Skip to main content
Michelin research engineers at Ladoux reviewing tyre simulation results, the reskilling and hiring evidence behind them
SKILLS DEVELOPMENT
5 min read

Does Reskilling Work? Michelin's Own Numbers, Fifteen Months After Cholet and Vannes

Michelin promised tailored support to every worker affected by the Cholet and Vannes closures. Fifteen months later its own mandated consultancy reported 522 placements at Cholet. This dossier tests the promise against the placement data, the 2026 vacancies at Ladoux, and the one independent study that says which kind of retraining actually changes an employer's mind.

Evidence reviewed · September 2, 2026 · 19 sources

Executive brief

On 5 November 2024 Michelin announced its intention to close Cholet and Vannes and committed to tailored support for every affected employee (Michelin, 5 November 2024). One year on, the group published a quantified review (Michelin, 4 November 2025), and La Tribune reported on 5 November 2025 that of 1,246 employees made redundant, slightly fewer than half had not yet found a durable solution. By the seventh monitoring committee, Randstad, the consultancy Michelin itself mandated, reported 522 employees with a professional solution (Les Echos, 17 February 2026). A promise of support is testable. This dossier tests it, then asks what the company hires for now.

One closure, four published totals

What was publishedValue as printedDatePublisher
Employees concerned across Cholet and Vannes at announcement1,2545 November 2024Le Parisien
Employees made redundant, as restated a year later1,2465 November 2025La Tribune
Jobs removed by the Cholet closure92717 February 2026Les Echos
Cholet employees counted by the prefecture at the monitoring committee94716 February 2026actu.fr

Sources 2 Le Parisien · 4 La Tribune · 7 Les Echos · 8 actu.fr

I. The commitment, and the number that measures it

Michelin's closure announcement of 5 November 2024 gave a commitment of individual accompaniment rather than a numeric placement target, and no breakdown of internal transfer against retraining against external placement was published then or since; we treat that breakdown as unverified. Le Parisien reported on 5 November 2024 that 1,254 employees were concerned across the two sites, and the figure most often used a year later was 1,246 redundancies (La Tribune, 5 November 2025). AEF info reported on 5 November 2025 that roughly half of the workforce at the two sites had already found a durable solution.

The Cholet site data is the most granular, and it comes with a discrepancy worth naming rather than smoothing. Les Echos reported on 17 February 2026 that the closure removes 927 jobs and that Randstad counted 522 employees with a professional solution. Actu.fr reported on 16 February 2026, from the prefecture, 522 of 947 employees with a solution, about 55%, alongside €403,000 of validated aid and €6.83 million committed. Two site totals, 927 and 947, appear in dated reporting from the same week. We publish both.

Read plainly: after fifteen months, with a mandated outplacement consultancy, a supportive prefecture and millions of euros committed, roughly one in two workers had a durable outcome. That is neither a scandal nor a success story. It is the realistic yield of industrial redeployment in France, and it is the single most useful number a worker facing a closure can hold in mind.

522

Cholet employees with a professional solution, counted by the outplacement consultancy Michelin itself mandated and reported by the prefecture as about 55% of 947, fifteen months after the announcement

Alongside it the prefecture reported €403,000 of validated aid and €6.83 million committed. About half is neither a scandal nor a success: it is the yield a worker facing a closure should expect, and it is a count, not a projection.

Sources 7 Les Echos · 8 actu.fr · 9 Le Journal des Entreprises

II. What the independent evidence says about retraining

There is a dated experimental answer to the question of whether retraining changes employer behaviour. A study run under the aegis of DARES and published in January 2025 as Rapport d'études no. 63 tested employer demand for candidates in career transition and found that long, certifying training increases employer interest, including in shortage occupations (Céreq document repository, January 2025; summarised by Mon Compte Formation on 9 May 2025 and by Centre Inffo on 21 May 2025).

The operational reading is sharp. Short awareness courses do not move a recruiter. Long certifying programmes do. Any redeployment offer should therefore be judged on a single question: does it end in a recognised certification, or in an attendance record?

III. What Michelin recruits at Ladoux in 2026

Michelin's own recruitment site states that Ladoux gathers half of the group's R&D teams and that 75% of Michelin tyres are developed there (Michelin recruitment site, consulted 2 September 2026). Dated 2026 vacancies published on that site cluster in a recognisable direction: theoretical chemistry and numerical simulation, finite-element simulation, scientific computing, and applied artificial intelligence in industrial operations, one such alternance vacancy in industrial AI data science was published on 22 July 2026. Ladoux headcount is not published, and we do not estimate it; nor do we assert the presence of specific research programmes at the site that we could not confirm from a dated source.

That recruitment pattern matters more than any slogan about the future of mobility. A materials company that develops tyres by simulation needs people who can model a polymer and a contact patch, not only people who can mix and test one. For a young engineer, the entry ticket at Ladoux in 2026 reads: continuum mechanics, numerical methods, physical chemistry, and code you can defend.

