# Cariad Job Cuts: The AI Shift Depends on Reskilling

Author: Marin Niehues, Niehues Consulting · Published: 2026-10-07 · Organisation Design
Source: https://www.niehuesconsulting.de/en/blog/cariad-restructuring-ai-roles/

The Cariad job cuts shrink Volkswagen's software subsidiary far more than they equip it with new AI positions: officially around 1,000 of roughly 4,300 positions in Germany are to go and the IG Metall union and the works council expect up to 1,600, while only around 200 positions for AI specialists are being created. Because a job security agreement rules out compulsory redundancies until the end of 2029, Cariad has to deliver its announced shift to artificial intelligence largely with the roughly 2,900 people who stay, so reskilling and role design will decide whether it succeeds. By our own calculation the net reduction saves around 180 million euros a year, which is less than a fifth of the target of lowering development costs by around one billion euros a year.

**Key facts at a glance**
- On 30 September 2026 Cariad's management board announced further job cuts at a staff meeting. On 1 October 2026 dpa reported around 1,000 of roughly 4,300 positions in Germany.
- IG Metall and the works council expect up to 1,600 positions to go according to an Automobilwoche report of 6 October 2026, including around 400 from an earlier programme.
- Cariad intends to align its work with the broad use of AI, creating around 200 positions for AI specialists and cutting 200 others in return.
- Development costs are to fall by around one billion euros a year on average. According to Handelsblatt investment will fall by around 6.1 billion euros between 2027 and 2031.
- The job security agreement rules out compulsory redundancies until the end of 2029. The framework and pay agreements with IG Metall have been terminated.

**Cariad is shrinking by up to 1,400 positions net and adding only 200 new AI positions, although all of its work is to be aligned with AI. Because redundancies are ruled out until the end of 2029, the remaining AI capability has to come from reskilling, so a company in this position first calculates its workforce balance per role and then links every new role to a person already on the payroll.**

## Cariad is officially cutting 1,000 jobs while IG Metall and the works council expect 1,600

Cariad's management board announced further job cuts to employees at a staff meeting on Wednesday, 30 September 2026. Handelsblatt reported first. The German Press Agency dpa confirmed the figure of around 1,000 positions from company sources, as [Autohaus](https://www.autohaus.de/nachrichten/autohersteller/vw-software-schmiede-cariad-mit-weiterem-stellenabbau-3911283) reported on 1 October 2026. That is almost a quarter of the roughly 4,300 people Cariad employs in Germany, down from 6,500 at the end of 2023. According to Handelsblatt, Cariad is to cut investment by around 6.1 billion euros between 2027 and 2031. The timing and instruments of the cuts are still open, because negotiations with employee representatives have yet to take place.

[Automobilwoche](https://www.echo24.de/welt/cariad-drastischer-umbau-betrifft-1600-mitarbeiter-in-deutschland-zr-94527314.html) reported on 6 October 2026 that Cariad intends to align its work "consistently with the broad use of artificial intelligence". Around 200 positions for AI specialists are to be created, while a further 200 are cut elsewhere. IG Metall and the works council fear that up to 1,600 positions will go, because around 400 employees from an earlier programme are also still due to leave. Rivian now supplies the basic architecture of the vehicle software for Western markets and an alliance with Xpeng does so for China, so Cariad is concentrating on automated driving, infotainment, cloud systems and chassis software. In a statement to [mobiFlip](https://www.mobiflip.de/?p=491848) Cariad describes its goal as a "new, smaller tech organisation geared to speed". According to dpa, works council chair Gerhard Retzer called the termination of collective agreements and the job cuts no plan for the future. IG Metall announced determined resistance.

