BMW Group to deploy humanoid robots in production in Germany for the first time
press.bmwgroup.com116 points by JeanKage 10 hours ago
116 points by JeanKage 10 hours ago
Whenever I hear german companies mention digitalisation, I get reminded that they still use pen and pencil in production environments to log data, pass those sheets to secreteries who enter the data into legacy systems so data analysts can enter it into another system that then has an integration with SAP. Data from SAP then flows onwards to some buzzword filled Azure product that costs a few million a month from which someone downloads an xls file and uploads it to Tableau where they run some simple calculations. Someone else downloads it as an xls and manually writes (not copy pastes) the numbers into a power point presentation and makes graphs by drawing shapes. This is then presented at some bi-monthly meeting.
I wish I was making this stuff up.
I used to work for the US side of a German multinational (one of the largest in the world) and discovered the same thing when it came to software.
The German side always had slick presentations (they always had good visual marketing) and impressive claims, but whenever I tried to work with their products, I always found the claims overstated and that they hadn't really executed deeply. This despite my German counterparts working hard (I visited HQ in Germany and when they work, they really work and clock the hours, no idle chitchat)... yet it doesn't translate to impact.
A lot of their products had impressive front-ends but half-baked back-ends (on the American side, it's the reverse -- our interfaces looked like crap, but our stuff actually worked and often delivered in less time).
A lot of their designs were also non-human friendly (if you've ever driven a German car, you'll realize that the car was built for engineers and not for end users -- weird little user-hostile features pop up everywhere). I don't understand why this is -- this is a nation that produced Dieter Rams. Tobi Lutke (CEO Shopify) likes to talk about how Germans grew up surrounded by good design, yet that design culture never permeated many German products. I own a Bosch in-unit washer/dryer and it's frustratingly unintuitive and has a "my (the engineer's) way or the highway" philosophy.
I went to a BMW talk once about the infotainment system (it was built on the latest Azure tech), but came away feeling that the work was not deep. It was skin deep.
I wonder what has happened to the German builder/tinkerer culture that made German manufacturing great. In the 1980s and 1990s, Germany was synonymous with excellence. But in the 2000s-present, not so much (except maybe in very narrow mittelstand verticals, e.g. Zeiss).
"if you've ever driven a German car, you'll realize that the car was built for engineers and not for end users "
cracks me up. I once leased a BMW for 3 years. By the time I returned the car, I still didn't what all the cryptic buttons for HVAC and other controls. They just refused to follow established automotive ergonomic conventions.
Anyway, my father used to do business with Germans for a long time. He had many interesting stories to share, but one that has always stayed with me is, his disdain for how cheap / penny pinching his German companies and their employees were when doing deals. This was in the 90s, so definitely passed the West Germany glory days.
My take is, in the era of global competition, Germans didn't know how to strike the right balance and effective allocate resources. Where to compromise, and where not to. I don't know if it's sheer stubbornness or they're just wired differently.
I wonder if anybody's manufacturing was great before the Japanese quality revolution. It took Germany longer than the US to adopt modern quality control. Granted, Germany did a lot of it, for instance their chemical industries were staggering.
I've formed the impression that every country's engineering and design cultures are essentially aesthetics.
BMW's most recent infotainment is a big step backwards to me in both areas - aesthetics and UX. Its previous generation was one of the best in the market also in both. I literally don't consider buying the newest 4-series just because of it, especially of the ultrawide driver's "monitor" - it's just so ugly, and I regret it, since on the outside the car appeals me so much more. I'd rather spend the same on previous year model with better specs.
agreed. IMHO, last good BMWs were E90 and F10s. That's when they still cared about craftsmanship even in areas not visible to customers.
> if you've ever driven a German car, you'll realize that the car was built for engineers and not for end users -- weird little user-hostile features pop up everywhere
I drive a German car, all be it one from 2015, I don’t recognise this statement at all.
I find it quite well designed.
This sounds more like over generalised FUD to me.
I think Germany got infected by the shareholder value and privatization virus but doesn't really understand it.