A profitable company reshaping its workforce

What is publishedValue as printedDatePublisher
Workforce adaptation project based on volunteering, no forced departuresUp to 1,500 posts over three years28 May 2026Michelin
French workforce, as stated by the groupAbout 17,000 people28 May 2026franceinfo
Plants still operated in France1329 May 2026La Montagne
Segment operating income for 2025, at constant exchange rates€2.9 billion11 February 2026Michelin
Free cash flow before acquisitions for 2025€2.1 billion11 February 2026Michelin

Sources 15 Michelin · 16 Radio France (franceinfo) · 17 Michelin · 18 La Montagne

IV. The context the candidate should not ignore

On 28 May 2026 Michelin announced a workforce adaptation project based on volunteering, with no forced departures, covering up to 1,500 posts over three years: reported as weighted about two-thirds towards office functions, with the group stating a French workforce of about 17,000 (Michelin and franceinfo, 28 May 2026). No execution data (acceptances, actual departures, redeployment against exit) was available in the public record as of 2 September 2026, and we do not invent any. La Montagne reported on 29 May 2026 that Michelin still operates 13 plants in France.

The financial backdrop is not distress. Michelin reported segment operating income of €2.9 billion for 2025 at constant exchange rates and free cash flow before acquisitions of €2.1 billion (Michelin, 11 February 2026). A profitable company reshaping its workforce is a different negotiation from a failing one, and candidates and works councils should read it that way.

V. The 2030 materials target, and why it drives hiring

Michelin's stated target is 40% renewable or recycled materials in its tyres by 2030 and 100% by 2050 (Michelin, Media Day 2021). We could not extract the latest reported actual share from the group's most recent registration document in this pass, and we therefore publish the target without a progress figure rather than guessing one. The target alone is enough to explain the recruitment pattern: substituting a material means requalifying a product, and requalification is a simulation, testing and characterisation workload before it is a marketing claim.

VI. What to do with this

Three conclusions follow from the dated record. First, treat any employer promise of accompaniment as a claim to be measured, and ask for the placement rate at twelve and fifteen months, because that number exists. Second, choose long certifying training over short familiarisation, because the DARES-commissioned experiment shows that is what shifts employer demand. Third, if you are aiming at Ladoux, aim at the competences its 2026 vacancies actually name (simulation, computational chemistry and industrial AI), rather than the competences the brand evokes.

Methodology and limits

Built only from dated public sources listed with this article. Placement outcomes come from Michelin's own one-year review of 4 November 2025 and from press reporting of the monitoring committees between November 2025 and February 2026; the two Cholet site totals of 927 and 947 are published side by side without reconciliation. The retraining-effectiveness finding comes from the DARES-commissioned study published in January 2025 and its two published summaries. Recruitment evidence comes from Michelin's own recruitment site and a vacancy dated 22 July 2026. Ladoux headcount, the internal-transfer against retraining against external-placement breakdown, execution data for the voluntary-departure plan announced 28 May 2026, and the latest actual share of renewable and recycled materials are stated as unverified rather than estimated. Corrections are recorded on this page.

The argument

The spine of this dossier

Every part below rests on citations this dossier carries, including the part that says what would prove us wrong.

  1. The question

    A closure promised tailored support for every employee and named no rate. Which number then measures the promise for a worker deciding whether to wait or to retrain?

    One decision a named reader faces, phrased as a question. A topic cannot be answered, so it cannot be wrong.

  2. The tension

    The commitment and the count come from different hands. The group announced the closures on 5 November 2024 with individual accompaniment and no numeric placement target. The only counted outcome was produced by the monitoring committee and the consultancy the company itself mandated: 522 employees with a professional solution, reported on 16 February 2026 as about 55% of 947, on a site whose total is also published as 927.

    Two things in the published record that do not fit, each citing a document this dossier carries.

    Citations [1] [8]

  3. The mechanism

    The asymmetry follows from the wording of the commitment. Because the promise names a method and not a rate, nothing in it can be missed, and the group published no breakdown of internal transfer against retraining against external placement even in its own one-year review. The measurable figures therefore come from the prefecture and the outplacement consultancy, which is why the promise is judged against a number its author never set.

    Why the tension exists: the causal step, stated so it can be disputed on its own terms.

    Citations [3] [7]

  4. What follows

    For a worker facing a closure the planning figure is roughly one in two after fifteen months, with a mandated consultancy and a mobilised prefecture in the room. What changes that odds ratio is documented: the study published in January 2025 found that long certifying training raises employer interest in candidates in transition. So a redeployment offer is worth what its certificate is worth, and a candidate should accept the long programme over the familiarisation course.

    What follows for the reader: which skill gets priced, which comparison stops working, which offer is not what it looks like.

    Citations [10] [11]

  5. What would prove us wrong

    This reading fails the day the employer publishes its own arithmetic. A breakdown of internal transfer, retraining and external placement in a group release, or a placement rate stated on one reconciled Cholet total instead of the two now printed, 927 and 947, would replace our count with the company’s and settle whether accompaniment delivered more than the committees recorded.

    The observation that would show this argument is wrong, and the source that would carry it.

    Citations [3] [4]

  6. Our position

    The industrial desk’s reading: we adopt 522 of 947 as the counted outcome, we print 927 and 947 side by side and recompute no rate on either, and we read the up to 1,500 posts announced on 28 May 2026 as a ceiling on a voluntary project rather than as departures, because no execution data is published.

    What this desk concludes, signed and dated, resting only on the citations named above.

    Citations [8] [15]

Signed by industrial-desk, 2026-09-19

Every argument in the corpus

Shape the future of career exploration

Want a simulation for this role or industry?

We're building a library of job simulations shaped by the roles readers ask for. Tell us what you'd like to practice, and we'll work with our partners to create it, completely free.

Request a Simulation

Topics tracked in this dossier