But how to get to that European cloud?

Original Article ↗ Hacker News Discussion ↗

Geopolitics, sovereignty, and who should build it

  • Some argue Europe (and possibly allies like Australia) needs its own cloud for legal and strategic independence, especially post‑CLOUD Act and as US defense guarantees look shakier.
  • Others say Australia should orient toward Asia or stay neutral, and that a truly “sovereign cloud” owned by a US firm (e.g. AWS “EU sovereign cloud”) is contradictory unless tech and control are fully localized.
  • A minority questions whether “sovereign cloud” is even coherent: renting “other people’s computers” is inherently non‑sovereign, so owning hardware and decentralizing might be safer.

Industrial policy, protectionism, and R&D

  • Proposals include tariffs on foreign clouds, “China-style” requirements for local partners owning local tech, and tying public spending to EU providers.
  • Counterpoint: tariffs are distortive and similar effects could come from subsidies or R&D support, but switching from incumbent US clouds still needs strong policy push.
  • Several want a “European DARPA” focused on long‑term, high‑risk tech, arguing current EU programs (Horizon, Gaia‑X) are too bureaucratic and consultant‑driven.
  • Disagreement over whether EU regulators and political culture (risk‑averse, lawyer‑led, fragmented financial markets) are capable of designing the right incentives.

Gaia‑X and previous “sovereign cloud” efforts

  • Gaia‑X is widely cited as a failure: design‑by‑committee, money to bureaucrats and large incumbents, little going to engineers.
  • Some describe it as covert subsidy for the usual big industrial players, with poor technical outcomes and even weak hiring signals.
  • Example from France: “sovereign” clouds that turned out to be Huawei‑run, later shut down, leaving customers repeatedly forced to migrate.

Existing European providers and product gaps

  • OVH, Hetzner, Scaleway, Infomaniak, Outscale and telecom‑run datacenters are mentioned as real, working European infrastructure.
  • Critique: many are essentially VPS/“lumber” sellers, lacking polished, self‑service, integrated services (autoscaling, managed DBs, S3‑like storage, CI/CD‑friendly patterns) that make hyperscalers attractive.
  • Others argue government workloads are often small and don’t need hyperscaler‑level scale; but reliability and operational tooling (rollouts, HA) still benefit from “cloudy” features.

Market dynamics, regulation, and free‑market debate

  • One camp says “if there’s demand, the market will provide”; government should cut regulation and taxes.
  • Another camp says cloud is a sticky, scale‑and‑network‑effects market where free entry fails; active regulation is needed to create demand for “subject only to EU law” providers.
  • Broader dispute over tariffs vs subsidies, and whether cloud is (or tends toward) oligopoly‑style concentration that justifies antitrust and industrial policy.

Talent, pay, and capital

  • Many see low European tech pay (e.g. ~€70k) as a core blocker: hard to “bend over backwards” to build massive infrastructure when US remote roles offer far more.
  • There’s debate over how big the EU–US gap really is once taxes, social benefits, and COL are included, but consensus that median EU tech salaries are far below US hyperscaler levels.
  • Some advocate aggressively poaching engineers who built AWS/Azure/GCP, combined with tax incentives, high‑speed rail connectivity (London–Paris–Randstad), and cross‑border startup funding.
  • Others emphasize Europe’s problem is risk‑averse capital: lots of pensions and industrial giants, very little true high‑risk VC.

Government procurement and practical experience

  • Several describe disastrous experiences with big “framework” hosting contracts for EU governments: slow, incompetent providers chosen on price, lots of red tape, poor K8s/HA designs, and attempts by dev teams to escape to AWS/Azure/Hetzner.
  • Lesson drawn: a European cloud for the public sector only works if it matches hyperscalers’ self‑service and developer experience.

Alternatives to a single Euro‑hyperscaler

  • Some think chasing an AWS clone is “fighting the last battle”; better to:
    • Standardize around simpler clouds (Hetzner‑like),
    • Encourage many smaller “drones” (local or sectoral providers) instead of one “aircraft carrier,”
    • Or push on‑prem / microcloud solutions for resilience against cyber or connectivity disruptions.
  • Cloud itself is questioned: for many workloads, traditional hosting with good ops may beat complex cloud constructs; cloud’s main value is operational convenience and managed services, not theoretical elastic scaling.

Soft power, ecosystems, and Microsoft/AWS dominance

  • US clouds benefit from entrenched ecosystems: universities teaching proprietary stacks, free credits for students, consulting firms tied to Microsoft, and “never fired for choosing X” brand equity.
  • Some suggest universities should stop teaching vendor‑specific stacks (e.g. ASP.NET/Azure) to weaken this lock‑in and create space for neutral or European alternatives.