The battlefield is becoming a testing ground for embodied AI.
San Francisco | July 2026
A United States robotics startup backed by Eric Trump is preparing humanoid machines for military applications and exploring the future integration of weapons, bringing a concept long associated with science fiction closer to operational defense technology.
Foundation Future Industries is developing the Phantom series of humanoid robots for logistics, reconnaissance, inspection and other tasks in dangerous environments. Chief executive Sankaet Pathak confirmed that the company is also evaluating what the defense industry calls “kinetic” capabilities, a term generally associated with systems capable of delivering physical force.
Pathak has not disclosed which weapons or targeting mechanisms are under consideration, but said the company could unveil further developments within months. The announcement does not mean Foundation currently possesses an autonomous humanoid soldier ready for combat. Its existing machines remain experimental platforms confronting substantial limitations in reliability, mobility, battery life and physical manipulation.
The company’s principal prototype, Phantom MK1, is designed to operate in environments created for human beings. A humanoid form could theoretically allow the machine to climb stairs, enter buildings, manipulate equipment and move through spaces where wheeled or tracked vehicles cannot function effectively.
Foundation says the robot has been tested with Ukrainian forces, although public information does not establish that it participated in combat or carried weapons. Ukraine has become an important testing environment for military robotics, drones, electronic warfare and artificial intelligence because systems can be evaluated under conditions that laboratories cannot fully reproduce.
Pathak has previously said the company ultimately wants robots capable of identifying targets and operating weapons. He has also maintained that any lethal action should remain subject to human authorization rather than being executed independently by the machine.
Under that proposed model, the robot could navigate, recognize objects and process battlefield information with a degree of autonomy, but a human operator would retain control over the final decision to use force. The distinction is central to the growing international debate over military AI.
Foundation was created in 2024 and later acquired Boardwalk Robotics, a company associated with humanoid research conducted alongside the Florida Institute for Human and Machine Cognition. That acquisition gave the startup access to technical expertise, intellectual property and existing relationships involving government-funded programs.
The company has promoted its connections with multimillion-dollar defense agreements, but reporting surrounding those claims indicates that several contracts were inherited from Boardwalk Robotics or conducted through institutional partnerships. Public evidence does not clearly demonstrate that Foundation independently secured every government contract attributed to it.
Eric Trump, the son of US President Donald Trump, is both an investor and Foundation’s chief strategy adviser. He has publicly praised the company’s robots and argued that the combination of humanoid hardware and artificial intelligence could transform military, industrial and hospitality applications.
His involvement gives the startup greater public visibility while raising questions about the relationship between political influence, private investment and federal defense procurement. No evidence presented in the reporting establishes improper government intervention, but transparency will be important if the company pursues additional Pentagon contracts while the president’s family remains financially connected to it.
The military interest in humanoid robots predates Foundation. The Defense Advanced Research Projects Agency funded major robotics competitions more than a decade ago, while the US Army has supported programs exploring technologies relevant to militarized humanoid systems.
The strategic logic is straightforward. Robots could enter contaminated facilities, inspect suspected explosive devices, transport supplies under fire or move ahead of soldiers into buildings where ambushes are likely. Replacing humans in the most dangerous stages of an operation could reduce military casualties.
Humanoid systems may also perform tasks outside direct combat. They could carry ammunition, evacuate wounded personnel, inspect damaged infrastructure, collect intelligence or operate inside environments affected by chemical, biological or radiological hazards.
Yet the difference between a controlled demonstration and a reliable battlefield machine remains enormous. Combat environments contain smoke, rubble, unstable surfaces, damaged stairs, blocked doors, poor communications and unpredictable human behavior.
Modern humanoids can display impressive balance and coordinated movement under prepared conditions, but they frequently require extensive training for specific environments. Picking up irregular objects, using tools and recovering from unexpected physical contact remain difficult engineering problems.
Battery capacity creates another limitation. A soldier can operate for extended periods while carrying food and equipment, but an electric robot requires access to energy, replacement batteries or charging infrastructure. A machine that becomes immobile after a short deployment could create an additional logistical burden.
Foundation’s planned Phantom MK2 is expected to improve environmental protection by making the machine resistant to water and dust. Those features would be basic requirements for real military deployment, where rain, sand and debris can rapidly damage exposed electronics and mechanical systems.
Robotics specialists caution that dependable humanoid soldiers may remain more than a decade away. Even after a successful laboratory demonstration, military certification and operational deployment require years of testing, maintenance planning, cybersecurity evaluation and integration with existing forces.
The prospect of arming humanoids nevertheless intensifies the ethical debate surrounding lethal autonomous weapons. The central concern is not the machine’s humanlike appearance, but the degree of authority delegated to its software.
A drone, ground vehicle or robotic vessel can become a lethal autonomous system if it selects and attacks targets without meaningful human control. A humanoid robot is therefore only one possible physical form of a much broader technological category.
The United Nations has called for an international prohibition on machines empowered to decide independently who lives and who dies. Secretary-General António Guterres has argued that lethal autonomous systems operating without human judgment are morally unacceptable and should be banned under international law.
Supporters of military automation contend that AI-assisted systems could process information faster, improve targeting precision and reduce errors caused by fear, fatigue or emotional reaction. They also argue that robots could protect soldiers by assuming missions with extreme casualty risks.
Critics warn that algorithms can misidentify civilians, misunderstand surrender gestures or act on incomplete data. Battlefield conditions may also disrupt sensors and communications, producing errors that cannot be reversed once lethal force has been applied.
Accountability presents another unresolved challenge. When an autonomous or semi-autonomous system kills the wrong person, responsibility could be distributed among commanders, operators, software developers, manufacturers and government agencies.
Military robots could also lower the political cost of initiating conflict. Governments may become more willing to use force when their own soldiers are not immediately exposed, even though civilians and opposing forces continue carrying the human consequences.
Cybersecurity adds another layer of risk. A weaponized humanoid connected to networks or controlled remotely could become vulnerable to interference, spoofed sensor data, software failures or hostile takeover.
Foundation argues that robotics can make warfare more precise and reduce collateral damage. That promise will depend not only on engineering performance, but on operational rules, human supervision and legal accountability.
The Phantom project does not yet represent a real-world Terminator. It represents something more immediate: a private company actively trying to move humanoid robots from controlled demonstrations toward military utility, including the possible use of weapons.
The decisive question is no longer whether artificial intelligence will enter armed conflict. It is how much authority societies will permit machines to exercise once they arrive there.
Phoenix24 | Technology can enter battle faster than ethics can establish its limits. La tecnología puede entrar en batalla más rápido de lo que la ética puede establecer sus límites.