Why developers are moving their stack to Europe — and what it means for Australia
For most of the last decade, the default answer to "where should I host this?" was a US hyperscaler. AWS us-east-1, a Vercel deployment, a Cloudflare account, maybe a sprinkle of Google Analytics — done. That muscle memory is breaking. A growing number of indie developers, agencies and even mid-size companies are quietly migrating their digital stacks to European providers: Hetzner in Germany, Scaleway in France, OVHcloud, Exoscale, Fastmail, Proton, Mailjet, and a long tail of smaller specialists.
The trigger is no longer just GDPR paperwork. It's a deeper recalculation of risk — political, legal and financial. And while the loudest voices in this shift are European, the implications are very relevant to Australian developers, who sit inside the Five Eyes intelligence-sharing arrangement and increasingly need to convince European customers their data won't leak into US jurisdiction.
The sovereignty argument grows up
The framing has shifted from "privacy" to "sovereignty". A recent essay in The Parliament Magazine bluntly described the EU as having become a "digital colony" — overwhelmingly dependent on cloud, search, social and operating systems run from California and Seattle. In The Guardian, Johnny Ryan made the same case more sharply: Europeans are "dangerously reliant on US tech", and the current geopolitical mood — tariffs, unpredictable executive orders, the CLOUD Act — makes that reliance look more like a vulnerability than a convenience.
In an interview with German site Netzpolitik.org, a developer pushing for European alternatives put the principle even more plainly: "Any dependency is a bad thing." That sentence is doing a lot of work. It captures something engineers already intuit about distributed systems — single points of failure are bad — and applies it to entire vendor relationships, jurisdictions and supply chains.
What "moving the stack" actually looks like
The migration isn't usually a single dramatic re-platforming. It's a gradual layer-by-layer replacement:
- Compute and storage: US hyperscalers swapped for Hetzner, OVHcloud, Scaleway or Exoscale. The economics are striking — a dedicated server in Falkenstein can cost less than a comparable EC2 instance by an order of magnitude.
- Email and productivity: Google Workspace and Microsoft 365 swapped for Proton, Tutanota, Fastmail, Mailbox.org or Infomaniak.
- Analytics: Google Analytics swapped for Plausible (Estonia), Fathom or Matomo self-hosted in the EU.
- CDN and edge: Cloudflare partially replaced by Bunny.net (Slovenia) or Fastly configurations restricted to European POPs.
- Payments: Stripe supplemented or replaced with Mollie or Adyen for European customers.
- AI and ML: OpenAI APIs swapped for Mistral, Aleph Alpha or self-hosted open-weights models on EU GPUs.
The work is unglamorous — DNS records, SPF/DKIM rewrites, data export scripts, re-issuing API keys — but the trend is real enough that European hosting companies are openly courting the traffic. Hetzner's order queues for new servers have become a regular discussion topic on Hacker News.
The regulatory pressure is real, not theoretical
For a long time, GDPR enforcement felt like a paper tiger to small developers. That has changed. After the Schrems II ruling and a string of decisions against Google Analytics in countries like Austria, France and Italy, lawyers stopped giving comfortable advice about "standard contractual clauses" being enough on their own. The Data Privacy Framework that succeeded Privacy Shield is widely expected to be challenged again.
At the same time, Europe is actively funding alternatives. The European Commission has just committed €307 million through its Digital Europe Programme to accelerate AI and digital infrastructure leadership — funding AI factories, cybersecurity, digital skills and deployment of European cloud capacity. That's not enough to outspend Microsoft on its own, but it signals where procurement money will flow. Government contracts, health systems and universities across the bloc are increasingly being told to prefer EU-based providers where they exist.
Why this matters from Sydney or Melbourne
Australia isn't in the EU, but Australian developers and SaaS founders are directly affected.
If you sell to European customers, you're now being asked harder questions in security questionnaires: where is the data stored, who has access, is there sub-processing through the US, can you guarantee no transfer outside the EEA? An Australian startup that hosts everything in Sydney AWS is technically processing data through a US-headquartered provider, which still triggers Schrems-style anxiety for some buyers. Picking a European region on AWS or Azure helps a little; using a fully European provider helps a lot more.
If you sell to Australian customers, the same logic is starting to apply in reverse. The federal government's hosting certification framework, the push for sovereign cloud through AWS's and Microsoft's local "top secret" regions, and the steady drumbeat of breach notifications (Optus, Medibank, Latitude) have made data location a board-level question. The European playbook — diversification, jurisdictional clarity, contractual transparency — translates well.
If you're a solo developer or small agency, the cost story is genuinely attractive. The Australian dollar's relationship with the US dollar makes AWS bills painful. Hetzner pricing in euros, paradoxically, often works out cheaper for raw compute than the equivalent AWS Sydney instance, even after factoring in latency and egress.
The honest trade-offs
It would be dishonest to pretend the move is free.
- Latency: Serving Australian users from Frankfurt adds 250–300ms of round-trip time. For a customer-facing app aimed at Australians, that's a deal-breaker without a CDN layer. For internal tools, backups, batch processing or APIs consumed by European customers, it's a non-issue.
- Ecosystem depth: AWS has a managed service for almost everything. European providers are catching up but you'll often roll your own — managed Postgres on Scaleway is fine, but the equivalent of DynamoDB, SageMaker or Bedrock is patchier.
- Tooling assumptions: A lot of "infrastructure as code" templates, tutorials and Stack Overflow answers assume AWS. Hetzner Cloud has a Terraform provider, but you'll do more reading.
- Support culture: European providers tend to be more email-and-ticket, less account-manager-on-Slack. For some teams that's a feature.
The realistic pattern most teams settle on is hybrid: keep latency-sensitive workloads close to users, move data-at-rest and back-office systems to Europe, and consciously reduce single-vendor dependency. The goal isn't ideological purity — it's optionality.
The bigger principle
The deepest insight from this wave of migrations is the one the Netzpolitik interviewee landed on: dependency itself is the risk. Whether the lock-in is to a US hyperscaler, a Chinese CDN or, eventually, a dominant European provider, the answer is the same — keep your data portable, your contracts readable, and your DNS in your own hands.
For Australian developers watching the European stack-shift from a distance, the lesson isn't necessarily "move everything to Hetzner tomorrow." It's that the default of putting every byte you own into one US account, with one credit card, in one jurisdiction, deserves to be re-examined. The companies that quietly diversified five years ago are the ones now answering procurement questionnaires with a shrug instead of a sweat.