Humanoid robots are moving from science fiction into military reality faster than most people realize. Foundation Future Industries, a US-based robotics company, has already deployed its Phantom robots in Ukraine for non-combat roles like material handling and reconnaissance, and the company's CEO says weaponization testing could begin as early as next year.

When Will Humanoid Robots Actually Enter Combat?

The timeline for armed humanoid robots on the battlefield is accelerating. Foundation Future Industries CEO Sankaet Pathak told Euronews Next that his company expects to start testing weaponization use cases "as early as next year," following successful pilot programs in Ukraine. The company is already building its next-generation Phantom 2 robot, which features significant durability improvements designed specifically for harsh battlefield conditions.

Sankaet Pathak

Pathak argues that humanoid robots make military sense not for widespread destruction, but for precision operations where avoiding civilian casualties and infrastructure damage matters. "Humanoids only make sense when your mission objective is more precision, essentially trying to make sure you don't destroy infrastructure, hurt civilians, do the best of your ability as you can accomplish a very complex mission," he explained.

Pathak

What Makes Humanoid Robots Different From Drones?

The key distinction, according to Pathak, is that humanoids fill a gap that drones cannot. While drones excel at aerial strikes, humanoids could handle ground-level tasks requiring dexterity and situational awareness. The company has already tested humanoids in Ukraine for roles including material handling (moving supplies between indoor and outdoor environments) and reconnaissance (mapping interiors and clearing buildings).

The military testing in Ukraine forced Foundation to completely redesign its hardware for outdoor, high-stress conditions. The improvements are substantial:

• Payload Capacity: Increased from roughly 25 to 30 kilograms in the original Phantom to about 80 kilograms in Phantom 2
• Impact Tolerance: Improved from 12 to 15 G-force to nearly 100 G-force, meaning the robots can survive much harder falls
• Environmental Durability: Phantom 2 is waterproof and dustproof, with a 3-kilowatt-hour battery for extended operations

Currently, Foundation leases Phantom robots commercially for about $100,000 per robot per year, with military customers purchasing at similar pricing.

How to Understand the Regulatory Landscape for Autonomous Weapons

The legal and ethical framework governing armed humanoid robots remains underdeveloped. Here are the key regulatory developments:

• Current Legal Status: Humanoid robots fall under existing international humanitarian law (IHL), which requires weapons to respect the distinction between combatants and civilians, but no specific treaty governs humanoid or autonomous robots on the battlefield
• UN Negotiations: Since 2023, the United Nations has been negotiating a dedicated treaty on Lethal Autonomous Weapons Systems (LAWS) through the Convention on Certain Conventional Weapons
• Push for a Ban: UN Secretary-General Antonio Guterrez is pushing for a binding ban on weapons that operate without human control by 2026, calling them "killer robots" that select and engage targets without human judgment

Pathak pushes back against the idea that humanoids should be treated differently from other precision weapons already in use, such as armed drones and unmanned ground vehicles. He argues that the comparison to existing weapons systems is more relevant than fears of rogue artificial intelligence.

The company's humanoid robots use artificial intelligence built on "world models," which learn from videos, simulation data, and spatial inputs to build representations of physical environments and predict what will happen next. This is fundamentally different from large language models (LLMs) like ChatGPT, which predict the next word in a sentence. Instead, world models predict how physical objects and environments will change over time.

"We're heavily focused on world models that can predict the future. We think that's going to be at the core of building intuitive AI that's robust," said Pathak.

Sankaet Pathak, CEO at Foundation Future Industries

What Are the Real AI Safety Risks?

Pathak distinguishes between theoretical existential risks and immediate practical threats. When asked whether AI could seize control of armed robots, he argued that such a scenario would more likely involve drones or nuclear arsenals rather than humanoids. "If an AI's mission objective is to destroy humanity, I guarantee you they're not sending 100,000 humanoids. I think they're just using drones or nuclear arsenals," he stated.

Instead, Pathak identifies "AI terrorism" as the more immediate danger, meaning misuse of widely available AI models for cyberattacks, disinformation, or arming consumer drones. He blamed open-source AI models for this risk, noting that when Meta released its open-source Llama 2 language model in 2023, people released uncensored versions within days that could answer questions about building nuclear weapons.

However, Pathak believes that scenarios where AI systems rewrite their own directives, self-improve, and self-replicate independent of large computing clusters remain several major technical steps away. "As these models become more efficient, as they can self-replicate with less compute, they can improve themselves and also demonstrate behaviours where they can bypass and update their own directives. That's when you're getting to that territory," he explained, but added that "we're probably like three, four, maybe five hops away from that".

Pathak

Foundation's investor base includes Eric Trump, the venture capital firm Stripe, and Define, among other backers. As humanoid robots move closer to military deployment, the regulatory and ethical questions surrounding their use will likely become impossible to ignore.