Executive brief
Crolles is simultaneously recruiting and shedding, and both facts are documented. On the hiring side: the 2022 announcements attached more than 1,000 jobs to the site extension, France 2030 committed to training 35,000 people in electronics and robotics by 2030, a €10 million microelectronics training programme runs from 2023 to 2028 with Grenoble INP among its partners, and a dedicated FD-SOI design academy opened in Grenoble in 2025. On the shedding side: STMicroelectronics announced in April 2025 a plan reported at about 2,800 positions worldwide and roughly 1,000 voluntary departures in France by the end of 2027. This dossier holds both, and then does the harder thing: it states which career figures a reader might expect and which of them do not exist in any source we could open — including an aggregate current vacancy count for the site, any semiconductor-specific attrition rate, and any published wage grid for the sector. The one pay figure we found attached to a named Crolles-area posting is printed exactly as advertised, with its scope and its limits.
I. The paradox to resolve, not to smooth over
A candidate reading regional coverage of Crolles in this window encounters two irreconcilable stories. One says the largest chip site in France is expanding, backed by national strategy and thousands of trained entrants. The other says it is cutting. Most coverage picks a side. Both are documented, and the resolution is not a compromise — it is a distinction between stocks and flows.
The cuts are a flow out of an existing cost base, concentrated on the group's engineering and support population and executed through voluntary departures over three years. The training programmes are a flow into a national talent pool that no single employer controls, aimed at a decade. Those two flows can run in the same direction on a chart and still describe entirely different labour markets: a plant can reduce its own headcount while the corridor around it absorbs more entrants than before, because the entrants are being trained for an industry rather than for a company.
What follows is the evidence for each flow, with dates and publishers, and a candid inventory of what nobody has measured.
II. The demand side: what was promised and what was published
| Figure | Scope | Publisher and date |
|---|---|---|
| More than 1,000 jobs created at the site | Site extension, announced alongside the €5bn national electronics strategy | French Presidency / national coverage, 11 July 2022 |
| Approximately 5,000 employees on site, with 1,000 additional jobs by 2026 | Site, forward-looking | Regional and trade coverage, 2022 |
| More than 4,300 employees at Crolles, more than 2,000 at the Grenoble R&D centre | Site and separate R&D campus | STMicroelectronics site page, consulted 4 September 2026 |
| Close to 3,300 employees, revenue €1,617 million | Single legal entity STMicroelectronics Crolles 2 SAS, FY2024 | Financial-registry data, 2024 accounts |
| About 2,800 positions worldwide; approximately 1,000 voluntary departures in France by end-2027 | Group, then France | EE News Europe 14 April 2025; Reuters, Les Echos, EE Times, Le Parisien/AFP 30 April 2025 |
| 300 to 400 jobs described as at risk at Crolles | Site, union estimate — not company-confirmed | Local union sources via actu.fr, 2025 |
Read the table as a scope problem, not a contradiction to be solved. Site-wide employment, one legal entity's payroll, a forward-looking creation target and a three-year voluntary-departure programme are four different measurements. The union figure is the one most often repeated as if it were a company number; it is not, and we label it every time.
III. The supply side: the training pipeline that exists
The pipeline into Crolles-type work is unusually well documented for a French industrial region, because the state has funded parts of it explicitly.
- France 2030 electronics strategy, 11 July 2022. Announced at Crolles, with €5 billion of national support and a stated objective to train 35,000 people in electronics and robotics by 2030. Read the wording carefully: the target covers electronics and robotics, not semiconductor manufacturing alone. We could not corroborate a semiconductor-only sub-target, so we do not present the 35,000 as a chip-industry figure.
- INFORISM, 2023–2028. A microelectronics training programme awarded a €10 million France 2030 subsidy over the period, with Grenoble INP among its partners; the associated national research funding record (project reference IA-23-CMAS-0024) confirms it sits inside the "compétences et métiers d'avenir" framework. This is the most concrete public commitment to microelectronics skills capacity in the region we could source.
- Grenoble INP – Phelma apprenticeship. A three-year microelectronics and telecommunications engineering programme delivered by apprenticeship, running since September 2013. Its Nanotech track is a separate, distinct programme — a distinction routinely blurred in coverage.
- FAMES Academy, 2025. Launched in Grenoble to train engineers in FD-SOI and advanced design, i.e. aimed squarely at the platform the companion technology dossier identifies as the site's spine.
- CEA-Leti cleanroom platform, with a €42 million renovation programme reported in 2024 — infrastructure that serves both research and hands-on training on industrial-grade equipment.
The mechanism worth naming: three of these five instruments train for a platform (FD-SOI, 300 mm processing, cleanroom practice) rather than for an employer. That is what makes the pipeline robust against one company's pause — and it is also the reason a candidate should evaluate a programme by which platform it certifies them on, not by which company recruits from it.
IV. The roles the site actually posts
Job-board evidence for the site, at the time of research, is concentrated in process and production work inside the cleanroom rather than in design: operator and process-technician roles on continuous shift patterns, equipment maintenance, and production support. That composition is consistent with what the companion technology dossier establishes — Crolles is a front-end manufacturing site, and front-end manufacturing is a technician-heavy activity.