Whats funny is when its privatized by publicly traded companies it becomes this weird nationalized-kinda but not really thing that turns the economy into a bifercated class of first class citizens and second class citizens.
This describes large companies everywhere
I encountered oil wells essentially controlled by post-it notes passed around an office.
Maybe they found the PostIt notes worked better than whatever software they tried.
I distinctly remember reading that the entire bill of materials for building a Boeing 747 was managed in Excel. I have not been able to find that claim since then but it was so amazing to me that I remember it.
It doesn't really make sense as I think about it now, because the 747 design predates Excel by many years so maybe it was BS.
Assume that later models of the 747 were managed in Excel. No trolling: What is wrong with that? You can write insanely powerful software using (1) formulas on the sheet (which are essentially functional programming) and (2) imperative logic in VBA (HTTP calls, database calls, file system, etc.). For years, I used this model and wrote pretty powerful software. Sometimes, I miss it for the encapsulated system. These days, in "biz dev" (internal software), it seems like the Excel model was replaced with an Electron front-end (HTML/CSS/JS) with Java back-end.
There's nothing particularly wrong with it, Excel works great. But almost any complex Excel file is riddled with obscure bugs and the nature of the tool makes it impractical to apply some of the most effective quality control techniques. Like you can't easily do code reviews or write an automated test suite.
I second this! Excel is a front end everybody knows and everybody can run. I always got laughed at when I say the biggest competitor of small apps (things like gym diary, meal planner etc.) is excel. Now that it even support python…
I've seen worse. For 2 years I received the results weekly, that I didn't ask for, of a $1m a year burn reporting stack. This was launched during a massive back patting ceremony like something out of Severance.
So one day I stared at it randomly and noticed that the pie chart percentages on one thing didn't even add up to 100. Looked back at history and it turned out this had been the case since day one. Spent a day taking it to bits and a good 50% of it made no sense at all and people had been making business decisions on it without checking it.
And to remediate it? They replaced it with some AI generated slop which is even worse.
It's always funny when HN users comment that there are no more opportunities for startups and it's too hard to compete against large, wealthy corporations. The reality is that most of them are so badly managed that competing against them is easy if you're actually competent.
> The reality is that most of them are so badly managed that competing against them is easy if you're actually competent.
The world is a graveyard littered with startups that thought this way. One of the consequences of wealth concentration and monopolies is that it is insufficient to be better than your competitors because your customers are also incompetent. To find product-market fit you not only have to be better, you have to be noticed by someone that cares that you're better and upon reflection confirms you solve a valuable problem.
By way of analogy, it's not enough to realize that MouseCorp makes shitty mousetraps and the local village spends $1M/yr on them. You can make a better mousetrap thinking its worth $1M/yr, or do the deeper look and realize the local village doesn't have a mouse problem but rather has a problem with too many feral cats, and has no interest in buying better mousetraps and once their attention is gained simply stops buying mousetraps altogether. Both parties lacked competence, but that didn't mean there was a market.
Right, so don't waste time trying to sell low-margin products to local governments. As the saying goes: it's like trying to shear a pig, too much squealing and not enough wool.
> The world is a graveyard littered with startups that thought this way. One of the consequences of wealth concentration and monopolies is that it is insufficient to be better than your competitors because your customers are also incompetent.
It's less that and more that governments and bureaucrats are corrupted to create barriers tot the market and to turn a blind eye to anti competitive behavior and outright illegal practices. For example huge banking corporations have been caught laundering money for drug cartels and got away with fines -- if your fintech startup tried that on, you would never see the outside of a prison cell.
I have seen this way to often in other areas. That is the push here as well AI can sort through it. Too many people are held to account for not meeting what amounts to made up numbers.
They make connecting SAP so difficult... this is the only way
It's not how it works. You suppose to contract a consulting company that contracts some offshore company to connect you to SAP.
And the customer being cheap doesn't pay for the proper modules and thus everything gets mapped to PSP elements -- to keep the same old garbage piles that get pushed around.
