Where Europe's capacity runs out.
Europe has decided to pay for its own security. The harder question is where that money will hit a wall, and what it takes to move the wall.
In 1932 Carl Anderson photographed a track in a cloud chamber that bent the wrong way. The particle that made it could not be seen. What could be seen was its trail: a line of droplets that condensed in supersaturated vapour wherever the particle had passed. That trail was the first evidence of the positron, the particle Paul Dirac's theory had predicted.
This series borrows the method. Most of what threatens Europe's security is invisible at the moment it matters: a drone over a runway at night, a vessel loitering above a cable, a dependency written into a contract a decade ago. What can be seen is the trail it leaves in incidents, budgets, statistics and decisions. Each week, Cloud Chamber reads one of those trails and asks what made it.
Money is no longer the limit
For most of the past thirty years, the honest answer to "why can't Europe defend itself?" was that it chose not to pay. That has changed. At The Hague in June 2025, NATO allies committed to spending 5% of GDP by 2035: at least 3.5% on core defence, and up to 1.5% on protecting critical infrastructure, defending networks, civil preparedness, innovation and the industrial base.1
When a system has money and still can't produce, the limit is somewhere else. Throughput is set by the narrowest point; spending anywhere else changes nothing downstream. So the useful question for the next decade is not how much Europe spends, but where that spending will hit a wall. We see four walls.
Wall 1 · The seam
Europe's most exposed infrastructure sits in the seam between civilian and military responsibility. A subsea cable is owned by a company, lies in waters patrolled by a navy and carries traffic for a whole continent. An airport is run by an operator, regulated by an aviation authority and defended, in principle, by the state. When something hostile appears in the seam, each party can reasonably believe it belongs to someone else.
The trail is easy to read. By one fact-checked count, drone disruptions at European airports roughly quadrupled between January 2024 and November 2025.3 In September 2025, sightings closed Copenhagen Airport, and a study published in July 2026 plotted 144 suspected drone sightings across northern and western Europe between 2024 and 2026.4 At sea, the Irish government estimates that about 75% of transatlantic subsea cables pass through or near Irish waters.5
The asymmetry is the point. A drone bought online can stop an airport for hours. The cost curve favours whoever is probing, not whoever is watching. Europe will not close that gap by buying military systems for every civilian site. It closes it when watching becomes cheaper than probing.
Wall 2 · Dependence
Between 2020 and 2024, 64% of the major arms imported by European NATO members came from the United States, up from 52% in the previous five years.2 That figure says nothing bad about the supplier. It says something about leverage: whoever builds a system controls its upgrades, its spare parts and, increasingly, its software.
Sovereign defence doesn't mean making everything in Europe. It means knowing which parts carry the load, and making sure those are built, maintained and changed here, under European control.
Wall 3 · Power
Every modern defence system runs on electricity and computing, and both are now contested. In 2025, data centres used 23% of Ireland's metered electricity, up from 5% a decade earlier.6 Ireland is an outlier, but it shows where the rest of Europe is heading: AI, reshored industry, electrification and defence will all draw on the same grids. A plan for sovereign AI that doesn't account for power isn't a plan.
Wall 4 · Time
The last wall is time: the gap between seeing a new threat and fielding a response that works. Cheap drones went from hobby to weapon of war in a few years, while major defence programmes still take many years to buy. The side that iterates faster sets the terms. Europe has the engineers to iterate quickly. Its institutions are still learning to let them.
What would move the walls
We hold a few convictions about solutions. Every essay in this series will test them.
- 01Watch persistently. Protection usually fails because nobody was looking at the right moment, not because nobody could act. Early warning comes first.
- 02Make watching cheap. Sensing has to cost less than the threat it watches for, or it will never scale to thousands of sites.
- 03Measure honestly. Claimed performance means nothing without field data: detection rates, false alarms and the conditions in which a system fails.
- 04Defend, and keep it European. Detect, protect, defend, under European control.
- 05Reason from first principles. Start from physics and cost, not from how things have always been bought.
About this series
Every week: one development at the frontier, the risk it reveals and what would fix it. We use public sources and cite them. We don't publish the weaknesses of specific sites, and we don't speculate about people. When we're wrong, we'll say so.
Europe moves at the speed of its narrowest point. This series is about finding it.
Deltrac
← All essays