SpaceX knew about the potential collision, Holger Krag, head of the ESA's Space Debris Office, told Forbes — but refused to do anything about it and wouldn't say why.
According to the Forbes story, the U.S. military noticed that one of SpaceX's Starlink satellites had a 1 in 1,000 chance of colliding with the ESA's Aeolus Earth observation satellite.
The Pentagon informed both agencies of the potential collision, and although the ESA's satellite had been occupying the region for nine months longer than the Starlink satellite, SpaceX told the ESA via email that it didn't plan to take any action to prevent the collision and didn't offer any explanation as to why.
This prompted the ESA to perform the avoidance maneuver, which it does regularly to avoid collisions with dead satellites or space junk — but rarely to avoid active satellites and never before to prevent a collision with a mega constellation of them. But the fracas raises the possibility that similar maneuvers might become more regular in the future.
While SpaceX's Starlink mega constellation currently consists of fewer than 60 satellites, it could eventually launch upwards of 12,000 as part of its plan to deliver broadband internet to the entire globe — and that could dramatically increase the risk of collisions.
"We are not upset by them saying [they wouldn't move]," Krag told Forbes. "My concern is how often will we have such events in the future? These are just two satellites. Now they will add several thousand, and they will also be disposed and end up at various altitudes. And there's no rule or law on how to react, it's all goodwill."
To prepare for a future in which potential Starlink crashes are far more common, the ESA is now exploring the use of artificial intelligence to automate its collision avoidance systems, tweeting that such a system is "becoming necessary to protect our space infrastructure."
According to Engadget, the Starlink satellites are already equipped with automated collision avoidance systems, so perhaps SpaceX's decision to do nothing was part of a plan to see if those systems would work as intended, by automatically adjusting the satellite's path without SpaceX manually intervening.
That could explain why SpaceX didn't provide the ESA with a reason for its inaction — it's hard to imagine the agency would've been OK with SpaceX risking the ESA's $560 million satellite in order to test its own AI.
More on Starlink: SpaceX Just Unleashed 60 Starlink Satellites Into Orbit