I wonder if it’s cheaper to just have an AI write the parts of SAP you actually need.
if the AI is a certified SAP consultant, sure. But then it would probably cost you $20K/month in subscription.
That might actually describe a pretty good implementation of an interface to SAP.
I think pencil is more efficient than SAP.
I agree and it's quite resilient to digital outages/downtime (at least in terms of hours, probably not more than a few days) so your manufacturing productivity won't drop to zero when the ERP system goes down. The paper logs can also be entered later when the system comes back up.
As we've seen in the Iran conflict, datacenters are a target and result in extended outages.
> I wish I was making this stuff up.
Lmao. Yes it's a pretty good summary of what happens in the corporate world, and not only in Germany.
SAP is truly terrible.
Well....
Before you say SAP is terrible, have you tried the competition?
It has some data entry screens that are super efficient, cursor always goes to the right place, tab moves you to the right place etc. 15 years and lots of ERPs later I have never seen better
When you are viewing a purchase invoice on the ledger you can see the PO it is matched to, click on it and it goes to the PO. Click a line on the PO you can see the GRNs related. Click on them you go to the actual GRNs
Oracle ERP can't do any of that.
A union in Germany is fighting Tesla over this same thing...
>. In 2026, Giga Berlin is the pilot site for the "Optimus" Gen-3 integration—humanoid robots performing repetitive tasks in the battery pack assembly area. IG Metall views this not as progress, but as a threat to job security.
https://www.teslaacessories.com/blogs/news/the-giga-berlin-s...
Not sure this counts as "humanoid" any more than the robots we've had in factories for a century... the hands and feet are nothing like a human's, and would not be improved by being more human.
It seems they just made the shape of their machine have a vaguely human silhouette so they could ride a hype wave.
I'm all for programmable humanoid robots, humans are an awesome human interface, but this ain't it.
That’s generally what it seemed like to me too. Seemed to be human shaped so they could say it’s humanoid… and nothing else.
Nothing in the video looked like it couldn’t be done by a more industrial robot shaped robot. And I bet that would be cheaper or easier to make.
Then I started reading the text. When I got to the part very early on about deploying “Physical AI” that confirmed it to me.
This all seems to be “humanoid washing”. Nothing terribly interesting that someone put a special coat of paint on to get attention.
I’d love to be proven wrong. But the video certainly didn’t show it. And I didn’t notice it in the press release, though it was hard to parse past the ridiculously over the top language that did nothing but obscure what was actually going on.
Probably because there’s not much going on.
I agree with your opening paragraph -- hot take! What would be more interesting, would be to see humanoid robots un/packing and moving boxes in their warehouses. That to me seems like one of the first logical places to deploy humanoid robots to replace (or assist) human workers.
Seems to be this European robotics company
https://robotics.hexagon.com/product/
https://blogs.nvidia.com/blog/hexagon-robotics-ai-software-a...
My prediction is that by the time humanoid robots actually make it to the factory floor, they'll be pretty un-humanoid.
90% of car manufacturing is done by oldschool industrial robots, and I've had people point out that heavy use of industrial robots are basically unique to the car industry.
You might see a robot arm here and there in other industries, but it's somewhat rare, usually its all purpose-built machines or humans.
Cars are in the intersection where you have large parts that need to be moved around where other things happen with them and large scale of standard designs that is worth automating.
I expect that when we manage to do automated construction, it will use robot arms in several places too. But it's very rare that those two happen at the same time. Usually, large things are not standard.
You reminded me of the hilarious SV pizza making robot startup which has its own robot arm.
In this video you see the unnecessary robot arm move the pizza to the oven: https://www.youtube.com/watch?v=bN45bTsBUW8
For contrast a How it is Made video of frozen pizzas being created at dozens (hundreds?) per minute: https://www.youtube.com/watch?v=0UrSIOtv8a0
The difference is they can be put into environments that do not allow for purpose robots. I don't think it is a bad idea that spaces are still made for humans for when something goes wrong. A frozen pizza factory is trying to solve and sell something different.
