Overnight into 14 October 2026, Ukrainian units reported that their Starshield terminals were degrading once vehicles or drones moved faster than about 160 km/h (100 mph). The Kyiv Independent reported that the restriction came without warning. A defence adviser to President Zelenskyy said SpaceX had "implemented a speed limit", and Zelenskyy said it "was not a technical problem; it was the will". Eight days later, on 22 October, ICEYE and Nokia signed a letter of intent with Ukraine's Defence Ministry to explore a satellite communications system under Ukrainian control. The sequence is a clean case study in infrastructure dependency, and also in how easily a policy response can be oversold.
The strongest case for the restriction
Start with the argument for the other side. Starshield is not an ordinary consumer product. Musk's own position, as reported by Al Jazeera, is that the systems are "owned and controlled by the US government" and that limits on speed or usage are decided by the US military and ultimately the president. On that view, a government that owns a military asset is entitled to set its terms, and a supplier that keeps an uncontrolled long-range strike capability on a commercial network could be exposed to escalation risk. Musk later claimed that Ukraine's military had violated Starlink's terms of service by using it for offensive drone strikes, and that a loophole was closed. Ukrainian officials disputed that account.
That is a legitimate argument about governance, even if the execution here was poor. A supplier or sponsor is entitled to set conditions. What it is not entitled to expect is that an ally will accept those conditions being changed silently, mid-war, with no notice.
What the outage actually showed
The weakness was not the cap itself but the absence of any process around it. ABI Research noted that the service is funded by Germany under a contract running until April 2027 and operated by Ukraine, yet neither was consulted. Militarnyi reported that strike drones lost connection at speeds as low as 80 km/h, and that at least three units reported regular dropouts at 75 km/h, and that some missions were affected. That outlet also relayed one analyst's view that a software glitch could not be ruled out, so the cause remains contested.
The blame was not resolved either. The Kyiv Independent reported on 24 October that nobody had publicly taken responsibility for the restriction, with Trump administration officials pointing at Musk. It also reported, citing the Wall Street Journal, that Trump had ordered Musk on 21 October to lift the limits. Trump himself said: "I've heard it's fixed. If it's not, I will speak to him." The practical lesson is that when an ally's battlefield connectivity depends on a private firm and a government that each point at the other, no one is accountable for an outage.
The diplomatic reaction was notable. The UK prime minister's readout of a 23 October call with Trump stressed "the importance for Ukraine to have full access to Starlink so they can continue to defend themselves". Governments that rely on a single provider ended up asking for access to be preserved, which is what dependency looks like.
Restoration, and what it does not prove
On 23 October Zelenskyy said the speed limits remained in effect. The next day Defence Minister Yevhenii Khmara said the ministry was recording speeds above 400 km/h with a stable connection, and that it would keep monitoring and documenting performance. Zelenskyy said service had returned to pre-restriction levels.
Restoration is good news, but it is not a reason to relax. It shows that the dependency can be reversed by a phone call, which also means it can be imposed by one. The outcome rested on a political decision, not on any durable rule or contractual guarantee.
The Finnish letter of intent: promising, but thin
The 22 October document was announced during Finnish Prime Minister Petteri Orpo's visit to Kyiv. According to Nordic Defence Sector, it commits the parties to explore cooperation on sovereign satellite communications under Ukrainian control. It is not an announced procurement agreement. The companies disclosed no financial commitments, procurement plans or timetable, and the outlet noted that the companies' separate low Earth orbit partnership, announced on 1 October, expects its first communications satellites to launch in 2028, with no corresponding schedule for Ukraine. TelecomsTechNews likewise reported that no pricing, delivery timetable or terminal volumes were disclosed.
This matters for analysis. ICEYE is best known for Earth-observation satellites, and a communications constellation with global coverage, ground segments and ruggedised terminals is a different undertaking. Ukraine may well be better served by a layered approach that combines European sovereign capacity with commercial services, than by a single replacement for a system that works at 400 km/h today.
A proportionate response
A pro-innovation reading avoids two errors. The first is treating this episode as proof that commercial satellite services are untrustworthy and must be replaced by state-run systems. Starlink's rapid deployment in 2022 was possible precisely because a private company moved faster than any procurement office could. The second is shrugging off the risk because service came back after a week.
The sensible policy is contractual and architectural, not protectionist:
- Written terms of use. Any government paying for or depending on military satellite service should have notice periods and escalation procedures written into contracts, not left to personal messages.
- Multi-vendor terminals. Hardware and software that can switch between providers reduces the cost of any single supplier's decision.
- Open competition. European projects should be judged on delivered capability and published schedules, not on announcements. A letter of intent without a timeline should be treated as a starting point.
- Transparent accountability. Who decides whether a military terminal's speed is capped should be a matter of record.
Ukraine's experience this October does not argue against commercial satellite innovation. It argues for treating connectivity as critical infrastructure, with the redundancy and rules that implies, and for judging the Finnish effort by what it launches and when.