“Cloud repatriation” has become the boardroom phrase of 2026. After a decade of cloud-first, cloud-everything, a growing number of South African organisations are quietly asking whether they moved too much, too fast and whether some of it should come home.
It’s a fair question. It’s also, in most cases, the wrong one.
The hangover finance teams are feeling right now is real. But the cause usually isn’t the cloud itself. It’s that most organisations migrated far faster than they learned to measure. We lifted, we shifted, we celebrated the data centre we decommissioned. Then the invoice arrived, and nobody in the room could fully explain it. That’s not a cloud problem. That’s a visibility problem wearing a very expensive disguise.
The numbers behind the headache
This isn’t a local anecdote. In Flexera’s 2026 State of the Cloud report, managing cloud spend was named a top challenge by 85% of organisations, the third year running it has held that spot. More tellingly, estimated wasted cloud spend climbed back to 29% of infrastructure and platform spend, reversing five years of improvement. The biggest culprits weren’t exotic: idle compute and over-provisioned instances, resources switched on, sized generously “to be safe” and then forgotten.
Sit with that figure. Nearly a third of cloud spend, on average, delivers nothing. The reversal is being driven by the very thing everyone is rushing into: AI and a growing sprawl of services, each with its own pricing logic, layered onto environments that were never fully mapped to begin with.
Here’s the part the global reports gloss over: in South Africa, that waste is denominated in dollars.
The rand makes a blind spot more expensive
Hyperscaler billing is overwhelmingly a USD affair. When the rand weakens, your cloud bill grows in real terms without a single additional workload being deployed. You can be running a leaner environment than last quarter and still present a larger number to your CFO purely because of exchange rates.
That’s a difficult commercial position because it muddies the one question that matters: are we becoming more efficient, or is the currency simply making us look worse? If you can’t separate those effects, you can’t answer the board. And when you can’t answer the board, every cost conversation defaults to the blunt instrument: a freeze, a culling exercise, or a panicked debate about moving workloads back on-premises.
Repatriation then gets sold as the cure. Sometimes it genuinely is the right call. But moving on-premises to escape a bill you couldn’t read simply relocates the blind spot, now with capex, hardware refresh cycles and load-shedding attached. Hybrid is already the reality for roughly three-quarters of organisations; the answer is rarely “all-in” or “all-out.” It’s “all-seen”.
What actually fixes it
You don’t optimise your way out of this with a spreadsheet and good intentions. You fix it with visibility: the discipline of being able to see, attribute and measure what your environment is doing.
1. Attribute before you optimise.
You cannot cut what you cannot trace. Before anyone debates repatriation, every meaningful chunk of spend should map to a team, service or workload. Tagging and cost allocation sound like housekeeping; they’re actually the difference between a decision and a guess. Most organisations discover their worst waste the moment they’re forced to put a name next to it.
2. Measure unit cost, not the rand total.
The headline figure lies in a volatile-currency economy. Track cost per transaction, per customer or per workload instead. Unit economics strip out the forex noise and reveal whether the thing you’re running is becoming more efficient or quietly bloating. It’s also the language a CFO is most likely to trust.
3. Instrument the hybrid seam.
The most expensive waste often hides in the gap between on-premises and cloud: the handoffs where on-premises tooling stops looking and cloud tooling hasn’t started. Neither side sees the whole path, so nobody owns the inefficiency. Unified visibility across both environments is where the biggest savings, and the biggest performance problems, usually surface.
The point
The organisations getting cloud right in 2026 aren’t the ones with the strongest opinion on cloud versus on-premises. They’re the ones that can see clearly enough to make the decision on evidence rather than instinct, and prove the result afterwards.
Splitpoint was built on a fairly impolite idea: most IT cost problems are visibility problems that have been allowed to grow up. You can’t repatriate, rightsize or defend a cloud decision you can’t measure. So before you spend another rand reacting to a bill you can’t fully explain, make it explainable.
The cure for the hangover isn’t drinking less. It’s finally turning on the lights.
Splitpoint Solutions builds observability into network, infrastructure and application environments across South African enterprise, so the people making the spending decisions can actually see what they’re deciding on.
Share