I agree and for me personally this is very easy to see and understand.
Why do you think the vast majority of people fail to see it like this? Guys like Musk obvious hype it up as he now has tied the valuation of the firms he owns and operates to this story.
I don't disagree with the general utility of humanoid (or other multipurpose) robots, just not in a factory setting.
I think automating stuff in the factory makes zero sense - its a controlled environment with purpose designed tooling where anything that makes sense to automate has been automated. All the extra work will only result in marginal gains.
It's automating the stuff that goes on outside of the factories - for example construction imo is about almost as labor intensive as it was a century ago, the marginal gains were offset by more complex building techniques and higher expectations.
Housing is also just about the most valuable thing that exists in every country.
Because so much infrastructure is in humanoid form. If you can make something that can manipulate two hands on arms that are positioned and moved like human arms, you could just put that torso into a lot of situations to replace a human without a lot of retooling. That's the dream I think.
Installing humanoid robots in a factory is like using regexes to parse data.
It makes sense if it's a one-off but there are better solutions.
Maybe it does make sense for small scale businesses that need just a little automation? Like a humanoid robot could restock shelves and do inventory in a grocery store at night, and you wouldn't need to retrofit anything to be able to do that.
Large scale factories seems like the wrong use case for humanoid robots.
I personally think humans make the mistake of thinking that we must create objects that emulate oneself. Imagination is tough, I know.
Does the computer 'memory' behave identically like human memory? Of course not. Does it look like the 'memory' of a human? Again, of course not.
Vertically upright humanoids have a lot going for them: they don't occupy a lot of floor space, they can pull an object right into their center of gravity to manipulate it, and because they're familiar they're relatively easy to prototype actions for because they're our actions.
People always asser without evidence that humanoid isn't the best design, but there's a paucity of alternatives that don't make some type of tradeoff: humanoid might not be the best at anything, but it's clearly very good at a lot of things.
I toured a major US OEM assembly plant recently, and there were a TON of humans working insanely repetitive tasks.
The most kind numbing of all were the easiest (sit in chair and put bolts upright in holder so robot can pick them up) and the highest paid thanks to union seniority.
The UAW will kill all the US OEMs before that let robots replace all the humans.
Hexagon is very prominent in precision manufacturing through their dimensional measurement robots (CMM Coordinate Measurement Machine) and other metrology software/hardware. This is most likely why they were chosen by BMW, as I imagine they already have a working relationship together, although the EU aspect could have contributed as well.
I wonder if this is a newly acquired subsidiary producing these robots (they've been doing a lot of acquisitions recently), or if these have been in development in-house for a while.
Let me first comment that this is just another publicity stunt and that there will be no useful humanoids in BMW factories in the near future. Then I will read TFA and get back here.
So, nothing to see here. From what I understand the robots are doing pick and place. In the video the robot also gets passed a part from a human worker. So nothing here makes any sense and the tasks could solved with conventional robotics. I guess it is good that the robot can be moved but for this you don't need a two-legged humanoid balancing.
It is a pilot project. German pilot projects rarely go anywhere. If this succeeds against all odds, I hope for BMW that the robots are buying cars, too.
It'll be the first time a BMW ever used turn signals!
Not if they trained on driver data. Will come with tailgating, lane swerving and flashing high beams as standard. Sonar will be used to judge the minimum distance you can ram behind someone and when to activate high beams.
Yeah. Feels kind of insignificant considering the amount of non-humanoid robots they've used on production lines for the last few decades and lack of any claims to be "fully autonomous" or for the humanoid robots to be performing particularly advanced tasks
Here is a 60 Minutes piece showing Boston Dynamics Atlas working in a car factory in the United States. https://www.youtube.com/watch?v=n6ISdRkS37I
Hyundai vs BMW, where is Tesla?
Tesla beat Hyundai and BMW to this meaningless announcement a year ago, and have already progressed from that to the inevitable “oh yeah, this doesn’t actually work yet.”
Give Hyundai and BMW time.
Posted elsewhere in this discussion...
