Jabil discusses the pace of humanoid robot development and production

The Apollo humanoid robot is designed for a variety of applications. | Credit: Apptronik

While much of the focus on humanoid robots has been on entertaining demonstration videos and massive amounts of funds raised, the work of development and finding a value proposition is quietly continuing. How are humanoids progressing from prototypes to mass production? Jabil Inc., a design, supply chain, and manufacturing services provider, offered its perspective.

Companies preparing to ramp up production of humanoid robots include Tesla , 1X Technologies , XPeng , and UBTECH . Jabil works with robotics developers to bring their products to market at scale.

The St. Petersburg, Fla.-based company has worked with Apptronik on bringing its Apollo humanoid to active manufacturing environments. Thomas Brown , senior director for engineering and technology, including digital commerce and robotics, at Jabil, replied to the following questions from The Robot Report .

Setting realistic expectations for humanoid volume

How are investor and customer expectations for mobile manipulators and humanoids evolving?

Brown: In terms of managing expectations, a lot of customers have gotten humanoids out in low volumes, but we are entering a phase of tens of thousands, where it’s no longer something that’s assembled by hand on site. You’re starting to get to where you need a cost structure, manufacturability, and to improve quality. These are areas where Jabil can really add a lot of value.

When a robot’s much more expensive, there’s only so many they can produce for gathering AI training data, and they need more customer use-case validation. We provide that ability to cost down their designs to actually hit those scales.

How is the current boom in compute enabling more capable physical AI ?

Brown: Compute is now sufficient for many of the experiences you see on the demo reels, but we do need more computing capacity at the edge. The NVIDIA Jetson series is a popular lineup for our customers developing their aspirations in the robotic space.

We’re getting more compute and memory, but prices are an ongoing pain point for scale. The foundation model makers are developing intelligence that’s outside that capacity, but expanding it is key.

We’ve seen Hyundai , Tesla , and BMW fund and test humanoids. What advantages do automotive partners bring to design for manufacturing?

Brown: Take wiring harnesses for humanoids. It needs a hybrid of automotive-type robustness with an extremely small form factor like high-grade consumer electronics.

Developers need to work with partners for rapid iteration. In addition to sensing and compute, there’s a wide range of power supplies , allowing for swapping.

What are some of the challenges in getting humanoids to market?

Brown: Prior to the smartphone, things were much more simplistic — you could just build up capabilities to the cost point and scale. Now, companies are resetting and diversifying their supply chains.

My company is working with major humanoid developers in the U.S. and abroad, but I can’t name them at this time.

Editor’s note: Physical AI and humanoids are among the session track topics at RoboBusiness 2026, which will be on Oct. 20 and 21 in Santa Clara, Calif. Register now to attend.

Register now and help us celebrate 20 years of RoboBusiness!

Where humanoids, mobile manipulators make sense

What is the commercial outlook for humanoids in factories and warehouses ?

Brown: In manufacturing , there are a ton of spaces where humanoids make sense. When you look at a production line today, most automation is built for specific tasks, but the applicability of humanoids to a wide range of tasks changes the math from a purpose-built line.

A lot of these humanoid companies have done a good job of partnering with end customers to gather clear requirements. That’s a solid approach. In the consumer space, it’s still the science fiction route of “I just want a robot to fold laundry” or whatever the vague use case might be.

In terms of iterating, do humanoids need to equal human capabilities, or should developers be aiming for superhuman ability?

Brown: I think it’s more about having a super set of capabilities that you can apply in bits and pieces to other markets. You see it already with some of the humanoids, where they have a wheeled base instead of being bipedal. It’s a much easier problem to solve, and I would say that’s the same thing with having the fully embodied humanoid versus, say, one arm with the camera.

The company that conquers that can conquer all other challenges, but the company that conquers just one is going to struggle to go up the stack. You see many of these companies going after the larger play. Maybe some of them will completely hit it; maybe some of them won’t.

I would say it’s truly the super-set approach that they can decompose into specific solutions that are more cost-optimized, depending on the use case.

Do you see mobile manipulators as an interim step toward humanoids, or more of a standalone form factor?

Brown: I definitely think there’s a place for mobile arms. Even traditional automation like a machine tender running at 1.5 mps [3.3 mph] is pretty fast. A humanoid going that fast would be tremendously fast, let alone if it’s carrying a load.

There’s just different needs, depending on what the facility requires and what the existing line looks like. The humanoid is going to be a supremely flexible device, but it’s not always the right device. Traditional high-speed automation will still have a big play for the foreseeable future.

You had mentioned consumer robots, when and if do you think we can expect them?

Brown: Maybe in five-ish years. It could be 20 years, but progress is going very rapidly. If the model development continues at the pace it has and if the computing continues at the pace it has, we’re certainly going to be there on the AI and electronics hardware side.

There’s going to be a market for multiple…

News is gathered automatically from public robotics & AI feeds on a schedule.