**The restructuring in numbers**
- 4,300 | people work for Cariad in Germany, against 6,500 at the end of 2023
- 1,000 | positions are to go under the new programme
- 1,600 | positions are expected to go according to IG Metall and the works council, including an earlier programme
- 200 | new positions for AI specialists, with 200 others cut in return
- €855m | operating loss in the first half of 2026

## Cariad loses up to 1,400 positions net while the shift to AI creates 200 new ones

Cariad shrinks by up to 1,400 positions net, because only 200 new AI positions offset the 1,600 expected departures. Automobilwoche names three items on the departure side: the new programme with around 1,000 positions, around 400 employees from an earlier programme and 200 positions cut in return for the AI positions. Together they add up to exactly the union's 1,600, Automobilwoche names the three items but does not add them up itself.

::: diagramm Cariad loses up to 1,400 positions net while the new AI positions offset one eighth of the departures
- New reduction programme | -1000 | −1,000
- Remainder of the earlier programme | -400 | −400
- Cut in exchange for AI positions | -200 | −200
- !New positions for AI specialists | 200 | +200
- !Net change | -1400 | −1,400
Source: Automobilwoche via echo24, 6 October 2026; dpa via Autohaus, 1 October 2026; own compilation of the three items named there, sum calculated by us; base: 4,300 positions in Germany

Of the 4,300 positions around 2,900 would then remain, which is 45 percent of the workforce at the end of 2023. The 200 new AI positions would equal around 7 percent of this future workforce, so the other 93 percent have to change their tasks with AI if Cariad wants to align all of its work with AI. According to dpa, Cariad handled earlier rounds of cuts through severance and early retirement programmes. Such programmes select departures by the willingness of employees, so system knowledge can be lost exactly where the integration of partner components needs it.

## The job cuts cover only a small part of Cariad's savings target of one billion euros

By our own calculation the net reduction of 1,400 positions saves around 180 million euros a year, because even the entire workforce in Germany costs far less than one billion euros. The calculation assumes fully loaded costs of 130,000 euros per position and year, because Cariad does not publish a figure. At 100,000 euros the saving would be 140 million euros, at 160,000 euros it would be 224 million euros. According to Automobilwoche personnel costs are the main cost item by a wide margin, although the reports do not give the number of employees outside Germany.

::: diagramm By our own calculation the net reduction saves around 180 million euros a year, less than a fifth of the savings target
- Investment cut 2027 to 2031, per year | 1220 | €1,220m
- Savings target for development costs per year | 1000 | €1,000m
- Personnel costs of all 4,300 positions | 559 | €559m
- !Personnel costs of the net reduction of 1,400 positions | 182 | €182m
Source: Automobilwoche via echo24, 6 October 2026; Handelsblatt via dpa and Autohaus, 1 October 2026; own calculation, assuming fully loaded costs of €130,000 per position and year; €6.1bn spread evenly over five years

The investment cut equals around 1.2 billion euros a year over five years and is therefore of the same order as the savings target, although the reports leave open whether both figures describe the same programme. The larger part of the savings therefore has to come from how the work is divided up: less in-house development, more partner components, lower investment and renegotiated collective terms. How to lower costs without losing the ability to deliver is described in [Efficiency programmes: cutting costs while protecting performance](/en/blog/efficiency-programmes-cut-costs-protect-performance/).

## Cariad's core capabilities are moving from in-house development to integrating partners

With the partner architectures from Rivian and Xpeng, Cariad's core capabilities shift from programming complete platforms towards architecture, interfaces, validation and partner management. How much AI changes the individual tasks depends on the task and on experience. Erik Brynjolfsson, Danielle Li and Lindsey Raymond measured in the customer service unit of a software company that an AI assistant raised the number of issues resolved per hour by 15 percent on average, with less experienced staff benefiting above all ([Quarterly Journal of Economics, 2025](https://doi.org/10.1093/qje/qjae044)). In a randomised study published in July 2025 the research organisation [METR](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/) found that experienced open-source developers took 19 percent longer with AI tools. In [February 2026](https://metr.org/blog/2026-02-24-uplift-update/) METR reported signs of a speed-up with later tools but considers the new data an unreliable signal because of selection effects. The benefit of AI therefore has to be measured at the level of the individual task before it enters workforce planning, as [AI in software development](/en/blog/ai-in-software-development-productivity/) shows.