>The Automation Factor Tesla’s response to labor pressure has always been more automation. In 2026, Giga Berlin is the pilot site for the "Optimus" Gen-3 integration—humanoid robots performing repetitive tasks in the battery pack assembly area.
https://www.teslaacessories.com/blogs/news/the-giga-berlin-s...
It's coming, next year, there will be a million of them.
On the moon or on Mars?
There are already 2M robots on Mars, Elon is working on a space mission to bring 1M back.
In the mountains of Iran and stepps of Ukraine, I fear… and before you downvote, consider that starlink are in both, already.
What advantages has a humanoid robot compared to automation for the tasks in a factory?
I mean: yeah, it's easier to think for a human how to operate a human. Is this the Pareto optimum? No low hanging fruits left?
Does anyone here with years of experience in robotics have a better explanation? (I hope so, because all the domain experts I met, always could tackle questions that pop into the mind of a layman without even breaking a sweat. Experts being experts, obviously)
That's excellent! I look forward to much cheaper cars now that the robots will be making them for the masses.
Oh, absolutely. Because history clearly shows that when multi-billion dollar corporations save money on labor, they immediately pass those savings directly to the consumer.
Give it time: at some point nobody will be consumer except for the equity lords. Savings will reach them.
Seems so funny to me that we are building llms to write in english code for computers. And building robots to perform some automated processes in the shape of humans.
When are we going to rip the bandaid off, and skip bothering with the ux layer built for humans? I guess that is just old fashioned 20th century factory style automation that doesn't get headlines written about it, at least not in these decades.
The humanoid-like robots are designed for existing prouction-lines that can't be easily shut-down and retooled/reconfigured.
Looks like they already have been testing it in the Spartanburg, SC, USA plant (just outside of Greenville SC [also I think the largest BMW factory in the world making most of their SUVs]). Still I don't get why a humanoid robot would be a thing for car making, a robot arm seems like it'd almost always be more efficient.
Auto manufacturers have been using robots for decades to paint, weld, machine, cast, and manufacture critical automobile components and even entire chassis.
Some auto robot porn: https://www.youtube.com/shorts/Zi8EWZptTfA
The idea of humanoid looking robots shuffling around a factory floor seems like a gimmick to me.
Meanwhile China has dark factories.
I think this is a myth - Chinese factories don't seem to be automated to a higher degree than European ones, and in any case are still full of Kuka, Fanuc and ABB robots.
I think there's a domestic brand or two that's gaining marketshare, but they're not there yet.
There's a myth of Chinese high-tech (esp in cars), that is not to say their stuff isn't technologically advanced, but the characterization that Chinese tech has left Europeans' behind just does not pass muster when one looks at mechanic videos of Chinese EVs.
Their cars look fancy and are full of futuristic screens and sensors, but the suspension setup and lot of engineering behind them is not exactly cutting edge.
That's why a lot of car reviewers say that a lot of their EVs don't drive particularly well
This take might have been true years ago, especially for cheaper makes, but modern EVs from china use high-quality components and designs that often surpass european automakers. The premium brands - Nio, Zeekr, Polestar, Lotus, etc - have design and R&D offices in Europe, and source parts from suppliers all over the world. Nio uses Nvidia Orin chips, Qualcomm SoCs, Brembo brakes, Bosch controllers, ZF suspension systems, Continental/Pirelli tires, and ClearMotion (in their flagship model); can't get any better than that.
The driving feel is definitely a thing, the chinese cars are very soft and 'boaty' which is not as desirable elsewhere. They are also on average much larger and heavier than their western counterparts, cities in China have road infrastructure built in the past 20-30 years with spacious lanes.
In a sense BMW has factories in China too (through Brilliance). I once heard the story that they built a 1:1 clone of the Dingolfing plant there.
The owner family did the right thing at the right time. If the Europe and US business tanks they will be fine. BMW as a brand not necessarily.
So their cars will get cheaper, right... right???
