A quantum computer is usually photographed from the outside: a gold chandelier of cables and plates hanging in a clean room, a few researchers in front of it. The chandelier is not the computer. It is the cooling system, and in most laboratories that run superconducting machines it was made by someone else, in another country, and bought from a catalogue. That is the subject of this piece. Not the qubits, which get the headlines, but the companies and capabilities without which no French quantum machine can be switched on. The Cour des comptes audit of July 2026 says as much: it still recommends growing a French leader in the refrigerators, a market where one Finnish maker has made close to 200 million euros since 2008.
This is the second piece of the quantum chapter of the CareerOn Industry Atlas. The first followed the money and found that the French State plays two roles, planner and buyer. This one follows the hardware underneath the money: the cold, the scarce materials and the electronics that must work inside the cold. For a candidate, these layers matter because they are where a large share of the engineering work sits, and because they are less crowded than the qubit itself.
French quantum depends on three layers it does not yet fully supply: the extreme cold of superconducting machines is bought mainly from foreign makers and a French leader remains a target, the scarce materials have French suppliers named only by the State programme's own report, and control electronics that operate in the cold are named by Brussels and by French industry as a capability still to build; the dependence is documented, the remedies are recommendations, and the careers are in closing the gap.
We read eight dated documents for this piece. Five were already in the series and were read again for what they say about suppliers, not money: the Cour des comptes audit of July 2026, the parliamentary science office's report of 20 January 2022, the national strategy's annual report for 2022 published in March 2023, the European Commission's quantum strategy of 2 July 2025, and the OECD and European Patent Office map of the global quantum ecosystem of 17 December 2025. Three are new to the series: the report of the mission led for the Prime Minister by Paula Forteza, Jean-Paul Herteman and Iordanis Kerenidis in January 2020, the preliminary research and industry agenda of the European Quantum Flagship from November 2022, and the May 2024 report of GIFAS, the French aerospace and defence industry association, on quantum technologies. We refused the Flagship's newer roadmaps, because they are published as consultation drafts whose content the publisher says may change. A draft is not an edition.
The cold is a product
Start with physics, because it sets the shopping list. The OECD and European Patent Office map is plain about what a cooled platform needs: such a platform "would require a cryogenic dilution refrigerator at 50 millikelvin to lower the thermal energy of each quantum bit about six thousand-fold compared with a room-temperature lab". That is a fraction of a degree above absolute zero. It is not reached with a better fridge. It is reached with a dilution refrigerator, a specialised instrument built by a small number of firms.
Who are they? The mission report of January 2020 answered in one sentence. It wrote that the needs of extreme cryogenics were covered mainly by three global actors: Bluefors in Finland, Oxford Instruments in the United Kingdom and Janis in the United States. It then listed a French ecosystem in cryogenics, including Thales, CryoConcept, MyCryoFirm, Absolute System, the Institut Néel and IRIG, and proposed to support a competitive compact cryogenics offer through national innovation competitions and calls for projects. The verbs matter. The foreign firms supply. The French firms are an ecosystem to be supported.
Six years later the audit court picked up the same thread. In its July 2026 report it singles out the Finnish firm: founded in 2008, Bluefors sells the cryostats needed to cool certain types of qubits and has since generated close to 200 million euros of revenue. In the same report it recommends making three industrial leaders emerge in enabling technologies, naming stable isotopes, dilution cryostats and quantum interconnect. A recommendation in 2026 to make a French leader in dilution cryostats emerge is, read plainly, a statement that one does not yet exist at that scale.
