✍️ Who signs first?

2.4 million qubits have already come off STMicroelectronics’ lines, etched on 300mm FD-SOI wafers using the same process the company runs for automotive chips. They belong to Quobly, a Grenoble spinout that has just closed a €115m Series A and promises its first commercial machine, accessible via cloud, before December. Standard silicon, in a fab that already runs at volume.

The cap table reads like a supply chain. STMicroelectronics and SEALSQ invest alongside Bpifrance, the French state fund, with the European Innovation Council and Air Liquide’s venture arm in support; Soitec and Orano anchor the industrial partnerships. Suppliers, shareholders and future customers overlap almost entirely. In hardware, that overlap is the only known way to get from prototype to production in under fifteen years.

Brussels has just adopted Chips Act 2.0: €120bn by 2035, including €30bn for an advanced foundry. Article 30 lets a public buyer require bidders to declare the European share of chips in their products, and to dual-source with at least one domestic supplier. The first Chips Act subsidised capacity. This one organises demand.

Magdeburg explains the pivot. €9.9bn in German subsidies sat waiting for Intel’s fab; the project died, the money went back to the budget, and Europe still has no leading-edge node on its soil. France has committed €8.7bn to the sector since 2018, by its own auditors’ count. Infineon, meanwhile, opened a €5bn plant in Dresden three months ahead of schedule, because the orders for power chips were already booked.

Half the ecosystem still reads a state-led round as a distress signal. On a quantum roadmap running to 2032, that reflex is absurd: the shareholder you want is the one who will not need to exit in year seven. As for the 5% of global compute Europe brandishes like a wound, a large enough cheque to Nvidia would restore it within a quarter, without one additional wafer leaving a European line.

Quobly has invoiced roughly €10m in two years, almost all of it under Proqcima, the French programme targeting two sovereign prototypes by 2032. Public customer, public shareholders, public money. The market test only begins when a bank or a compute operator signs of its own accord, and European quantum has watched excellent teams end up on American balance sheets before. Ask Oxford Ionics, sold to IonQ last year.

You do not close a compute gap by buying compute. You close it by manufacturing, and manufacturing requires someone committed to buy. Europe has finally put the two in the right order. The customer, for now, is still Europe itself.