What Humanoid Robots Really Mean for Founders Right Now

For most of the last decade, humanoid robots lived in viral videos. They danced, did backflips and occasionally fell over, and most founders reasonably filed them under “interesting, but not my problem.” That is starting to change. Humanoid robots are now clocking real shifts in real warehouses and factories, the companies building them are signing multi-year commercial contracts, and one of the biggest carmakers in the world has started early production of its own robot.

That does not mean a robot is about to walk into your office and take over the job of your operations manager. The truth is more nuanced and, for founders, more useful. Understanding where humanoid robots actually are today, what they can and cannot do and how the economics are shaping up will help you make better decisions about automation, hiring and the opportunities that will open up around this technology over the next few years.

Where Humanoid Robots Are Actually Working Today

The clearest evidence comes from deployments that have been running long enough to produce real numbers. Agility Robotics says its Digit robot has moved more than 100,000 totes at a GXO logistics facility in Georgia since June 2024, a deployment the company describes as the first formal commercial humanoid deployment. In an update released this week, Agility said GXO will expand Digit to sites in North Carolina and Tennessee starting in early 2027. It also listed Toyota Motor Manufacturing Canada, Schaeffler and Mercado Libre among its customers, said three manufacturing customers are in paid programs and reported more than $300 million in committed multi-year orders for its next-generation Digit 5, which is a measure of potential value rather than revenue already earned.

Figure AI has published some of the most detailed results so far. In a report on its BMW deployment, the company said its Figure 02 robots worked ten-hour shifts, Monday through Friday, for about eleven months at BMW’s plant in Spartanburg, South Carolina. Over that time they loaded more than 90,000 sheet-metal parts, logged over 1,250 hours of runtime and contributed to the production of more than 30,000 BMW X3 vehicles. The job demanded placing parts within a 5 millimeter tolerance inside tight cycle times. Figure was also candid that the robot’s forearm was its top hardware failure point, and it is now retiring Figure 02 in favor of a redesigned Figure 03.

To see what this looks like inside a working factory, this 60 Minutes report goes behind the scenes as Boston Dynamics trains its Atlas humanoid to work at a Hyundai plant.

Separating the Signal From the Hype

Every big technology wave comes with bold forecasts, and humanoids are no exception. In October 2026, Tesla said it had started early production of its Optimus robot at its Fremont factory. Elon Musk has claimed production could reach 20,000 units a week by the end of 2027, yet he also described the ramp as “slow and somewhat bumpy” and admitted, according to TechCrunch, that “there’s still quite a lot of work to do with regard to the AI and the robot for it to be truly useful for open-ended tasks.”

That second quote is the one founders should remember. Today’s humanoid robots are proving themselves in structured, repetitive jobs like moving totes, loading parts and feeding machines, often in facilities designed around clear workflows. They are not yet general-purpose workers you can drop into any messy environment and trust to figure things out. The deployments that work share a pattern. The task is well defined, the environment is controlled, success is measurable and humans remain close by to manage exceptions.

The wider robotics market tells a similar story of steady, practical growth rather than a sudden revolution. The Association for Advancing Automation reported that North American companies ordered 10,989 robots worth $723.5 million in the third quarter of 2026, up 24.8% in units from a year earlier. Collaborative robots, the kind built to work safely alongside people, made up about 16.7% of those orders. Most of that demand is for conventional industrial machines, not humanoids, which is a useful reminder that the best automation choice is often the simplest one that gets the job done.

The Business Model Is Changing Faster Than the Robots

One of the most important shifts for smaller companies has nothing to do with legs or hands. It is the move toward paying for robots the way you pay for software. Instead of a large upfront purchase, many providers are offering subscriptions or robots-as-a-service arrangements, where you pay a monthly or hourly rate and the vendor handles maintenance and updates. On the consumer side, 1X lists its NEO home humanoid at $20,000 to buy or $499 a month on subscription, with deliveries tied to the company’s production ramp.

For founders, that pricing model changes the math. A robot you rent by the month can be tested against a specific problem without betting the company on it, and it can be returned or upgraded as the technology improves. It also means the comparison is no longer “robot versus a year of wages” in the abstract. It becomes a practical question of cost per task, uptime, safety and how much management attention the robot requires compared with the alternatives you already have.

Where the Opportunities Are for Founders

Most founders will never build a humanoid robot, and that is fine. Some of the biggest opportunities in every platform shift sit around the core technology rather than inside it. When the personal computer arrived, fortunes were made in software, training, repair and services. When smartphones arrived, the app economy dwarfed the hardware makers in sheer number of businesses created. Robotics is likely to follow a similar path.

• Integration and deployment. Companies that buy robots need help mapping workflows, preparing facilities, training staff and measuring results.
• Software and data. Robots need task-specific training data, fleet management tools, analytics and safety monitoring, and that work increasingly looks like software.
• Maintenance and support. Every robot in the field needs servicing, parts and on-site technicians, which creates local service businesses.
• Training and workforce transition. People will need to learn how to supervise, troubleshoot and work alongside robots, and someone has to teach them.

If you run a business with physical operations, such as a warehouse, a production line, a kitchen or a distribution center, your opportunity looks different. You do not need to be first. You need to be informed. The founders who benefit most from new technology are rarely the ones who buy the flashiest version. They are the early adopters who understand the problem well enough to know when the tool is finally ready to solve it.

How to Prepare Without Getting Burned

Start by mapping the physical work inside your business. Write down the tasks that are repetitive, physically demanding, hard to staff or prone to injury, and estimate what each one actually costs you today in wages, turnover, errors and downtime. This exercise is valuable even if you never buy a robot, because it forces you to see your operations clearly and often reveals cheaper fixes, from better layout to simpler equipment.

Then match the problem to the right tool. A conveyor, an autonomous mobile robot or a collaborative arm may solve a task more cheaply and reliably than any humanoid. Humanoids make the most sense where your space, tools and processes were designed for people and redesigning them would be expensive. Before you think about physical robots at all, it is worth working out what to automate with AI agents first, because the software side of your business is usually where the fastest and cheapest gains are hiding.

When you do test a robot, run it like any serious experiment. Choose one narrow task, agree on clear success metrics with the vendor, insist on a pilot before a long contract and ask hard questions about uptime, safety certification, support response times and what happens to your data. And bring your team into the process early. The deployments that work treat robots as a way to remove the worst parts of a job, not as a threat hanging over the people doing it.

Stay Curious, Stay Grounded

Humanoid robots are at an awkward, fascinating stage. They are real enough to be loading car parts and moving totes on paid contracts, but early enough that their makers are still openly redesigning hardware and admitting how much work remains. That is exactly the moment when informed founders gain an edge, because they can learn while the costs of experimenting are falling and before the technology becomes a commodity everyone uses.

You do not need to predict exactly when humanoids will become common in small businesses. You just need to keep watching the real deployments, understand your own operations deeply and be ready to move when the numbers make sense. The future rarely arrives all at once. It shows up one shift, one pilot and one practical use case at a time, and the founders paying attention are the ones who get to shape it.