Sponsored by
In the autumn of 2022, Ukrainian forces requested that SpaceX extend Starlink coverage over Sevastopol to guide a sea drone operation into the harbour.
Elon Musk refused. Coverage over Russian-occupied Crimea had not been active, partly due to US sanctions, and Musk declined to switch it on, citing concerns about SpaceX becoming directly complicit in a major act of war and the risk of nuclear escalation.
The operational fact was clear regardless of the reasoning: a private American company had made a battlefield decision that affected a sovereign military operation, unilaterally, without advance notice, and with no formal recourse for the party depending on the network.
European defence ministries read that decision carefully.
The dependency Starlink revealed was not new.
Commercial satellite infrastructure has always been owned by someone. What changed in 2022 was that the someone made a visible, consequential, unilateral decision in a live conflict, and the party depending on the infrastructure had no contractual or legal mechanism to prevent it.
That is a different kind of risk than service outage or pricing change. It is the risk of the infrastructure becoming a policy instrument in the hands of a third party whose interests may not align with yours at the moment you need it most.
European defence and space policy has spent the years since working through what that risk means in practice.
AI Executive Assistant for Normies
Catch is a smart, proactive AI admin agent that handles your scheduling, inbox, emails, follow-ups, and day-to-day admin across Gmail, Outlook, Slack, WhatsApp, and phone.
No setup. No learning curve.
From swamped to sorted, instantly.
The European answer is not a single satellite network.
It is a layered approach across three separate programmes, each addressing a different part of the dependency.
IRIS² is the most significant. The Infrastructure for Resilience, Interconnectivity and Security by Satellite is a European Union programme to deploy a sovereign broadband constellation by 2030. The consortium building it includes Airbus, Eutelsat, SES, Thales, and Telespazio. The design requirement is that European governments retain operational control over the network, including the ability to prioritise government and defence traffic independently of commercial allocations. That last clause is the direct response to the Starlink decision: the network cannot be switched off for government users by a commercial operator acting on their own judgment.
Galileo already operates as the European alternative to GPS, with higher accuracy for authorised users and encryption keys held by European institutions rather than the US Department of Defense. It is not new. It has been operational since 2016. What changed after 2022 was the speed at which European military doctrine began treating Galileo dependency as a baseline requirement rather than a preference.
Eutelsat OneWeb provides an existing low Earth orbit constellation, operating at roughly 1,200 km altitude, that several European governments have contracted as an interim capability while IRIS² is built. It is not sovereign in the same sense as IRIS² but it is European-owned, headquartered in Paris, and not subject to US CLOUD Act jurisdiction over its operational decisions.
The pattern across all three programmes is the same pattern visible in the French Teams ban and the ECB Digital Euro decision.
The trigger is not ideology. It is a specific moment when the dependency became visible as a risk rather than invisible as a convenience.
For Starlink, the moment was October 2022 and a switched-off coverage zone in the Black Sea. For Microsoft, it was the CLOUD Act being cited in enough procurement disputes that legal teams stopped treating it as theoretical. For the Digital Euro, it was the explicit demonstration that American cloud infrastructure could be legally compelled to produce transaction data regardless of where it was stored.
Each of these decisions happened separately. Each was framed in local terms by the institution that made it. None of them reference each other officially.
The infrastructure layer underneath all of them is moving in the same direction.
For builders developing products for European public sector, defence-adjacent, or critical infrastructure clients, the satellite question is immediately practical.
Applications that depend on commercial satellite connectivity for European government or military users will face procurement questions about which network that connectivity runs on and who controls the switching decisions. IRIS² is the long-term answer. Until 2030, the question is which interim infrastructure has acceptable sovereignty credentials.
The broader signal is about the architecture of dependency itself. The organisations moving fastest to replace American infrastructure are not doing so because they dislike American companies. They are doing so because a specific moment revealed that the dependency had an owner, and the owner had interests.
Every piece of infrastructure your product touches has an owner. In a stable period, that owner is invisible. The European experience of the last three years is a study in what happens when the owner becomes visible.
404 Found covers AI developments from a European Insider, three times a week.
Our Founder Was Early to AI Winners. Now He’s Tracking What Comes Next.
Over the last 18 months, MavSource’s founder tracked and invested around major AI-related names like Micron +100%, Nvidia +74%, Sandisk +130%, Western Digital +74%, TSM +22%, Broadcom +27%, Okta +35% and Lam Research +39% — an average return of approximately +63% across the group.
Now, he’s bringing the sources and ideas behind those insights to you in one daily email digest. MavSource aggregates the most important AI updates from newsletters, podcasts, company news, AI labs, funding rounds, and more — then summarizes what matters in a simple 5-minute brief.
Past performance does not indicate future results. Informational only; not investment advice.