On Sunday July 5th, Atlas®, Boston Dynamics humanoid robot, stepped onto the field during halftime of the FIFA World Cup 2026™ Norway-Brazil round of 16 match at New York New Jersey Stadium. In the first public appearance of our production Atlas—and the first ever robotic halftime activation of the FIFA World Cup™—the robot presented the ceremonial match ball to the referee to start the second half of the game and showed off its range with a series of goal celebrations inspired by some of the world’s most celebrated footballers as part of Hyundai Motor Company’s ‘Next Starts Now’ campaign.

For a live performance on a stage as big as the FIFA World Cup™, there’s nowhere to hide. Atlas had to be able to move dynamically and perform reliably. We trained these behaviors in our lab with a rigorous Reinforcement Learning pipeline for whole body control and tested them both in the lab and in the real world. Coordinating a demonstration like this activation requires Atlas to have the precision and adaptability to perform reliably in a bright, noisy, chaotic stadium—the same skills that drive success in a busy, unpredictable factory.

We sat down with Seth Davis, senior program manager on the Atlas team, to learn more about how this demonstration came together and what it takes to succeed in the field (or on the pitch).

What does a senior program manager do?

I work with all of the stakeholders, both the technical teams within Atlas, as well as leadership and other teams across the company, to help coordinate getting Atlas out of the lab and into the real world. A typical day for me is a lot of plate spinning for all our Atlas projects, making sure that every one of those gets the attention it needs, that the right things are happening at the right time, that people have the information that they need to do their job well.

I’ve been with Boston Dynamics almost 15 years and historically my focus has been on allowing the engineers to focus on the work they need to do, to give them the support, the infrastructure, the tools, the team they need to make the robot perform in the real world.

What goes into a deployment like the one at the FIFA World Cup™?

There’s so much work. There’s obviously a lot of technical work to make the robot able to do these kinds of unique and novel behaviors, where we’re manipulating a ball and walking on challenging terrain like a soccer turf. There’s all these logistics and little bits and pieces that go into executing an event and getting one of our robots from the lab out into the real world. We need to plan for different contingencies, ensure the communications will work, and test under a range of conditions to make sure the demonstration will go off as planned on game day and that the robot will do something exciting that no one’s seen before.

There’s also a lot of work of coordinating with the different stakeholders, the Hyundai team, the FIFA team, and understanding what each team needs to be successful. With our robots in particular, but also robotics in general, people ask for what they think the robots are and not always the thing that they actually want. When we’re trying to plan out demos, people will ask for the moon. If I can understand the story they’re trying to tell or the problem they want to solve, we can usually find a way to accomplish it, even if we can’t accomplish the specific thing that the person asked me for.

How does a public demonstration like this one relate to Atlas’ eventual work with customers?

I think it really speaks to how versatile both the robot can be and the team we’ve built here at Boston Dynamics. It was really important with Atlas to build a robot that could kind of do anything that people can do to actually make it valuable for a customer in manufacturing or warehousing or other industries. The fact that we can take that robot and use it to do soccer moves in a show like this is really exciting. 

We have a team that can pivot from getting really deep into the technical details of how to make a robot walk, how to make a robot pick something up and manipulate it really accurately to making the robot do things like kicking a soccer ball, carrying a soccer ball, interacting with people at these kinds of events. The robot’s performance isn’t just precise in an engineering way, but also communicates something about what that robot is capable of and what that robot means to the future.

What lessons have we learned from going through that deployment process with Spot and Stretch that we’re carrying forward into the Atlas program?

I started at Boston Dynamics when our quadruped robots were first getting out of the lab and going into the field with LS3 and then the Spot® program. I spent a lot of time building up the Field Engineering team that actually deploys Spot to our customers’ sites and helps the customer get the robot integrated into their facility and doing useful work. Now we’re at the beginning of that journey again with Atlas.

One of the biggest lessons is that the robot is just one part of a whole ecosystem that’s necessary to actually provide value to customers. A walking robot is the exciting and really visible part of what we’re doing. But we have to deliver more than that.

When a customer buys a Spot, what they’re really investing in is inspections. They want to understand the health of their plant. The robot helps do that, but customers also need specialized sensor payloads and systems like Orbit™, our orchestration and intelligence software, to manage all that inspection data and oversee Spot’s work. 

That’s becoming more and more important with Atlas as we go from the prototype stage, where we’re just making the robot do interesting things to the production stage, where it needs to be an actual, useful member of a factory team.

The robot is really cool, but it needs to talk to the warehouse management system. It needs to communicate to supervisors about what work it should be doing and what the status of that work is. If it makes a mistake, it needs to be able to communicate that. So the robot is really just one part of a larger system. At Boston Dynamics, we’ve gotten really good at building that whole system around the robot.

How do we test our robots to ensure they’ll perform in the real world?

We learned pretty early that we had to get robots out of the lab, so we didn’t get stuck in the mode of what works in a controlled environment. For our commercial applications that usually means testing at customer sites or similar facilities, like our first Atlas deployment at the Hyundai Metaplant last year. But for something like the FIFA World Cup™, we took Atlas to a soccer field at a local city park.

It’s actually not the first time we’ve run tests at that particular park. When we were working on LS3, we would do monthly tests in Prospect Hill Park, which was just up the hill from our old lab. Going back and forth with the city of Waltham, they decided that the easiest way to manage us using the park was to give us picnic permits. Every few weeks I would be at the Recreation Department office requesting a picnic permit for the Valley Shelter area. They knew exactly what we were doing, but we were just having a picnic with a box truck full of robot parts and a 900-pound four legged robot walking on the trails.

My original contact actually still works for the city. I talked to her just a few weeks ago for the permits to use the soccer fields to test Atlas.

What is your favorite thing about working at Boston Dynamics?

It’s been really exciting to be part of how the company has grown from, “Can we make a robot walk?” to “Can we make a robot that’s useful to the point that a customer wants to buy it?” 

Working with Atlas now is really just the next version of that. It’s really exciting to be part of both being part of activations like the FIFA World Cup™ with Atlas, but also in how we bring these robots into Hyundai’s factories and make them do useful work

Humanoid robots are so new. At Boston Dynamics we have the opportunity to be part of building the future.

Do you have any advice for aspiring roboticists?

One of the things that I’ve noticed about the people who are really successful at Boston Dynamics is they are experts in something that they’re really passionate about, but that’s not the only thing that they’re interested in. They’re excited about robotics, but they’re also they have some hobby that might have nothing to do with robotics. Having that sort of broad perspective can be really helpful when it comes to problem solving. You can bring a perspective that other people don’t have.

The parts of my job that I’ve always really enjoyed have been when I’m doing something that isn’t really understood and there’s a lot just figuring it out—How do we do this? How do we make this work? How do we get from where we are to where we need to be?

There’s always been a ton of that at Boston Dynamics. We know what we want to build, and we have to figure out how to build it to make these incredible things happen.