Last updated: August 5, 2026
By Tom Kenaley, Co-Founder and President, KORE1
BMW is cutting roughly 8,000 jobs by the end of 2027, about 5 percent of its 154,540-person workforce, through a voluntary severance program aimed at office and development roles in Germany. Production is excluded. No compulsory redundancies. The program does not even open until October 2026, and that single scheduling detail is why most of what you have read about this event is wrong for your purposes.
I run the engineering desk conversations at KORE1 and we get paid when a company hires someone we found, so read the rest of this knowing there is an invoice at the end of a good outcome. Two of the sections below argue against a hire. One of them argues you should stop trying to recruit the exact people this headline is about. That costs us money and it stays in, because the alternative is watching a client burn a quarter chasing a bench that does not exist on this continent.
KORE1 has placed engineering talent since 2005 through our engineering staffing agency practice, and the automotive side of that work has been strange for about eighteen months now. Reqs open. Reqs close. The candidates moving between them are almost never the ones in the news. Our broader 2026 tech layoffs roundup tracks this event alongside thirty-odd others, and the closest structural cousin in that set is Tesla, where the headline number also had almost nothing to do with the bench that eventually reached the market.

What BMW Actually Agreed To
On July 29, alongside first-half results, BMW confirmed an agreement with its works council on what the company has called the largest voluntary redundancy program in its history. Six weeks of negotiation produced it. The shape matters more than the headline number. Read the shape.
Severance offers go out from October 2026 through December 2027. They go to roughly 40,000 of BMW’s 85,000 permanent German employees, which is to say almost half the German workforce gets an envelope, and BMW expects about 8,000 of them to take it. Around one billion euros has been set aside to fund those packages. A billion out, a billion back. The target is a matching one billion in annual savings starting in 2028. Administration, research and development, planning, and corporate functions are in scope. Plants are not.
The financial backdrop is not subtle. Per the BMW Group’s own H1 2026 results release, group revenues fell 8 percent to 62.27 billion euros, pre-tax earnings dropped 29.4 percent to 4.05 billion, and the second-quarter automotive EBIT margin came in at 2.3 percent. Free cash flow for the half fell 45 percent. Milan Nedeljković, who took over as chairman of the board of management from Oliver Zipse in May, framed it as reshaping the organization to stay competitive. Fair enough. A 2.3 percent automotive margin at a company that has historically run four times that is not a communications problem.
Munich moved last on this, incidentally. Wolfsburg and Zuffenhausen got there first, and the supplier tier beat all three of them by a year.
| Company | Scale of cuts | What is actually being reduced | Runs through |
|---|---|---|---|
| BMW Group | ~8,000 (5% of 154,540) | Administration, R&D, planning, corporate. Production excluded. | Dec 2027 |
| Volkswagen Group | ~50,000 in Germany | All areas. 35,000 at the VW brand alone, plus Audi, Porsche, and the Cariad software unit. | 2030 |
| Porsche AG | 5,000 more, ~9,000 cumulative | Zuffenhausen and the Weissach development center. R&D complexity reduction is explicit. | 2035 |
| Bosch (supplier) | ~22,000 announced in total | Mobility division, chasing a 2.5 billion euro cost gap across five German sites. | 2030 |
| ZF Friedrichshafen (supplier) | 7,600 in one division, up to 14,000 overall | Electrified drivetrain. EV product development halted in favor of plug-in hybrids. | 2030 |
Sources: electrive on the BMW program, electrive on the VW Group target, electrive on Porsche, and Automotive Manufacturing Solutions on the supplier tier, which puts total German automotive job losses at roughly 55,000 since 2023.
Office and Development. Not the Line.
Every one of those programs protects the plant and goes after the desk. That is the pattern, and it is the only part of this story with a clean read-across to American hiring.