The cold is a product, and it is bought abroad
| What the document prints | Status | Publisher, date |
|---|---|---|
| A cooled platform needs a dilution refrigerator at 50 millikelvin | Physical requirement | OECD and EPO, 17 December 2025 |
| Extreme cryogenics supplied mainly by three makers: Bluefors (Finland), Oxford Instruments (UK), Janis (US) | Market reading | Prime Minister’s mission, January 2020 |
| Bluefors, founded in 2008, close to 200 M€ of revenue since | Reported revenue | Cour des comptes, July 2026 |
| Make French leaders emerge in dilution cryostats | Target, not a result | Cour des comptes, July 2026 |
Sources 5 Organisation for Economic Co-operation and Development · 6 Mission parlementaire confiée par le Premier ministre (P. Forteza, J.-P. Herteman, I. Kerenidis) · 1 Cour des comptes
Two cautions keep this reading honest. The first is that the 200 million euros is one company's cumulative revenue since 2008, as the court printed it. It is not a market share, and it does not tell us how many French laboratories bought from Bluefors. The second is that the mission report is a reading of the market in January 2020. It names no shares, and the market may have moved. What the record supports is narrower and still strong: two public documents, six years apart, describe the extreme cold as supplied from abroad, and the later one still treats a French leader as something to build.
The counter-case deserves a hearing. CryoConcept is a French maker of dilution refrigerators, and the mission report names it. A reader could argue that France already has the capability and lacks only scale. That may be true. But none of our eight documents gives CryoConcept's volumes, customers or share, and we will not supply them from memory. Capability without a figure is printed as capability, not as supply.
Scarce materials, named once
The second layer is materials. Some qubits are built in silicon, and the best of them want silicon made almost entirely of one isotope, silicon-28, because the other isotopes disturb the qubit. Some cooling cycles need helium-3, a rare isotope of helium that reaches temperatures below one kelvin. Neither is bought at a hardware store.
Here the French record names suppliers, and names them clearly. The national strategy's annual report for 2022 states that "Orano Stable Isotopes has developed an isotope separation line using gas centrifugation that allows to enrich Silicon-28 up to 99,99%". It then describes Air Liquide's ambition to anticipate the needs for industrial cryogenic systems operating from room temperature down to 20 millikelvin, and lists among those needs "the supply of critical raw materials such as Silicon-28 or Helium-3". Two large French industrial groups, one in nuclear fuel and one in industrial gases, positioned on two of the scarcest inputs in quantum hardware.
Two French material suppliers, named by one document
| What the document prints | Who says it | Publisher, date |
|---|---|---|
| Orano: a centrifuge line enriching silicon-28 up to 99.99 % | The programme, about its own ecosystem | France 2030 quantum strategy, March 2023 |
| Air Liquide: cryogenic systems down to 20 millikelvin, supply of silicon-28 or helium-3 | The programme, about its own ecosystem | France 2030 quantum strategy, March 2023 |
| Support for cryogenics, lasers and stable isotopes; no supplier named | Parliament | OPECST, 20 January 2022 |
Sources 2 France 2030 · 3 Sénat, République française
This is where our first hypothesis had to change. We set out to show that the material suppliers are named by the record. They are, but by only one document, and that document is the programme describing its own ecosystem. The parliamentary report of January 2022 names the same materials as a need, in a programme of support for cryogenics, lasers, stable isotopes and other enabling technologies, but names no supplier. The European strategy of July 2025 writes, at continental scale, that a structural weakness "limits both private investment and the emergence of critical supply chains", and also names no supplier.
So the honest sentence is this: Orano and Air Liquide are named as French suppliers of silicon-28 and helium-3 by the State programme's own annual report, which states capability and ambition, not delivered volumes, prices or customers. No independent document among the eight confirms what they deliver. That is not a reason to doubt the companies. It is a reason not to write that France is self-sufficient in quantum materials, which the record does not say.
Electronics that work in the cold
The third layer is the least visible and, for engineers, perhaps the most interesting. A qubit is controlled and read by classical electronics. Today much of that electronics sits outside the refrigerator, at room temperature, linked to the chip by a thicket of cables. Every cable carries heat into the coldest stage, and every added qubit adds cables. Past a certain size, the cabling, not the qubit, becomes the limit. The proposed answer is to move part of the control electronics into the cold, close to the chip.
The European Quantum Flagship's preliminary agenda of November 2022 names this directly. It recommends: "Support pilot lines for the integration of quantum circuits and control electronics, including the fabrication of classical control chips able to operate at cryogenic temperatures (cryo-CMOS)". In May 2024 the French aerospace and defence industry association reached the same point from the user's side. Its report on quantum technologies for aerospace, space and defence notes that progress is also needed in the field of cryo-CMOS.