One pay figure is directly attributable. A staffing-agency posting for a production role at the Crolles site advertised €24,000 to €28,000 gross per year. We print it because it is real and sourced, and we immediately bound it: it is a single agency posting for an entry-level production role, it is not a company pay scale, it is not an average, and it says nothing about shift premiums, night and weekend uplifts, or the collective agreements that materially change take-home pay in continuous-shift semiconductor work. Anyone reading it as "what Crolles pays" is over-reading a single advertisement.
Not published, and we looked: no aggregate count of current open positions at Crolles could be verified; no semiconductor-sector wage grid from a public statistical source (INSEE, DARES or equivalent) could be sourced at the granularity this dossier would need; no attrition or turnover rate specific to the semiconductor industry in France could be found from a credible statistical publisher; and no union-published pay or turnover dataset for the site could be located. Four figures a reader would reasonably want, and four absences we will not fill with estimates.
V. Transferability: what the evidence supports and what it does not
The most valuable question for a candidate is not "will Crolles hire me" but "if Crolles stops hiring, where does this experience travel?" Here we must be more careful than most career writing is.
What the evidence supports. The disciplines that govern 300 mm front-end work — cleanroom protocol, contamination control, statistical process control, metrology, equipment qualification and maintenance, documented deviation handling — are generic to high-purity manufacturing. They are the same disciplines named in the operating requirements of photovoltaic cell lines, sterile pharmaceutical production, flat-panel display fabrication and the microLED front-end documented in our Aledia dossiers. On that basis, cross-sector mobility is technically plausible, and the corridor around Crolles contains several of those destination industries within commuting distance.
What the evidence does not support. We could find no credible source documenting actual worker mobility between Crolles and photovoltaic, pharmaceutical, display or microLED employers — no cohort study, no employer survey, no placement data. Plausibility is not evidence. So the honest formulation is: the skills are structurally portable, and the portability is undemonstrated in published data. A dossier that told you "process technicians routinely move from Crolles into pharma" would be asserting a fact nobody has measured.
We also could not verify the existence of a Grenoble-specific "école des métiers du semi-conducteur" of the kind sometimes referenced in commentary. It does not appear here.
VI. Reading the restructuring correctly
The April 2025 programme matters to a candidate less for its headline than for its structure. Three features are documented and each changes the risk calculation.
It is framed around voluntary departures in France — roughly 1,000 by the end of 2027, per the 30 April 2025 reporting. Voluntary schemes fall disproportionately on populations with outside options and on those near retirement, which is a different exposure profile from compulsory reduction. It is multi-year, which means the site is not in a single shock but in a managed glide. And it is group-wide: the 2,800 figure is worldwide, and no company document we could open allocates it to sites, which is precisely why the Crolles-specific 300–400 remains a union estimate.
The industrial-relations record is dated and public: the CGT delegation at Crolles issued a formal unfavourable opinion at the works council of 18 July 2025, criticising public announcement ahead of employee-representative information; the CFDT metals federation traces the change of industrial model to statements made to representatives in November 2024. A candidate in due diligence can ask about both by name.
VII. The five questions to put to a Crolles employer
- Which programme funds this role — running 300 mm production, the paused joint fab, or the silicon-photonics expansion still under decision?
- Is the position inside the perimeter of the voluntary-departure plan announced in April 2025, and if not, why not?
- Which platform will I be certified on — FD-SOI device integration, generic 300 mm process, optical test — and who else in the corridor recognises that certification?
- What is the shift pattern and the full compensation structure, including continuous-shift premiums, rather than the base band advertised?
- What proportion of the team was hired in the last twenty-four months? It is the only turnover indicator a candidate can obtain, given that no published attrition data for the sector exists.
VIII. The analytical conclusion
The strategic finding on this anchor is that the Grenoble semiconductor corridor has decoupled skill demand from employer demand, and Crolles is the proof. The site's own headcount is under a managed multi-year reduction while, within the same twenty kilometres, a €10 million microelectronics training programme runs to 2028, an FD-SOI academy opened in 2025, a €42 million cleanroom renovation was reported in 2024, and a national target of 35,000 electronics and robotics trainees by 2030 stands. Those instruments are not betting on one company. They are provisioning a platform that Soitec, CEA-Leti, Aledia, Quobly and STMicroelectronics all draw from.
The career implication is unglamorous and, we think, correct: optimise for platform-portable competence, not for employer prestige. On the published evidence, the durable asset in this valley is certified 300 mm process discipline with metrology and equipment depth — a skill set the corridor keeps funding regardless of which building gets built next. The undemonstrated part is the one we refuse to dress up: nobody has measured whether those workers actually move between sectors when a line stops. Until someone does, treat portability as a well-founded expectation rather than a documented outcome.
The capital record behind the pause and the aid contradictions is documented in the first dossier on this anchor; what the site actually manufactures is documented in the second.