Filenews 19 December 2025
By John Koetsier
Most of the hundreds of start-ups around the world that make humanoid robots target their products for working in factories, in communities, or in homes. The Silicon Valley start-up Foundation is doing the same, but is targeting the US military as its customer for the Phantom, the humanoid robot 1.75 m tall and weighing 82 kg. In fact, the Phantom will be armed and fighting. The company's CEO, Sankaet Pathak, wants to build 50,000 Phantoms in record time: by the end of 2027.
"We need to work hard to give the U.S. military smarter tools so it's more effective," Pathak said in a recent interview with TechFirst. And he does not hesitate to equip the Phantom with an M4 rifle. Using humanoid robots for military purposes is not the Foundation's only plan.
The company built the Phantom in record time. Within 18 months of its inception, Pathak had a robot that worked hard performing actual production tasks at (unknown) partner locations. The venture came to fruition so quickly thanks to two smart acquisitions in the field of AI and next-generation enablers that significantly accelerated progress.
The Foundation's expansion plan is even more aggressive: 40 robots this year, 10,000 in 2026 and 40,000 humanoid robots in 2027. Pathak acknowledges that his plan is extremely aggressive, but there are chances that it will become a reality.
That's why he's betting big on the Foundation team, which includes top talent hired by companies like Tesla, 1X, Boston Dynamics, and SpaceX.
"Our head of production is a former Tesla production manager," Pathak pointed out. "He worked on the X and Y models. He knows his lesson: don't try to automate too fast."
If the Foundation manages to handle the frenzied growth, Pathak is confident that customers will show up. He intends to rent robots, not sell them, and he doesn't need hundreds of customers.
"We don't need 50 agreements," he stressed. "We need about five high-quality deals, which can yield hundreds of millions of dollars in annual recurring purchase orders."
In fact, if the Foundation meets its production targets and sales come out as Pathak envisions, then the 50,000 robots that will be built in 2026 and 2027 will be leased for about $5 billion. in annual recurring revenue, as he estimates.
With a little math, we understand that the Foundation plans to lease robots for about $100,000. per year each. It seems expensive to plan, given that labour costs in restaurants, warehouses or production lines can cost an average of about $40,000 per year, but in reality there is a discount, as robots can operate almost 24/7, meaning that each robot does the work equivalent of three to five people. Assuming that maintenance and repairs are included in the $100,000 mark, it's a significant cost reduction.
With almost full utilization, robots could save employers $166,000 each year. With a more realistic calculation that includes both the downtime of the robots and some supervisors who will be humans, the annual savings amount to about $90,000.
All of this presupposes, of course, that the robot is as fast as humans and as good as humans: something that no humanoid robot manufacturer has been able to implement so far. Therefore, the smart investment can wait a few years according to the best-case scenario of the Foundation's estimates, or it can significantly reduce the value of robots' work, until their hardware and software reach the same levels of human capabilities.
And this despite the high-tech "muscles" of the Foundation robot, which can work for many shifts without overheating, using high-performance, reversible actuators that make the Phantom powerful and safe for humans.
But there are also military applications.
Whether we like it or not—and anyone who's seen the movie "Terminator" probably doesn't—humanoid robots are of enormous utility in military applications. They can be used to do heavy work that soldiers or support personnel can't do, they can carry weapons and ammunition, and they can take risks that no good officer wants their soldiers to take. We are already seeing uses of wheeled robots in the Russia-Ukraine war.
Among other things, the robots will be able to carry weapons to buildings where there will be battles in the streets around them, or reach the top of a hill first or enter a cave in search of the enemy.
In fact, Pathak argues that humanoid robots could highlight military forces as more valuable. Instead of launching airstrikes or using heavy weapons, the military could use a humanoid ground robot that can enter buildings, instantly assess a situation and make the right decision. While the Phantom will not be fully autonomous in making decisions involving the use of weapons, the Foundation envisions a model similar to today's military drones, where robots handle movement and navigation themselves, while people at the base (or nearby, but hidden in a safe place) retain control of the targeting. If this approach is confirmed, humanoid robots like the Phantom may eventually change not only the way wars are fought, but also when they are avoided.
Pathak argues that humanoid drones of the "Terminator" type would actually be highly effective peacekeepers.
"I think if the U.S. military had 100,000 robots that could just demonstrate their function, that would put an end to a lot of wars before they even started," he says.
This makes sense: the U.S. would be more likely to send these robots to a potential battlefield than humans who could be killed and followed by political reactions. And so, insurgents, terrorists or other hostile fighters would find it more difficult to decide to start a conflict.
Equally likely, however, is the opposite scenario: the military uses humanoid drones for combat, and wars are more common.
Either way, the ethics of humanoid robots are about to get even more complicated. And if you think that China or Russia are not thinking of building military robots... think again. Essentially, although not in humanoid form, they are already doing so. So has the US military, which has had Predator drones for three decades.
The first automated and unmanned weapons date back to World War II: the German V-2 rockets. They had very primitive and rudimentary autonomous guidance, but perhaps we can describe them as the first military robots. Now, we're just increasing the level of complexity. And, of course, mortality.