**In-house development and integration of partner components**
Criterion | In-house development | Integration of partner components
Core capability | programming complete platforms | architecture, interfaces, testing and validation
Bottleneck | capacity of in-house development | management of partners and quality of acceptance
Typical roles | development teams per component | integration teams, platform teams, partner management
Role of AI | tool for programming | tool for testing, review and documentation as well as a product field in itself
Typical risk | delays from building everything in-house | dependence on partners and loss of system knowledge

Whether the 200 AI specialists work as a shared platform team or as part of the four field teams depends on how similar the AI applications of the fields are. Which roles agents change in development workflows is covered in [Agentic AI and the organisation](/en/blog/agentic-ai-roles-organisation/).

## Under job security, AI roles at Cariad are filled mainly through reskilling

Under a job security agreement a company fills new AI roles largely from its own workforce, because it can hardly acquire skills by replacing staff. In its [Future of Jobs Report 2025](https://www.weforum.org/publications/the-future-of-jobs-report-2025/) the World Economic Forum expects around 39 percent of workers' existing skills to be transformed or become outdated by 2030. Reskilling works under cost pressure when it is tied to specific target roles, for example from software development to integration and testing or from project control to partner management. Management plans external hiring only for the remaining gap, as described in [Strategic workforce planning for three years](/en/blog/strategic-workforce-planning-three-year/) and [Building AI capabilities in the organisation](/en/blog/building-ai-skills-in-the-organisation/).

The works council is involved, because in Germany a restructuring of this size is usually an operational change under section 111 of the Works Constitution Act, for which management negotiates a reconciliation of interests and agrees a social plan. Under section 92a the works council can make proposals to safeguard employment, explicitly including training. Where tasks change so that employees' skills are no longer sufficient, it has a right of co-determination under section 97(2) when in-house vocational training is introduced. This overview does not constitute legal advice. Training therefore belongs in the negotiation from the start, whose process is described in [Operational change in reorganisations](/en/blog/operational-change-social-plan-germany/).

## Executive teams should calculate the workforce balance first and fill AI roles second

Executive teams in mid-sized companies that combine a shift to AI with cost cutting start with a workforce balance per role and measure the new roles after twelve months:

1. Calculate the workforce balance: within four weeks management, HR and controlling set out departures, turnover, new hires and reskilling per role family for three years. Success means a balance that shows which part of the cost target comes from personnel and which from the way work is divided up.
2. Measure the benefit of AI per task: each unit head tests AI in two to three tasks over eight to twelve weeks. Success is measured by time per case and error rate before and after.
3. Match target roles internally: within six weeks HR and unit heads describe the tasks and training path of each new role and name people from the existing workforce. Success is the share of roles filled internally.
4. Agree training with the works council: from the second month management includes training and voluntary programmes in the negotiation on the reconciliation of interests and the social plan. Success means a signed agreement and the participation rate.
5. Review after twelve months: management records the fill rate, delivery performance and costs of the new roles. Success is the gap to plan, as described in [Measuring reorganisation success after twelve months](/en/blog/reorganisation-success-after-twelve-months/).

## Niehues Consulting combines organisation design and AI management

[Niehues Consulting](/en/) supports executive teams and heads of department in building organisations that deliver reliably. Organisation design creates structures, roles and decision paths for day-to-day business and for transformation. In delivery management, a delivery manager or product owner steers the programme until the change has arrived in operations. Where AI enters the workflows, a business case and a prototype are added, followed by introduction in the business units.

The approach follows the Niehues Delivery Model with its phases Understand, Build and Introduce. Before a shift to AI it asks three questions. In Understand: which tasks does AI change and who will do them in future? In Build: which roles, teams and interfaces do operations require? In Introduce: are employees trained and workflows adapted? Founder Marin Niehues, a management consultant from the delivery side, builds organisations that deliver. His projects range from venture-backed scale-ups to DAX 40 groups, with assignments in more than 20 countries on six continents. In the free 30-minute initial consultation you assess together which roles in your organisation change through AI and how to prepare your workforce for the new roles: [book a consultation](/en/#kontakt).

