Humanoid Robots Are Coming to Work. Here’s What You Need to Know Now.

Last year, investors put $4.3 billion into humanoid robot companies, according to Bank of America — a six-fold increase from 2018. This year, those robots are showing up for work.

Agility Robotics’ Digit has reportedly moved more than 100,000 totes at a GXO Logistics facility in Georgia and recently signed a commercial Robots-as-a-Service agreement with Toyota Motor Manufacturing Canada. Figure 02 completed an 11-month pilot deployment at BMW. Apptronik’s Apollo is under commercial agreement with Mercedes-Benz. This isn’t just a technology story anymore. It’s a business story — and if you run a company of any size, you should be paying attention.

From demo to deployment

Separate the hype from the reality. Humanoid robots are not yet walking into your local coffee shop or staffing a retail counter. What they are doing — through pilots, field tests and commercial agreements with major industrial partners — is handling specific, repetitive physical tasks that don’t require human judgment.

Figure 02 performed production-support tasks at BMW, contributing to the production of more than 30,000 vehicles at BMW’s Spartanburg plant and logging over 1,250 hours positioning sheet-metal parts for welding. These are no longer just stage demos; they are early pilots and field tests in real industrial settings.

The tasks are, for now, relatively focused: moving totes and containers, transferring components between workstations, loading and unloading machines. That’s because the industry wants to establish reliability and safety before expanding the scope. Agility’s Digit has moved over 100,000 totes at GXO Logistics and is now under a commercial agreement at Toyota’s Canadian assembly plant. Bank of America projects 90,000 units shipped globally in 2026, climbing to 1.2 million by 2030.

The technology making it possible

A humanoid robot is only as capable as its ability to perceive and navigate the physical world. That requires AI models, sensor fusion, compute and training data working in concert — but at the foundation is accurate spatial perception. Without the ability to see in three dimensions in real time, nothing else works.

That’s where 3D vision technology comes in. Orbbec, which has been building depth-sensing cameras since 2013 and reported 66% revenue growth in 2025, is one of the suppliers contributing to this stack, providing the spatial perception hardware that gives robots the ability to locate, recognize and respond to objects in the physical world.

The company’s engineers have written about where stereo vision technology is headed, and the trajectory is steep. At CES 2026, Orbbec launched a camera engineered specifically for humanoid robot wrists, and global technology brand HONOR debuted its first humanoid robot at MWC 2026 with Orbbec’s stereo vision cameras providing its spatial perception.

Perception is just one piece of a more complex stack. Humanoid robots also depend on advanced actuators — the motors and joints that translate AI decisions into physical movement — as well as onboard compute, battery technology and the AI models that tie it all together.

Swiss manufacturer Maxon supplies precision actuators used in Hexagon’s AEON humanoid platform. Google DeepMind has demonstrated its Gemini Robotics models on Apptronik’s Apollo platform and announced a partnership with Boston Dynamics at CES 2026 to integrate the same foundation models into Atlas. And NVIDIA’s Isaac GR00T platform supports synthetic data generation and accelerated robot training, dramatically compressing the time it takes to teach a robot new tasks.

The convergence of all these components is what’s making the current generation of humanoid robots viable in ways that earlier generations simply weren’t.

Some analysts now see factory use becoming economically viable across a broader range of applications in the 2025-2028 window, a timeline that would have seemed optimistic just two years ago.

What changes, and for whom

The first wave lands in industries with high labor intensity and predictable physical tasks: logistics, automotive, light manufacturing. Companies deploying humanoid robots in these environments are betting that labor costs will drop in ways that reshape pricing power and margin structures. Whether those bets pay off at scale remains to be seen, but the direction of the investment is clear.

For businesses further from the front lines of automation, the effects are less direct but still real. If physical labor becomes cheaper and more scalable across the economy, the pressure migrates. A logistics company that automates warehouse operations doesn’t just cut costs — it changes what it can charge, how fast it can scale, and what its competitors have to do to keep up.

Businesses built around doing things that robots will eventually do well need to be honest about that timeline. Businesses built around judgment, relationships, creativity and context — things that remain genuinely hard to automate — are better positioned than they might think.

The $4.3 billion invested in humanoid robotics last year wasn’t a bet on science fiction. It was a bet that the cost of physical labor is about to change in ways that ripple through every industry that touches it, which is most of them. The only real question is whether you’ll be ready when it reaches yours.