There has been a rash of articles this year pertaining to how the automotive industry has begun testing humanoid robots. The subtext is that they may soon be used to replace actual humans on the assembly line. But are these bipedal robots actually ready to take over or are they simply being leveraged to prime tech-focused investments and keep employees on their toes?
Automation is nothing new. Decades after standardized tooling and implementation of the assembly line created a surge in manufacturing jobs, workers grew concerned that robots would eventually come to replace them. While there is still some debate about whether or not widespread automation actually resulted in fewer jobs overall, it certainly has required less human hands for repetitive tasks. We likewise see a gradual decline in the number of assembly jobs since the automotive sector became the first industry to both develop and implement robotic arms at factories (e.g. General Motors’ “Unimate”).
With more part suppliers than ever before, the industry has managed to retain and even grow jobs in certain regions of the world. However, the general trend has been to supplant workers with robots wherever possible. Businesses are perpetually seeking ways of reducing overhead and automation has always been something accounting departments are willing to explore.
Industrial robots often have a high initial costs, typically need to be built into the facility, and may also require expensive maintenance vs human employees. But they’re frequently cheaper to operate at scale when you just need an easily repeatable task done several million times. They also make for a handy boogeyman to help discourage workers from asking for more money or added benefits. Jobs requiring higher levels of physical dexterity or sustained mental processing have long been assumed to be safe from automation. But that has started to change, too. As the hype around artificial intelligence and humanoid robotics has grown, so have the fears that they would displace more living, breathing employees.
Automotive News recently published an article outlining how automakers and parts suppliers are growing increasingly concerned over a growing shortage of workers and rising labor costs. But the industry has been engaging in sweeping layoffs over the past several years, making the former claim rather suspect. In fact, large swaths of the automotive sector have long been eager to pare down the number of employees to further pad profit margins.
Modern solutions have included pivoting to all-electric vehicles, which allows companies to outsource more work to suppliers while streamlining the final assembly process, and increased levels of automation. Practically every automaker is already heavily reliant on automation and believes that the further implementation of robots will be necessary to remain competitive. Leveraging robots with an anatomy that resembles ours is still a novel concept, albeit one the industry appears eager to investigate.
German manufacturers are presently testing humanoid robots on a limited basis to see if they might work for regular assembly jobs held by real people. The assumed advantage is that, unlike large stationary robots, humanoid variants could be implemented without needing to redesign entire factories or clear up floor space. They’d just slide in wherever people used to be, potentially saving truckloads of retooling money in the process.
In Changchun, China, Audi and its Chinese partner FAW are reportedly leveraging around 800 robots to inspect all-electric vehicles. The facility uses an automated transport system that shuttles parts from warehouses or partially assembled vehicles around the factory without any human input. While there are still real people on site, the facility is supposed to be heavily integrated and automated using a singular IT architecture and cloud-based data system to handle all information throughout the plant. According to Automotive Logistics, the site is now testing humanoid robots to see how they might be adapted for use on the assembly line.
“Technology helps amplify the synergies between the teams from Germany and China,” Helmut Stettner, CEO of Audi-FAW, said late in 2024. “It also enables us to be fast and highly efficient and speed up processes — both during construction and in day-to day-operation. We create an ideal synergy between our highly skilled workforce and cutting-edge technology at the site.”
Earlier that same year, Automotive Logistics noted that Dieter Braun, head of supply chain at Audi, suggested that the transition to battery electric vehicles necessitated a flexible approach to so-called sustainable manufacturing methods. This was allegedly supposed to include “embracing new technologies such as automation and applying those technologies quickly, measuring the results and using them to stay competitive.”
BMW began trialing the “Figure 02” humanoid robots at its American plant in Spartanburg, South Carolina, last year. It claimed to have learned some valuable lessons pertaining to their capabilities and what would need to be done if they were to become a permanent fixture in the manufacturing process.
Mercedes has similarly been testing humanoid robots at its Kecskemét plant in Hungary. Most of those are the Apollo series bots from Apptronik, with formal trials starting in March of 2025. The German brand has stated that implementing robots has been part of its ethos since the 1970s and that will be “increasing the use of AI and advanced robotics” in order to remain a pioneer in advanced manufacturing.
But there’s a sense that aspects of the technology are being over-hyped, which shouldn’t be all-that surprising if you’ve been following the industry over the years. Automakers love to over-promise on the latest technologies.
