Public cloud has genuine strengths — but for a large class of workloads in 2026, the case for it has weakened on three fronts at once. The economics often don't stack up. The data sovereignty posture has shifted as US executive policy has become less predictable. And the FX exposure of USD-denominated hyperscaler bills against a volatile AUD is now a board-level conversation in its own right. Repatriation to a Proxmox-based private cloud addresses all three.
We don't recommend repatriation for everything — only where it demonstrably improves your position. If the numbers and the risk model don't support a move, we'll tell you that.
When repatriation makes sense
Steady-state workloads
Workloads with predictable, consistent resource usage get little benefit from cloud elasticity — and pay a premium for it anyway.
Data-heavy services
Databases, media processing, and analytics workloads that push large volumes of data out of the cloud accrue egress fees that compound quickly.
Sovereign & regulated workloads
Healthcare, finance, government, and critical infrastructure operators that require AU data residency, auditable controls, and Australian-owned operating infrastructure.
VDI, dev/test & internal tooling
Virtual desktop infrastructure and internal tools rarely need the elasticity of public cloud — but they run 24×7 and accumulate significant costs.
Geopolitical risk exposure
Workloads where a foreign-state subpoena, sanctions regime, or executive order against a vendor would materially impact your operations or customers.
FX-sensitive cost base
Organisations with AUD revenue but USD-denominated hyperscaler bills — where a 10–15% AUD depreciation translates directly to a budget blow-out you can't hedge cheaply.
Why repatriation matters more in 2026
Sovereignty: who can reach your data, and under whose law
The two big hyperscalers operating in Australia (AWS, Azure) and the third with growing local presence (Google) are all US-headquartered. Their AU regions host your data physically, but the legal control surface still includes the United States. The US CLOUD Act obliges US-incorporated providers to produce stored data in response to lawful US process, irrespective of where it physically sits. FISA Section 702 and Executive Order 12333 sit alongside it.
For most workloads this is theoretical. For some — Defence-adjacent suppliers, critical infrastructure operators under SOCI, IRAP-assessed workloads, healthcare and legal practices holding sensitive client data, organisations with active or potential M&A activity — it isn't. Repatriating to a Proxmox cluster on hardware you own, in a colo you contract with directly, operated by an Australian provider, puts the entire legal control surface inside Australian jurisdiction.
The Broadcom era at VMware compounds this point. The platform vendor for the dominant private cloud hypervisor is now a US conglomerate that has demonstrated willingness to change commercial terms mid-cycle. The argument for picking a platform from a vendor whose business model isn't a single executive decision away from changing applies as much to political risk as to pricing risk.
Currency exposure: hyperscaler bills in USD, your revenue in AUD
Hyperscaler billing for Australian customers is denominated in USD (with AUD invoicing applied at the prevailing rate). Over the last three years, AUD/USD has moved 12–18% inside any rolling twelve-month window. That isn't noise — that's a budget blow-out hiding inside what looks like a stable infrastructure line item.
Currency hedging instruments exist, but they're typically:
- Imperfect — they cover a forecast bill, not the actual cloud consumption you'll have
- Costly — the spread eats most of the protection at the volumes most businesses run
- Untimely — they're typically set up by finance teams after FX already hurt, not before
A repatriated platform converts an ongoing FX-exposed opex stream into a one-time AUD CAPEX (hardware purchase) plus a predictable AUD opex line (Proxmox subscriptions billed locally, Australian engineering support, colo at AUD rates). The FX risk doesn't disappear, but it moves from "uncapped, unpredictable" to "bounded, hedgeable if you want to".
For organisations with a multi-year horizon, the question isn't whether AUD weakness is a real risk to next year's infrastructure budget — it's whether you want that risk priced into your architecture or quietly compounding inside it.
How we approach it
TCO, sovereignty & FX risk assessment
We model the real cost of your current cloud footprint against a private cloud equivalent — including compute, storage, networking, licensing, operational overhead, and the AUD/USD exposure baked into your current bills. The same assessment maps your workloads to their sovereignty profile so the recommendation respects both economic and regulatory constraints.
Platform design & build
We design a Proxmox VE environment sized for your actual demand profile, with Ceph or ZFS storage, appropriate HA, and Proxmox Backup Server integration. Built in an Australian colo (or your own DC), under Australian operating control, sized to absorb your workloads on day one.
Migration with clear cutover windows
Workloads are migrated in planned batches with defined cutover windows and rollback paths. Sensitive workloads cut over first or last depending on risk profile. We don't cut over until the platform is validated and your team is comfortable.
Typical TCO outcome
Organisations repatriating steady-state workloads typically see 40–65% reduction in annual infrastructure spend compared to their public cloud equivalent — with full ownership of the platform at the end of Year 1, AUD-denominated ongoing costs, and the entire legal control surface inside Australia.