that residency matters. You need your data to stay onshore. But you are comfortable with it living on a cloud provider, as long as that cloud has Australian data centres and contractual guarantees.
This is where most regulated and government-adjacent businesses land. Healthcare, finance, parts of education, anyone touching sensitive consumer information at scale.
Here is the part most leadership teams skip. If you say residency is a hard requirement for AI, you have to look at your existing tech stack and check whether it is already true.
Because if your CRM, your file storage, your collaboration tools, your support platform, your marketing automation are all already sending data overseas, then your residency position is wishful thinking, not a policy. You are not actually operating the way you think you are. Either fix the underlying stack so it matches the position you are claiming, or accept that the position is softer than you thought and stop pretending otherwise.
You cannot have a "data must stay in Australia" rule for AI and a "data can go wherever" reality for everything else. Pick one and live consistently.
Position Three: Data Must Stay In Australia, And Inside The Organisation
The third position is the strictest. Residency matters, and so does keeping the data inside your own walls. You do not want it on someone else's cloud at all.
Same trap as position two, just sharper. The moment you start using any mainstream SaaS tool, your data is already leaving your organisation. Your knowledge is sitting on someone else's infrastructure. That is not a future risk, that is current reality.
So this position is only honest if you already operate a private cloud or fully self-hosted environment where everything is bounded inside your organisation. If you do, fantastic, your AI choices flow naturally from that. You will be looking at private deployments, locally hosted models, your own infrastructure.
If you don't, then this position is not actually available to you yet. What you can do is decide whether you want to get there. That is a much bigger conversation, with cost and capability implications well beyond AI. But at least you are having the right conversation.
Where People Get Stuck
The mistake I see most often is leadership teams trying to choose between Copilot, Claude, an internal AI build, or a hybrid model, without having done this work first.
It is impossible to choose well from that position. Every option looks reasonable in isolation. Every vendor has a polished answer. Every consultant pushes the stack they know best.
But the answer to "which AI" is downstream of "where does our data live, and where do we need it to live." Until that is settled, you are picking between options that may or may not even be compatible with your actual constraints.
The Practical Move This Week
Get your leadership team in a room. Forty-five minutes is enough.
Answer two questions on a whiteboard. First, which of the three positions do we hold? Second, does our current tech stack actually reflect that position, or are we kidding ourselves?
That is it. You don't need a strategy document. You don't need a vendor. You need a shared answer. Once you have it, every AI decision from architecture to vendor selection becomes dramatically easier, because you finally have something to test those decisions against.
A Calm Takeaway
A lot of the noise around AI right now comes from skipping foundational questions and arguing about tools. Tools are downstream. Position is upstream.
Data sovereignty is one of those upstream questions. It is not glamorous. It will not show up in a vendor pitch. But it quietly governs every meaningful decision you will make about AI for the next three years.
Pick your position. Check it against reality. Then build on top of it.
See you next week,
— Aamir
📲 Resources & Links
🎧 Listen to the Podcast Episode on: Spotify | Apple Podcasts | YouTube
📘 Book: The CEO Who Mocked AI (Until It Made Him Millions) by Aamir Qutub