I think this is going to be bad for BMW, and bad for the current robotics-summer. I _hope_ that’s not the case, I’d love for robotics to get deployed more widely in manufacturing. But I’m pretty sure it will be. I think the chances of meaningful success would be higher with non-humanoid robotics
Robots are widely deployed in manufacturing, just not humanoid robots. The biggest benefit of stationary robots is that their behavior is 100% predictable and they are always where the humans expect them to be. There are certain areas in factories which can be nearly 100% automated (pick & pull in warehouses, for example), but there are a lot of areas where, without a human in the loop, there are too many edge cases to reasonably expect humanoid-robots-as-replacements-for-humans to anticipate or react to.
(I have 15 years experience in high tech manufacturing, with most of that building test automation & manufacturing execution systems, and an advanced degree in operations research.)
This is top-tier vagueposting.
Feel free to ask for more details if you have specific questions! I worked in robotics for many years, I have some decent familiarity with this space. Here’s some more detailed thoughts “for free”:
Humanoid robotics are largely a publicity stunt. Our actuators, sensors, and algorithms are better adapted to other form factors. The nice thing about humanoids is that you (in theory) don’t have to change the interface, since they can use the same interface humans can use. In practice that doesn’t hold well, because we don’t have great force/pressure sensors to cover large areas like human skin. Likewise, it’s difficult to apply the fine forces that are sometimes needed (grabbing an egg, moving a joystick, etc). And there’s risk of the robot doing something unpredictable, so you always have to set a good safety bound around it anyways. In the end it’s often better to adapt the process to modern robotics, rather than the other way around.
There are many good practitioners that write about these and other limitations, I think Rodney Brooks has some good discussion of it, eg. https://rodneybrooks.com/why-todays-humanoids-wont-learn-dex...
Apart from dexterity, bipedal machines are unstable and require dynamic adjustment to stay upright, as I understand it.
The mechanism humans use to stay upright after an unexpected loss of balance, flailing etc., would not be safe to be around when a robot employs them.
There's also the idea that a humanoid robot can learn to imitate human action just by watching it, thanks to AI magic!
It would be hard to believe that BMW doesn't have many industrial robots already deployed and in fact they do on a serious scale [1]. Now, to my mind, adding humanoid-robots to the existing mix of standard-robots and people seems not completely terrible.
The article's vacuous AI gloss language indeed makes it seem like they are indeed engaging in, crudely put, baloney. But your own language is weird here, like you don't realize robots are a standard thing in modern manufacturing. I mean, modern manufacturing "succeeds" massively using nonhumanoid robots at large scale.
[1] https://www.bmwgroup-werke.com/spartanburg/en/our-plant.html
This doesn't feel like it needs to be humanoid shaped. It does not appear ambulatory. Why not just tracked chassis with some robot arms. That said, humanoid robots with food tracks very anime.
These humanoid robots making cars is always a bit disappointing. If you look at videos of an assembly line the humans do manually dexterous complex tasks like bolting in car seats which are move in by hand, or helping place windshields into position.
The problem humanoid robots solve would be addressed better by simply altering the design of the car to not require humans to do that stuff.
What we actually see is humanoid robots deployed to do tasks that can already be done with simpler robot arms...
They will deploy robots , but their infotainment system is crap. Entire pricing model is to sell extra volume in engine for $10k on each measurable step , even though electric cars has a solved performance that is only limited by tires. Not to mention their gas cars are way more complex vs electric. Sure that will save bmw .
Not sure what the drawers are on the robot but one of the humanoid robots I saw changed its own battery that was pretty cool (I think it had 2).
Why doesn't anybody do the shoulder complex right? It gives me itches to scratch.
Looking forward to using one of those robots as a butler.
According to Figure, their robots had already been deployed in production
The robots featured in the embedded promotional video appear to be mostly useless. This is the opposite of impressive.
Will they dance? I've yet to see someone demo a humanoid robot doing something useful. Clearly, making them dance can't be that difficult.
And this is how it starts in EU
How they work? Without indication
They communicate through tailing each other and flashing bright lights from behind.
[dead]