Control chips that work in the cold: a capability still to build
| What the document prints | Status | Publisher, date |
|---|---|---|
| Support pilot lines, including control chips able to operate at cryogenic temperatures (cryo-CMOS) | Recommendation | European Quantum Flagship, November 2022 |
| Progress is still needed in cryo-CMOS | Industry judgement | GIFAS, May 2024 |
Sources 7 European Quantum Flagship (Strategic Advisory Board, CSA QUCATS) · 8 Groupement des industries françaises aéronautiques et spatiales
Two documents, one European and one French, one from research and one from industry, name the same capability as unfinished. Neither names a supplier, and neither says whether the pilot lines exist today. We therefore name none. The point the record does support is the direction of travel: control electronics that operate in the cold are treated as a capability to build, not as a component to buy. That has consequences for who gets hired.
What the three layers have in common
Set side by side, the three layers tell one story with three different degrees of evidence. On the cold, the dependence is documented twice and the remedy is a recommendation. On materials, French capability is documented once, by an interested party. On cold electronics, the gap is documented twice and no one is named as closing it. In none of the three does our record show a French supplier at industrial scale with a printed figure behind it.
That is a sharper finding than a general worry about sovereignty, and it should change how the first piece is read. The national plan audited in July 2026 funds research and buys machines. The layers in this piece are what those machines stand on. A national plan can buy a computer from a French start-up and still pay, through that start-up, a foreign maker of the refrigerator inside it. Nothing in our documents measures that flow, and we do not estimate it. But a reader should know the flow exists.
There is a reasonable objection. Dependence on a Finnish supplier is dependence on an ally inside the European single market, and the European strategy treats supply chains at the scale of the Union, not of France. On that reading the issue is European capacity, not French. The audit court's recommendation for French leaders in dilution cryostats shows that the French State does not yet accept that reading for itself. Both views are in the record. We print both.
What this means for your career
For a candidate, the useful reading is that the scarce skills sit below the qubit. Qubit design attracts the most applicants and the most press. The layers in this piece, cryogenic engineering, isotope and materials processing, and electronics designed to operate in the cold, are named by public documents as gaps. A gap named by a State audit, a European agenda and an industry association is where demand for engineers tends to outlast the hype cycle.
Three profiles follow from the record. A cryogenic systems engineer who understands dilution refrigerators, their thermal budget and their integration with a chip is working on the layer the mission report and the audit both describe as supplied from abroad. A materials or process engineer with experience in isotope separation or ultra-pure silicon is working on the layer where two French groups are named. An electronics engineer who can design circuits that keep working at cryogenic temperature is working on the layer that Brussels and French industry both call unfinished.
Our practice simulations follow the same rule as our sources: we place one only where an admitted document names its technology. The simulation Cryogenic electronics engineer: integrated quantum control at 4 K is placed here, because the European agenda names cryogenic control chips as a capability to build, and that is the work the simulation rehearses. The simulation Superconducting qubit engineer: designing and characterising a cat qubit was placed in the first piece, where the audit names cat qubits; none of the eight documents in this piece names them, so we print it as unplaced here rather than stretch it. We have no simulation yet for cryogenic systems or for isotope materials work. Both are recorded as gaps in our catalogue.
If you are choosing where to specialise, ask a prospective employer one question this piece makes possible: which parts of your machine do you make, and which do you buy? The answer tells you where the company's engineering actually happens, and whether the role you are offered sits on a layer the State is trying to bring home.
What we could not prove
We could not find, in any admitted document, a market share for dilution refrigerators in France, delivered volumes of silicon-28 or helium-3 from French suppliers, or the existence of a European cryo-CMOS pilot line. Each of those would change the strength of a sentence above. The Flagship's newer supply chain roadmap may hold some of them, but it is a consultation draft and we did not admit it. When a dated edition appears, we will read it and update this piece with a recorded correction.
The next piece of the chapter reads what Europe’s quantum industry is building in 2026: which machines, lines and programmes are dated for delivery, and which have slipped. After it come the arguments inside the field and, last, the people: which quantum jobs exist in France today, which are announced, and which documents tell the two apart.