The org chart that built a car in 2016 is not the org chart that builds one in 2026. The German version of that chart ran on enormous vehicle-program organizations. Hundreds of engineers per program, coordinating mechanical packaging, supplier validation, homologation, calibration, and a release cadence measured in model years. A software-defined vehicle does not need that org chart. It needs a much smaller number of people who own platform code that ships continuously, plus the functional-safety and cybersecurity specialists who keep that code legal, plus the high-voltage and power-electronics people who make the expensive part of the car work. What it does not need is the coordination layer in between. Porsche said the quiet part out loud when CEO Michael Leiters framed the deeper Weissach cuts around reducing complexity.
So the 8,000 is not really a downturn number. It is a re-composition number. Ask what it recomposes into. The answer is on the next page of every one of these announcements, and almost nobody reads that far.
BMW’s Research and Innovation Centre in Munich, the FIZ, holds something like 25,000 engineers, developers, designers, and specialists. Regensburg, Dingolfing, and Leipzig carry more. Those are the buildings the severance offers land in. Now look at what the same company did five weeks earlier. It closed out a $1.7 billion investment in South Carolina. The iX5 starts coming off the Spartanburg line before December. Six US-assembled electric models by 2030. And BMW puts its American footprint at more than 120,000 supported jobs and $43.3 billion contributed annually.
Cutting in Munich. Expanding in Spartanburg. Both true, same quarter, same company, and if you only read one of those two press cycles you will draw exactly the wrong conclusion about where automotive engineering demand is headed.

The Bench You Are Picturing Is Not Yours
This next section is the one that loses us business.
If you clicked this expecting a roster of newly available BMW engineers to poach, there isn’t one. Not this year. Possibly not next year either. The mechanics of a German voluntary program make it close to the least accessible talent event in the industry, and the reasons stack.
Nobody has left yet. Offers start going out in October 2026. Acceptance windows, notice periods, and negotiated exit dates push real separations well into 2027, and BMW has explicitly said there will be no compulsory redundancies, which removes the forcing function that normally pushes people onto the market fast.
The people who do take it are being paid to take it. A billion euros spread across the acceptances, sized on salary and tenure, in a country where a senior engineer with twenty years in can walk out with a package that funds a long, unhurried search. That is the opposite of the urgency you need on the other side of an offer.
Then there is the part American hiring managers skip. These are German nationals in Munich, Regensburg, Dingolfing, and Leipzig with no US work authorization, no plan to acquire any, and in most cases family, mortgage, and school arrangements built around a metro area they have lived in their entire adult life. Relocating one to Michigan is a nine-to-fourteen-month project involving immigration counsel, a real relocation budget, and a spouse who also has a career. We have run it. Five times since 2023. It works maybe one time in five, and the four failures cost you the req.
That one-in-five ratio is from memory, not a spreadsheet, so treat it as directional. The direction is right.
Where this genuinely does become your problem, and your opportunity, is one tier down the supply chain. Bosch, ZF, and Continental all run substantial American engineering operations. Plymouth and Northville in Michigan. Farmington Hills. Auburn Hills. Anderson and Charleston in South Carolina. When a German parent takes 22,000 positions out of a mobility division globally, US technical centers are inside that math, and those people are on American payroll, with American work authorization, forty minutes from your office. Northville, not Munich. Nobody writes headlines about a technical center in Northville. That is precisely why the talent there is reachable.
Where the Reachable Talent Actually Is
The real American automotive engineering bench got built over the last four years and almost nobody outside Detroit noticed, because it happened in slow increments rather than one Saturday-night memo. No memo. No tracker entry.
The Detroit Three have collectively cut more than 20,000 US salaried jobs from their recent employment peaks, roughly 19 percent of their combined white-collar workforces, according to CNBC’s analysis of company filings and employment data. General Motors accounts for about 11,000 of that, dropping from 58,000 US white-collar workers in 2022. Ford came down roughly 5,300 from its 2020 peak to about 30,700. Stellantis went from 15,000 salaried people to around 11,000. Combined, the three peaked near 102,000 salaried jobs in 2022 and closed last year at 88,700.