**What executives can take from the Cariad case**
Cariad is cutting up to 1,400 positions net and creating 200 new AI positions, while the job security agreement rules out compulsory redundancies until the end of 2029. Because by our own calculation the net reduction saves only around 180 million euros a year, the larger part of the savings target has to come from less in-house development, partner components and investment cuts. The AI capability will come mainly from reskilling the roughly 2,900 employees who stay, so a similar restructuring starts with the workforce balance per role.

**Sources**
- [Automobilwoche via echo24: Cariad: Drastischer Umbau betrifft 1600 Mitarbeiter in Deutschland](https://www.echo24.de/welt/cariad-drastischer-umbau-betrifft-1600-mitarbeiter-in-deutschland-zr-94527314.html), 6 October 2026
- [dpa via Autohaus: VW-Softwaresparte: Cariad mit weiterem Stellenabbau](https://www.autohaus.de/nachrichten/autohersteller/vw-software-schmiede-cariad-mit-weiterem-stellenabbau-3911283), 1 October 2026, citing Handelsblatt
- [dpa via Volksstimme: VW-Softwaretochter steht vor weiterem Stellenabbau](https://www.volksstimme.de/deutschland-und-welt/wirtschaft/kreise-vw-softwaretochter-steht-vor-weiterem-stellenabbau-4329705), 1 October 2026, with statements by works council chair Gerhard Retzer
- [mobiFlip: Volkswagen stellt sich bei Software (mal wieder) neu auf](https://www.mobiflip.de/?p=491848), 6 October 2026, with a statement from Cariad
- [Erik Brynjolfsson, Danielle Li, Lindsey R. Raymond: Generative AI at Work](https://doi.org/10.1093/qje/qjae044), Quarterly Journal of Economics 140 (2), 2025
- [METR: Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/), July 2025, with an [update](https://metr.org/blog/2026-02-24-uplift-update/) from February 2026
- [World Economic Forum: The Future of Jobs Report 2025](https://www.weforum.org/publications/the-future-of-jobs-report-2025/), January 2025
- [German Works Constitution Act (BetrVG), sections 92a, 97 and 111](https://www.gesetze-im-internet.de/betrvg/)
- Own calculation: workforce balance and personnel costs with the assumptions stated in the exhibits

::: faq Frequently asked questions

### When did the Cariad job cuts become public?

Cariad's management board announced the job cuts at a staff meeting on 30 September 2026. Handelsblatt reported first and dpa confirmed the figure on 1 October 2026. On 6 October 2026 Automobilwoche reported on the shift to AI, the 200 new AI positions and the expectation of IG Metall and the works council.

### How many jobs is Cariad cutting?

Under the new programme around 1,000 of roughly 4,300 positions in Germany are to go. IG Metall and the works council expect up to 1,600, because around 400 employees from an earlier programme are still due to leave and 200 positions are cut in exchange for AI positions. After deducting the 200 new AI positions this gives a net reduction of up to 1,400 positions.

### Can Cariad make compulsory redundancies?

According to the works council and IG Metall, the job security agreement still applies and rules out compulsory redundancies until the end of 2029. The cuts therefore run through voluntary programmes and negotiations. This overview does not constitute legal advice.

### Are the job cuts enough for the one-billion-euro savings target?

By our own calculation with fully loaded costs of 130,000 euros per position, the net reduction of 1,400 positions saves around 180 million euros a year. Even the entire workforce in Germany costs around 560 million euros under this assumption, so the larger part of the savings target has to come from investment, partner components and collective terms.

### What can companies in Germany learn from the Cariad case?

Companies in Germany can learn from the Cariad case to calculate a workforce balance per role before a shift to AI and to measure the benefit of AI per task. They fill new roles from their own workforce wherever possible, because a job security agreement limits replacing staff.

### When does reskilling pay off compared with a new hire?

Reskilling pays off when training for a target role costs less and takes effect faster than hiring externally. It also pays off when employees bring system knowledge that new staff would first have to build up. In integration roles this knowledge of existing systems is often the bottleneck.