Two opposing forces tug at EU AI policy. On the one hand, Brussels politicians paint EU dependence on US AI tech as a source of concern and vulnerability. Not least given America's volatile politics, Europe must stand on its own AI feet, so the thinking. On the other hand, the EU cannot afford to cut the umbilical cord to US-developed hardware and funding if it wants its firms to keep progressing. In this seminar, Daniel Mügge unravels the distinct threads that run through EU AI policy and options for the way forward. At the same time, he questions whether the terms of the "EU AI sovereignty vs transatlantic alliance"-debate do not miss something essential: a public debate about which AI future we want in the first place.