Now put the other number next to it. Bureau of Labor Statistics data cited in the same reporting shows motor vehicle manufacturing employment overall fell just 0.2 percent across that stretch, to 285,800 workers.
Read those two facts together and the whole shape of the market falls out. The line held. The office did not. Roughly thirteen thousand experienced, US-authorized, automotive-fluent salaried professionals came off the Detroit Three payroll in three years while total industry employment barely moved, which means they did not get absorbed by a shrinking industry. They scattered into adjacent ones. Some went to Toyota, which grew its American white-collar headcount about 31 percent from 2020 to 2025, to roughly 47,500. Some went to suppliers. A lot went to aerospace, industrial automation, and medical devices, where nobody was competing for them.
That is your bench. It has been sitting there the whole time, in Detroit, Dearborn, Warren, Auburn Hills, and increasingly in Greenville and Spartanburg, Chattanooga, and Huntsville as the southeastern manufacturing corridor absorbs the migration. Our post on reshoring and US manufacturing jobs covers that geographic shift in more detail.
Four Skills Getting Repriced Right Now
Not all automotive engineering is moving in the same direction. Some of it is genuinely contracting and some of it is scarcer than it has been in a decade, and a recruiter who files all of it under “automotive engineer” will send you the wrong person twice before you correct them.
Functional Safety and Automotive Cybersecurity
Scarcest thing on the list, and it is not close.
ISO 26262 functional safety at ASIL C and D, ISO/SAE 21434 cybersecurity engineering, UNECE R155 and R156 compliance work. These people sit at the intersection of engineering and regulatory, they are the reason a software update can legally reach a vehicle, and there were never very many of them. German programs are cutting around them rather than through them. Every US EV and ADAS program wants one. Nobody has two. We have had a functional safety req open for a client in the Midwest for eleven weeks and the constraint is not compensation. It is arithmetic.
High-Voltage Systems and Power Electronics
Battery pack architecture, thermal management, inverter and DC fast-charging design, BMS. Demand for this in Germany is genuinely softening because ZF just halted EV drivetrain development and redirected toward plug-in hybrids, and Porsche has pushed its electrification returns out by years.
Here it inverted. BMW’s new battery assembly plant in Woodruff, South Carolina, needs staffing. So does the Spartanburg expansion. So do the plants around Chattanooga and across the Southeast. The same skill is deflating on one continent and inflating on the other, which is the single most useful arbitrage in this entire story if you are hiring in the US.
Embedded and Software-Defined Vehicle Engineering
AUTOSAR Classic and Adaptive, C and C++ against real-time constraints, CAN and Automotive Ethernet, over-the-air update infrastructure, SOME/IP service architecture.
Strong demand, and a caution I would give any hiring manager. A traditional automotive embedded engineer and a software-defined vehicle platform engineer are not the same hire, and the resume language barely distinguishes them. One has shipped calibrations against a model-year deadline. The other has shipped code to a fleet on a two-week cadence with a rollback plan. Ask which one they have done. The answer takes thirty seconds and saves you a month. We run these searches through the embedded systems engineering desk for exactly this reason.
Vehicle Program and Product Engineering Generalists
The profile actually being reduced. Program managers, vehicle integration engineers, component owners, the coordination layer that made the old model work.
These are not weak engineers. Several of the best hires I have been involved with came from this background, because someone who has held a vehicle program together understands cross-functional dependency in a way that no amount of individual contribution teaches. But the roles they held are not being backfilled anywhere, in Germany or here, and the honest career advice is lateral into manufacturing engineering, supply chain engineering, or technical program management outside automotive. If you are hiring in those functions, this is an underpriced pool right now and it will not stay underpriced.

What This Talent Costs in the US
Public salary data on automotive engineering is a mess, and it is worth understanding why before you use any of it to build a band.