Most of the tasks these robots are being given could be carried out by a motorized cart with clamps or are already being done by other robots. While we’ve seen in-house demonstrations from the robotics companies showing off their units engaging in high-dexterity tasks or running through obstacle courses, the automakers don’t seem to be asking for much more than having them hold things.
That could simply be down to the robots not yet having been programmed to engage in tasks related to automotive production. But it could likewise mean that the technology hasn’t reached a point where the robots could actually rival human workers. There is a sense that some of what the public is being shown exists primarily to drive investment and draw attention to the relevant businesses. A lot of this technology has likewise been around longer than some companies would care to admit and it’s sometimes difficult to determine just how far it has come.
Honda retired its adorable ASIMO robot in 2022. But the company had already moved away from claims that it was supposed to be any more than a marketing endeavor as it worked on more practical robotic designs. ASIMO was poorly suited for industrial applications or home use and effectively became a mascot for the Honda brand after proving that bipedal robots were technically feasible 25 years ago.
Since then, Honda has actually revised its development strategy to pivot away from bipedal helper robots with a diverse range of capabilities and toward units that are purpose built to perform a singular job. The company even attributes this change to the lessons it learned from ASIMO, noting that the unit was invaluable in teaching Honda how to design robots that can share space with humans. It’s refreshing to see a company be this candid about its robotics program. But it’s also incredibly rare, as you’re about to see.
Tesla’s Optimus robot seems to be taking the generalized approach. Rather than focusing on how to make the units work on the assembly line, Tesla is framing them as an all-purpose robot that will eventually be sold to the public for under $20,000. While there are impressive demonstrations and viral videos featuring the robots engaging in numerous tasks, many have remained skeptical about claims made by CEO Elon Musk asserting that the current generation can “basically do whatever you want” — especially after it was revealed that the units at the “We, Robot” event were being remote controlled by humans.
This isn’t something that’s been limited to the Tesla robots either. Many of the world’s premiere robotics companies release videos that have been heavily edited or showcase the very best take of a robot performing a pre-programmed task in order to make the units look more impressive than they actually are.
Boston Dynamics has been viewed as the de facto leader in the field for decades. However, the R&D company has changed hands repeatedly without its new owner ever seeming to take full advantage of the tech it’s developing. Google (Alphabet) bought Boston Dynamics in 2013, then it was purchased by SoftBank Group in June of 2017. Hyundai purchased a majority stake in the firm late in 2020, with the intent of putting its robotics expertise to work for the auto brand in terms of manufacturing, logistics, and automation.
Since then, Toyota has also partnered with General Dynamics to develop general-purpose robots akin to what Honda originally envisioned with ASMIO. But they’ll have to confront the same obstacles as everyone else vying to compete on the same landscape.
Battery power has been a major issue. The hardware required to make modern robots move like humans is extremely heavy and energy intensive. In most cases, even the most advanced units can only run for a couple of hours on battery power alone. That undermines their usefulness, especially if the goal was to supplant human laborers and serve as an alternative to stationary robots that are already commonplace on assembly lines.
The issue is further complicated when you consider that modern humanoid robots often cannot lift as much weight as the humans they’re trying to replace. This is actually the direct result of designing them to replicate our anatomy. Were they built with a lower range of motion and equipped with locking wheels (or stationary), they’d be stronger and more energy efficient. But that would be far less impressive when it came time to show off the technology and source investment capital.
Humans have long believed that the most-advanced automatons would mirror their creators. That concept has only been reinforced by representations of robots in media, which has been highly influential in terms of the shape technology takes in the real world. We’ve spent years teaching robots how to walk and even talk in convincing ways, perhaps at the expense of other functions.
Robots ultimately need to be told what to do, whereas humans have a natural curiosity and can automatically adapt to their environment. Coding a machine to handle a nearly infinite number of variables is a monumental task, even with help from today’s large language models.
The most plausible scenario seems to be that robotics companies (whether established or novel) are using their humanoid robots primarily to draw attention. Then, that attention can be used to sell their proven, non-humanoid robots or convince investors to funnel money into development programs. Automakers presumably know this and likewise engage in partnerships with those brands to show themselves as likewise embracing cutting edge technology. So, while you may still end up losing your assembly job to a robot in the next couple of years, it probably won’t be to one that has two arms, two legs, and a head.
[Images: Mercedes-Benz; General Motors; BMW; UBITECH; Honda; Tesla; Hyundai; Toyota]
Become a TTAC insider. Get the latest news, features, TTAC takes, and everything else that gets to the truth about cars first by subscribing to our newsletter.
