Pentagon Gave a Startup $24M for Humanoid Robot Soldiers. Here Is What It Actually Has.

Foundation Future Industries, a San Francisco startup founded in April 2024, has secured $24 million in Pentagon research contracts to test its Phantom MK-1 humanoid robot for military applications, including enemy position breaching. Two units were deployed to Ukraine in February 2026, making them the first humanoid robots tested in any active combat theater, albeit in non-combat support roles.

The contracts span the US Army, Navy, and Air Force, and include an SBIR Phase 3 designation that qualifies Foundation as an approved military vendor. The company is seeking $500 million in new funding at a valuation exceeding $3 billion, a significant jump from its roughly $21 million in total funding raised to date. The announcement has drawn both industry attention and sharp political scrutiny, with Senator Elizabeth Warren calling the contracts "corruption in plain sight" after disclosing that Eric Trump, son of the sitting president, serves as the company's chief strategy adviser.

AI-generated image

Foundation Future Industries secured $24 million in research contracts from three US military branches. Source: Biped.News / AI illustration

Foundation Future Industries: The Full Picture

Foundation was founded in April 2024 by three people with very different backgrounds. Sankaet Pathak previously ran Synapse, a banking-as-a-service platform that filed for bankruptcy in 2024 while tens of millions in consumer deposits were unaccounted for. Arjun Sethi, the CEO of Tribe Capital, led the company's $11 million pre-seed round. Mike LeBlanc, a 14-year Marine Corps veteran and co-founder of Cobalt Robotics, supplies the military credibility and has publicly stated his belief that there is "a moral imperative to put these robots into war instead of soldiers."

The company's stated mission is to build general-purpose humanoid robots optimized for military environments, not the household chore market that has attracted most competitor investment. Pathak told Business Insider he considered home-use applications "kind of stupid," arguing that the technology's highest-value application is reducing soldier casualties. The Phantom MK-1 targets logistics, reconnaissance, and, eventually, precision operations such as target extraction.

In June 2024, CNBC reported that Foundation had been fundraising with marketing materials that exaggerated ties to General Motors, including claims that GM had committed to invest and placed a $300 million purchase order. GM flatly denied all of it. LeBlanc confirmed the denial and said he was "embarrassed" the materials existed. For a company seeking Pentagon trust and a $3 billion valuation, that credibility gap remains a live issue.

Under the Hood: The Phantom MK-1 and MK-2

The Phantom MK-1 stands 5 feet 9 inches tall and weighs 176 pounds. It has 19 upper-body degrees of freedom and five-fingered hands capable of dexterous manipulation. The vision stack uses eight cameras with no LiDAR, keeping the profile compact and reducing failure points in dirty environments. The AI system translates high-level task instructions through an LLM pipeline, blending autonomous operation with supervised teleoperation. Operators retain final authority over any lethal decisions, in line with current Pentagon policy.

The standout hardware choice is proprietary cycloidal actuators delivering up to 160 newton-metres of torque. Cycloidal drives offer high torque density and backdrivability in a compact form factor, useful for tasks requiring force sensing during manipulation. Walk speed is 1.7 metres per second (roughly 3.8 mph), comparable to a brisk human walking pace. Unit cost is approximately $150,000, with a lease model available at $100,000 per year.

AI-generated image

Foundation's Phantom MK-1 uses proprietary cycloidal actuators delivering up to 160 Nm of torque. Source: Biped.News / AI illustration

The Phantom MK-2, expected this month, addresses the main MK-1 weaknesses. Consolidated electronics reduce short-circuit risk. Waterproofing and a larger battery pack extend operational range in adverse conditions. Payload capacity jumps from 44 to 175 pounds, a significant change for logistics tasks. Cast-moulded bodywork replaces the MK-1's approach to reduce cost and accelerate production. The overall goal is a robot that can operate in rain, dust, and shock conditions without field failures.