ZipRecruiter put the national average for an automotive embedded software engineer at $153,383 in early 2026. Glassdoor says $165,615 for the same title. Comparably says $80,833. The spread is the whole story here. The high figure is more than double the low one, on a single job title, and the reason is that the aggregators are averaging Tier 2 supplier calibration roles in Ohio against OEM platform software roles in Silicon Valley and reporting a single number as though the two markets talk to each other. They don’t.
For the wider frame, the Bureau of Labor Statistics reports a median annual wage of $102,320 for mechanical engineers as of May 2024, with 9 percent projected employment growth from 2024 to 2034 and about 18,100 openings a year. Faster than average. Which is a genuinely odd thing to read in the same week as an 8,000-person European cut, and both numbers are correct.
The bands below come off our own desk.
| Profile | US base range | Direction of demand | Typical time to close |
|---|---|---|---|
| Functional safety / automotive cybersecurity | $165K to $215K | Sharply up. Supply is the binding constraint. | 8 to 14 weeks |
| High-voltage systems and power electronics | $145K to $195K | Up in the US, down in Germany. | 5 to 8 weeks |
| SDV platform / embedded software | $150K to $205K | Up, with a wide quality distribution. | 4 to 7 weeks |
| Traditional automotive embedded / calibration | $115K to $150K | Flat. Deep pool, easy fill. | 3 to 5 weeks |
| Vehicle program / product engineering | $125K to $170K | Down. Underpriced against adjacent functions. | 2 to 4 weeks |
Base salary only, excluding bonus and equity. Ranges reflect KORE1 placement and offer data across US markets in 2026, cross-checked against the public aggregator figures cited above. Detroit, Silicon Valley, and Seattle run 10 to 18 percent higher. The southeastern manufacturing corridor runs below. For a band on one specific title in one specific metro, our salary benchmark assistant will generate one.
If You Have a Req Open Against This
Stop writing “OEM experience required.” It is the most common mistake on automotive job descriptions right now and it is filtering out the exact supplier and technical-center engineers who are both available and, frequently, deeper on the specific subsystem you care about. Describe the system instead. High-voltage battery architecture. Pack thermal management. ISO 26262 ASIL D decomposition. OTA update pipelines on production fleets. Those phrases pull the right inbound and cost you nothing to write.
Now the interview loop. If your req is functional safety or automotive cybersecurity, get it under three weeks or accept that you will lose every finalist. Eight to fourteen weeks is the time to close on that profile and a large chunk of it is candidate scarcity, not process. Adding your own five-week interview cycle on top is how a req turns into a quarter.
One more, and it has a shelf life. The vehicle program and product engineering pool is genuinely underpriced today because the market has decided those roles are going away, and the market is half right. The roles are going away. The people are not. Those are two separate facts, and the transferable skill into manufacturing engineering and technical program management is real. In eighteen months this will be common knowledge and the discount will be gone. Today it is still sitting there.
Last thing, on structure. A good share of these searches close faster as contract or contract-to-hire than as permanent roles, because engineers coming off a corporate restructuring are more risk-tolerant about a six-month engagement than the conventional wisdom suggests, and it lets both sides test the fit. The ones that should be permanent from day one are the safety and cybersecurity roles, where a direct hire is the only offer that competes.
KORE1 has placed technology and engineering talent since 2005, across more than 30 US metros, with recruiters who average over 15 years on the desk. Our 2026 placements are running 17 days to hire on average, and 92 percent are still in seat at twelve months. Automotive and mobility work runs through our automotive engineering staffing practice, with plant-side and process roles handled by the manufacturing engineering desk. If you want someone to pressure-test a req against this market before it goes live, talk to a recruiter and we will tell you which of the four profiles above you are actually describing. It is two of them more often than you would think, and that is usually why the req has been open since March.
The Questions We Are Getting on This
Is BMW laying people off, or is this a buyout?
