If you are running a warehouse or factory in 2026, you are probably struggling to hire qualified labor. While not yet impossible, it’s getting harder to find good people willing to do the tough jobs. If you live with this challenge every day, you’ve probably spent some time wishfully looking at YouTube clips of humanoid robots doing backflips or stacking boxes.
The newspapers are full of recent headlines like “Humanoid Beats Usain Bolt’s 100m Sprint Record” or “European Unicorn “Humanoid” Lands 34,000 Pre-Orders”. Just last week, SpaceX and Tesla CEO Elon Musk told the 2026 G20 summit in North Carolina:
“10 years from now, I would say there are well over a billion humanoid robots. The productivity per robot will be probably five times that of a human.”
That is an impressive prediction! Two years ago, he announced the new humanoid Tesla is designing “in the long term—at scale—the robot should cost somewhere between $20,000 and $30,000. Less than a car.” A worker that is five times more productive and costs less than $30k? Sign me up!
Mr. Musk has made bold statements in the past. In 2019, he predicted that Tesla would have “1 million robotaxis on the road next year”, but it was only last week that they had their first fully ‘steering-wheel free’ rides in service, with a reported few dozen Cybercabs registered to Tesla’s Robotaxi fleet in Texas. That is a far cry from one million, and six years too late.
I have clients constantly asking me about the humanoid robots they see in the news, so let’s have a “Quick Check” on where the technology is these days with respect to our warehouses, factories, and supply chains.
Are They Available and Making an Impact?
The short answer is “not really”. According to market research firm Smart Analytics Global (SAG), worldwide shipments of humanoid robots reached about 19,000 units in the first half of 2026. This is a large increase (272%!) from a year earlier, with Chinese makers accounting for 97% of those sales. According to Reuters, most of these units went to research, education, entertainment, and data collection rather than industrial use. There are around 5 million warehouse workers globally that move materials, so we can safely say this small number of robots is not yet having an impact.
If you live in the US, you are now essentially blocked from buying any newly released Chinese humanoid robots. On July 28, 2026, the FCC restricted the importation, marketing, and sale of any new foreign-produced mobile robots in the United States, including humanoid robots. It remains to be seen whether Chinese manufacturers will be able to establish U.S.-based operations to work around these restrictions.
There are at least 10 companies in the US that are trying to build full humanoid robots. So far, most of these robots have been only tested in laboratories or very limited pilot programs in a small handful of real-world industrial situations. At this time, we can say that humanoid robots are not readily available to help improve our supply chain operations.
What Do We Need from a Humanoid Robot?
For humanoid robots to succeed in industry, three factors must align: Capability, Reliability, and Economics. To achieve widespread adoption, we will need to add Safety and Multiple Vendors to our criteria list.
One of the leading humanoid robotics companies that is focusing on industrial use is Agility Robotics. They are one of the few companies currently selling humanoid robots, and you may have seen them in action at some recent trade shows like ProMat or Automate. In preparation for going public, the company published an Investors’ Presentation that includes the claim that they have 1,000 orders pending for their next generation robot. Let’s see what we can learn using their robot called “Digit” as an example.
Capability: So far, all demos we can see on YouTube or at the trade shows have Digit moving plastic totes, by using simple “pinch grip” hands to grab and manipulate them. When we take a step back, these demonstrations are amazing to see, but when we take a closer look, we should wonder how many tasks in our operations involve only moving plastic totes or using simple grips. The famous roboticist Rodney Brooks stated last January that, in his estimation, “deployable dexterity will remain pathetic compared to human hands beyond 2036.” While there are tasks in our warehouses that don’t need perfect humanlike dexterity, there are many that do, and this will limit the potential of humanoid robotics until some further fundamental breakthroughs are made.
Reliability: One of the main things we need from our automated systems is reliability. So far, the reliability of humanoid robots has not yet been proven. For example, Agility’s Investor Presentation gives two pieces of information: 1) the robot is designed to have a service life of 31,000 hours, or roughly 5 years based on Agility’s assumptions; and 2) their robots have run for 65,000 hours in real-world conditions. These hours were spread over at least 9 sites with multiple robots in each, meaning that no one robot has come near its service life limits in real operational conditions. The long-term reliability of humanoid robots under continuous industrial use remains largely unproven.
Economics: Realistic pricing for humanoid robots has been hard to find since most companies are still in development mode. Agility’s Investor Presentation shows that they expect to rent their robots to customers at a monthly price of $8,500 USD, after a $25k USD deployment fee. This leads to a cost of over $100k USD per year. To show that their robots are affordable, they are assuming that their robots will work 20 hours per day, 6 days per week, replacing more than one human worker each. Many warehouses work only one shift per day, so this robot will not be economical for them at this time. Based on the currently limited capability of these robots, it may be difficult to find tasks that they can comfortably do that also require a person 20 hours per day, 6 days per week. A final point: one of their initial pilot tests reported that the robots worked at 65% of the speed of a human worker, meaning you may need two robots to replace one person, making the business case even more difficult to justify.
Safety: Currently there are no widely accepted safety standards specifically addressing humanoid robots. Most of the pilot programs keep robots behind metal fences and away from human workers, limiting the tasks that humanoid robots can do. There is a committee developing ISO 25785, the first global safety standard for “dynamically stable” robots, of which Agility is a part of. That standard is not expected to be released until late 2027 and will likely force robots to shut down in a safe way if people come close to them. Can we effectively deploy humanoid robots if they keep shutting down every time we walk by? How will insurance companies and local governments look at the risks inherent in these systems? Will corporate health and safety leaders be willing to take on those risks? Right now, there are no clear answers to any of these questions.
Multiple Vendors: To recommend core technology to our clients, we prefer that there are multiple vendors for them to choose from, each with a track record of success. Right now, we don’t see any one vendor who sells a humanoid robot that meets the criteria of Capability, Reliability, and Economics, let alone several to choose from. Once successful humanoid robots have been developed, it will still take many years for a community of financially stable competitors to develop that have strong experience, reliable systems, and robust service networks in place. When that happens, we will be able to finally say that humanoid robots are ready for action!
Conclusion
Right now, robotics is a rapidly evolving and exciting field to be part of. We are really excited about the current possibilities and future potential of the newly emerging technologies we see entering the market. 2026 is probably not the year to buy your first humanoid robot, but the potential for a breakthrough is stronger than ever before. There are hundreds of startup companies spending billions of dollars on research and development right now, it’s only a matter of time before one of them unlocks the future. St. Onge Company will keep watching these developments, as always, and do our best to help our clients jump in at the right time.
—Tom Bonkenburg, St. Onge Company