The Ukraine Deployment: What Actually Happened

In February 2026, Foundation sent two Phantom MK-1 units to an undisclosed location near Kyiv for a closed pilot demonstration. The test focused on a single scenario: supply pickup, demonstrating that a bipedal robot can carry supplies from an exterior location to an interior one while avoiding areas where a soldier would be exposed to fire. Pathak described the location as active wartime environment where "you can hear bombs go off" and power cuts are routine.

Pathak was direct about not wanting to "overstate" the Ukraine pilot. The Phantom did not fire weapons, did not engage in combat operations, and did not operate without supervision. The test demonstrated logistics capability in an austere environment, not combat capability. Foundation is using field feedback from Ukraine to inform the MK-2 design, particularly around waterproofing, battery life, and payload handling.

For context, NATO-backed ARX Robotics has already scaled production of autonomous land drones to 1,800 units per year at a UK plant. These wheeled systems perform logistics and reconnaissance without the engineering complexity of bipedal locomotion. Foundation's bet is that the human form factor becomes necessary for tasks requiring access to human-built environments: breaching structures, navigating stairs, operating equipment designed for human hands.

The Production Math: 250x in Two Years

Foundation's production targets are 40 units in 2025, 10,000 units in 2026, and 50,000 units by the end of 2027. Those numbers require a 250x manufacturing scale-up in two years on roughly $21 million in total funding. For comparison, Agility Robotics, with more commercial humanoid production experience than almost any other company, deployed seven Digit units at a single Toyota plant as a notable milestone. Tesla plans Optimus production with billions of factory investment behind it.

The military procurement path is also longer than commercial sales. An SBIR Phase 3 designation means Foundation is an approved vendor for research contracts, not a supplier of production units to front-line forces. The gap between a research agreement and a battlefield deployment system is measured in billions of dollars and years of safety certification. Shield AI's Hivemind program raised $2 billion specifically to bridge that gap for a far simpler system than bipedal locomotion with dexterous manipulation.

Autonomous Weapons Policy: Where 2026 Stands

Foundation's emergence arrives at a specific moment in international policy. The UN General Assembly First Committee voted 156 to 5 in November 2025 in favor of negotiations on autonomous weapons. The Group of Governmental Experts on lethal autonomous weapons has sessions scheduled throughout 2026 and is expected to deliver a final report to the Convention on Conventional Weapons in November. This is the last year of the GGE's mandate, making 2026 a pivotal window for international regulation.

The United States and Russia are among the minority of states blocking a binding instrument. The Campaign to Stop Killer Robots, a coalition of more than 250 NGOs, has advocated for such an instrument since 2013. Foundation's policy is that human operators retain final authority over lethal decisions, consistent with current Western military doctrine. What that means in practice as autonomy capabilities improve is a question no one has answered cleanly.

Pathak told Business Insider: "In ordinary circumstances, you probably want a human in the loop before any kind of kill action is invoked." It is a reasonable starting position, but operational pressure during fast-moving engagements tends to erode human-in-the-loop constraints over time. The ethical and regulatory questions around autonomous humanoid weapons are not resolved by stating that humans currently retain authority over lethal actions.

Frequently Asked Questions

What's Coming Next

The Phantom MK-2 is expected to debut this month with weatherproofing, a larger battery, 175-pound payload capacity, and cast-moulded bodywork for faster production. How quickly Foundation moves from prototype to manufactured units will be the first real test of whether its production targets have any grounding in operational reality. The MK-2's performance in military evaluations will also determine whether the Pentagon's research interest converts into anything resembling a production order.

The $500 million fundraise is the other near-term signal to watch. If Foundation closes that round anywhere near its $3 billion-plus valuation target, it signals that defense-focused humanoid robotics with political connectivity can attract serious capital despite credibility concerns. If the round prices lower or stalls, it indicates the market is applying stricter scrutiny to manufacturing claims and leadership history. For comparison, Apptronik recently raised a $935 million Series A at a $5 billion valuation with years of deployed hardware behind it.

The political controversy around Eric Trump and the Pentagon contracts will continue to drive media attention. But the actual test of Foundation's viability is simpler: how many MK-2 units does it ship in 2026, and at what cost? That number, when it becomes available, will tell investors far more than any government contract announcement.