Structurally it is a buyout, not a layoff. BMW negotiated a voluntary severance program with its works council and ruled out compulsory redundancies. Around 40,000 German employees receive offers between October 2026 and December 2027, and roughly 8,000 are expected to accept.
The distinction is not semantics. A voluntary program self-selects for people who were already close to retirement, already planning a move, or holding enough tenure to make the math attractive. It does not produce the involuntary, urgent, immediately-available candidate flow that a mass layoff does. Different event, different talent consequence.
Can I hire one of these BMW engineers?
Not realistically, no. They are German nationals working in Germany with no US work authorization, and they will not begin separating until 2027. Severance sized on twenty-plus years of tenure removes any urgency to relocate across an ocean.
The version of this that does work is narrow. A specific, senior, hard-to-replicate specialization, an employer with existing immigration infrastructure, and a nine-to-fourteen-month timeline you have actually budgeted for. If that describes your situation, it is worth a conversation. If you need someone in Q4, look at Michigan and South Carolina instead.
Does any of this touch BMW’s American operations?
BMW’s US operations are expanding, not contracting. The company completed a $1.7 billion investment in Plant Spartanburg and the new battery facility in Woodruff, South Carolina, starts building the iX5 there in late 2026, and plans at least six US-assembled electric models by 2030.
BMW’s US engineering footprint sits mainly in South Carolina and at the BMW Group Technology Office in Mountain View, California, which opened in 1998 as the company’s first R&D site outside Munich. None of that is inside the severance program. The company reports its US operations support more than 120,000 jobs, and it just told the market that South Carolina is central to its global manufacturing plan. Those are expansion signals sitting inside a contraction news cycle.
Where is the automotive engineering talent that I can actually reach?
Detroit, mostly, plus the German supplier technical centers on US soil. The Detroit Three have cut over 20,000 US salaried roles from their peaks, about 19 percent of their combined white-collar workforces, while total motor vehicle manufacturing employment fell only 0.2 percent.
Add Bosch, ZF, and Continental’s American engineering operations in Plymouth, Northville, Farmington Hills, Auburn Hills, and the Carolinas, which sit inside global reduction programs their parent companies announced in Germany. Same skills. US work authorization. A commute instead of a visa. That pool has been quietly available for three years and it is nowhere near picked over.
Is automotive engineering a shrinking career now?
Automotive engineering is not shrinking as a category, but its composition is changing fast. BLS projects 9 percent employment growth for mechanical engineers from 2024 to 2034 with roughly 18,100 annual openings, faster than the average across all occupations.
What is shrinking is a specific mid-layer. Vehicle program coordination, component ownership, and calibration-heavy roles built around a model-year cadence are being compressed by software-defined vehicle architecture, and that compression is what all these European announcements are actually funding. Safety, cybersecurity, high-voltage, and platform software are all growing through the same period. Anyone telling you the whole field is contracting is reading one column of a two-column table.
We need embedded automotive talent inside six weeks. Realistic?
For traditional embedded and calibration work, three to five weeks is realistic. Software-defined vehicle platform engineers run four to seven, and that one moves with your comp band. Functional safety at ASIL D will not happen in six weeks at any price.
The variable nobody controls for is interview length. A four-stage loop with a take-home adds two to three weeks and knocks out the candidates holding competing offers, which on the scarce profiles is all of them. If six weeks is a hard constraint, cut a stage before you raise the number.
Does KORE1 work automotive, or is this just commentary?
We work both sides of it, filling automotive, mobility, and manufacturing engineering reqs on contract, contract-to-hire, and direct hire while representing displaced engineers at no cost to the candidate. The hiring company carries the fee.
Engineers reading this from inside a supplier technical center or a Detroit salaried org should send a note through the contact form with the subsystem you own and what shipped on your watch. Program name, volume, and what broke are more useful to us than a skills list. We route to the right desk inside a day.

