A bigger constellation, a slower clock
On August 7, 2026, the European Commission and the SpaceRISE consortium — Eutelsat, SES, and Hispasat — signed the implementation agreement that expands IRIS², the EU's sovereign satellite constellation, from its original design to 348 satellites: 330 in low Earth orbit and 18 in medium Earth orbit, including a dedicated 66-satellite layer built specifically for defense and security users (European Commission). The Commission says the expansion delivers 60% more secure governmental capacity within the EU and 54% more worldwide, and Space Commissioner Andrius Kubilius has pledged to bring the network into operation "ahead of schedule" (EUSPA). First launches are still targeted for 2029.
Neither official announcement mentions Ukraine or Starlink by name. But the timing is not subtle. The agreement landed ten days after Volodymyr Zelenskyy, in White House talks on July 28, 2026, asked Donald Trump to persuade Elon Musk to loosen SpaceX's acceptable-use restrictions so Ukraine could route drone strikes against Russian ballistic-missile launchers through Starlink — a request Trump reportedly left noncommittal, after Musk himself declined an in-person meeting with Zelenskyy to discuss it (19FortyFive). SpaceX's policy is explicit: Starlink service "cannot be used to remotely pilot drones or for any military activity unless SpaceX explicitly permits it" — a line Gwynne Shotwell drew as early as February 2023 and that Musk enforced personally when he declined to extend coverage over Crimea that same year.
The steelman for Brussels
The case for IRIS² deserves to be made on its own terms before it's criticized. A single American company controlling the satellite links that Ukrainian artillery spotters, hospitals, and municipal governments depend on is a real strategic exposure — not a hypothetical one. It is entirely reasonable for European governments funding a war effort to want infrastructure whose usage terms they set, rather than infrastructure whose usage terms are set unilaterally by a foreign CEO who has already refused one battlefield request and left another unanswered for weeks. Sovereign governmental connectivity, procured through a consortium with EU oversight, is a legitimate hedge against exactly the kind of discretionary veto Kyiv is currently negotiating around.
That argument just ran headfirst into its own timeline problem this month.
The dependency IRIS² can't touch yet
While Brussels was finalizing satellite numbers for a constellation that won't fly before 2029, Ukraine's actual battlefield connectivity spent mid-August wobbling over a dispute that had nothing to do with Russia. Poland funds roughly $50 million a year covering about half of the 42,000 Starlink terminals in active use across Ukraine. When Starlink's new European roaming rules excluded Poland from a 30-country zone starting August 12, Polish Foreign Minister Radosław Sikorski threatened publicly to "re-think paying you $50 million p.a." unless SpaceX reversed course, and Prime Minister Donald Tusk demanded assurances the spat wouldn't degrade the Starlink terminals Poland funds inside Ukraine, which he said "help detect aerial threats more quickly." SpaceX backed down and restored Poland's roaming access by August 13 (Kyiv Independent), but the episode is the clearest illustration yet of the actual failure mode: Ukraine's frontline internet is one billing dispute, one policy change, or one executive decision away from disruption, and it has no operational alternative today.
This is the gap IRIS² doesn't close. Ukraine isn't even an IRIS² member yet — Kubilius said in mid-2025 he "would not be opposed" to Ukrainian accession given the country's Soviet-era space heritage, but formal negotiations haven't concluded, and the program's own members won't have satellites in orbit for another three years (United24 Media). Kyiv is separately pursuing its own path — the Ukrainian firm Stetman has floated a $1.1 billion, 360-satellite national network targeting a 2030 rollout — but that, too, is a promise for the next decade, not this one.
The real lesson is diversification, not substitution
The instinct to read IRIS² as "Europe's Starlink" gets the policy backwards. Swapping one monopoly provider for another sovereign-but-singular one doesn't eliminate concentration risk — it just relocates the veto point from Musk's office to a Brussels consortium managing 27 member states' competing defense priorities, which has its own track record of moving slowly. The proportionate response to what Ukraine is living through isn't a single European alternative to Starlink; it's genuine multi-vendor redundancy — Starlink, IRIS² once it exists, OneWeb, and national systems like Stetman operating in parallel, so that no single commercial or political relationship can degrade frontline connectivity by itself. That's a market-competition argument, not a sovereignty argument, and it's the one EU policymakers should be making explicitly rather than letting the IRIS² rollout imply it's already solved a problem it won't touch until the decade turns.
In the meantime, the immediate lever remains diplomatic, not orbital: getting SpaceX to clarify, rather than negotiate case-by-case, what battlefield uses Starlink will support in an active war. That's a policy Musk can change by email. IRIS²'s 348 satellites can't reach orbit before 2029 no matter how many implementation agreements get signed this August.