Microelectronics is to the 21st century what electricity was to the 20th: the invisible substrate of the industrial economy. A washing machine contains about twenty chips, a car around a thousand, a plane several thousand.
Without them, no technological leap—artificial intelligence, quantum, defense—would be conceivable. The Court of Auditors has scrutinized seven years (2018-2025) of public policy in this sector. Its verdict: extraordinary financial efforts, real technological results, but an incomplete strategy and insufficient management.
| Indicator | Value |
|---|---|
| Public aid programmed (2018-2025) | €8.7 billion |
| Aid actually disbursed over the period | €5 billion |
| Employees in the French supply chain | 53,600 |
| EU share of global chip production | 7% |
A Strategic Sector: Where Does Europe Stand?
The European Union accounts for only 7% of global semiconductor production, far surpassed by Asia and the United States. Even more worrisome: it is almost absent from the segment of the most advanced chips, essential for artificial intelligence and breakthrough innovations.
In the face of shortages revealed by the health crisis of 2020-2021 and under pressure from geopolitical tensions, the great powers launched a frantic race for subsidies. The American Chips Act (CHIPS and Science Act, 2022) committed $52 billion; China has been investing heavily for years; the European Union responded in 2023 with its own Chips Act, endowed with €43 billion.
France is not without strengths. It hosts CEA-Leti, one of the three most advanced microelectronics research centers in the world, alongside Belgium’s IMEC and Germany’s Fraunhofer. Its sector, organized around about a hundred companies and 53,600 employees, accounts for 11% of European production. STMicroelectronics, the Franco-Italian flagship, provides an unusually strong industrial foothold on the continent. In 2024, the sector even posted a trade surplus of €1.8 billion.
Almost €9 Billion Without a Consolidated Dashboard
This is one of the most striking revelations of the report: until this inquiry, there was no consolidated consolidation of the public support poured into the sector. The Court reconstructed, with the help of a data science expert, every financial flow since 2018. The result: €8.7 billion programmed, of which €5 billion actually disbursed over seven years.
The State accounts for the vast majority of the effort (€7.7 billion programmed, €4.3 billion disbursed). Local authorities contribute €220 million and European funds, paradoxically, account for a minority share of €715 million. In addition to these subsidies, there is indirect support via the research tax credit (€1.6 billion over the period) and public equity through Bpifrance and the CEA, totaling an extra €3.6 billion not counted in the total.
These amounts are exceptional. When scaled to the number of supported companies, public aid to the microelectronics sector ranks among the highest in the history of France’s industrial policy.
Yet the Court notes weak conditionality. Unlike the United States or Japan, France imposes few requirements regarding domestic production, job localization, or taxpayer payback on investment.
Liberty: a €7.5 Billion Bet Undermined by a Partner
The Liberty project crystallizes the ambitions and contradictions of France’s strategy. Rolled out at the Crolles site in Isère, it pairs Franco-Italian STMicroelectronics with American GlobalFoundries to build new factories intended to double France’s chip production capacity by 2028.
The total announced investment amounts to €7.5 billion, with French public funding of €2.9 billion (€1.8 billion for GlobalFoundries, €1.1 billion for STMicroelectronics), i.e., about one third of all public support to the sector during the period.
Problem: as of the end of June 2025, GlobalFoundries had not yet started its portion of the project. No payments had been made to it. As a result, the production capacity targets underpinning France’s entire strategy will have to be revised downward.
The Court also notes that the legal framework governing the two grant agreements is insufficient. The state pays each company separately, without a mechanism ensuring the project’s overall coherence if one partner disengages.
As of end-June 2025, €574 million had been paid to STMicroelectronics, but no payment had been made to GlobalFoundries, which had not started its portion of the project.
The Court also recalls that Article 17 of the Public Finance Programming Law requires a prior socio-economic assessment for any public investment exceeding €20 million, and a counter-expertise beyond €100 million. For Liberty, this assessment is incomplete. A clear shortcoming for a project of this scale.
Neither Mapping Nor Quantified Goals by Chip Type
The sages on Rue Cambon identify a major blind spot: France does not have a precise map of its supply and demand for electronic components. Three fundamental questions remain officially unanswered: Is France planning to enter the race for the finest chips? Which application markets does it intend to prioritize? Is its ambition in design, manufacturing, or both?
Without this compass, public support risks being scattered without coherence or prioritization. Coordination with European partners, notably Germany on production capacities, is judged insufficient.
Each member state lays out its own capacity strategy, fueling bidding wars of subsidies to attract foreign investors, while no single Union country possesses the resources to rival Washington or Beijing on its own.
Aids Too Slow, Too Complex, and Poorly Evaluated
The effectiveness of the aid is hampered by delays incompatible with the sector’s speed. Companies had to wait on average three years between filing a dossier and the contracting for the first PIIEC (Important Project of Common European Interest). In a market where technological cycles are counted in months, this slowness constitutes a real competitive handicap.
Governance was streamlined under France 2030, with centralized leadership by the Secretariat General for Investment (SGPI). But overall visibility remains blurred by the layering of multiple schemes.
SMEs, in particular, suffer from a disproportionate administrative burden. PIIEC dossiers require volumes of documentation such that they can be de facto edged out in favor of large groups able to absorb these management costs.
As for evaluating the effects of the aid, it remains embryonic. Admittedly, activity indicators are broadly rising. The turnover of the French sector grew by 46% between 2018 and 2024, versus 35% for the global sector.
Technological advances from the Nano 2022 plan are real. But the link between these results and the subsidies received is not demonstrated. The impact on employment remains undetermined. And above all, there is no indicator allowing measurement of progress in industrial sovereignty, yet the stated objective of public policy.
An Existential Challenge for Industrial Sovereignty
At the end of its analysis, the Court does not question the necessity of public support for the sector. In a world where the United States and China generously subsidize their chip industries, inaction would amount to an abdication. But it calls for a fundamental transformation of the method.
The public money is there, abundant and growing. What is still missing is a clear vision of what France wants to produce, for whom, and with which success indicators. Without this strategic compass, spending risks being generous without being sovereign.
