Humanoid Robots Are Going to War — and to Warehouses: The Industry's Two Very Different Futures
The humanoid robotics industry is splitting into two distinct trajectories this week, and the contrast couldn't be sharper: one company is quietly building robots that move totes for Amazon and Toyota; another is designing them to kick down doors in combat zones. Both developments landed within 24 hours of each other, offering a rare snapshot of just how wide the possibility space for bipedal machines has become.
Agility Plants a Flag in Silicon Valley's New Robot Belt
Agility Robotics announced a new 60,000-square-foot training facility in Fremont, California — a deliberate provocation, given that the address sits just up the highway from the factory where Tesla is expected to begin manufacturing its Optimus humanoid robots this year. TechCrunch broke the story on July 17th, and the symbolism is hard to miss.
But Agility CEO Peggy Johnson isn't framing this as a rivalry. "It's great to have [Tesla] in the same area as us, because really, for a long time Agility was out there alone, and it's good to have others in the humanoid space," she told TechCrunch. What she's really communicating is this: Agility already has something Tesla doesn't — a commercially deployed robot generating real revenue in real facilities right now.
Digit, the company's six-foot bipedal robot, is already moving totes and bins for customers including Amazon, GXO, Schaeffler, and Toyota Motor Manufacturing Canada. Agility has secured $300 million in contract orders, and Digits have reportedly moved over 100,000 totes at a GXO logistics facility alone. More than 30 additional customers are reportedly in deployment talks.
The Fremont facility is where Digit will learn new skills in environments that mirror the warehouses and factories it will actually work in — a kind of physical simulation layer that mirrors how software engineers use staging environments before production deployment. It's a grounded, methodical approach to scale, and it reflects Agility's founding DNA: the company was started in 2015 by researchers who developed novel techniques for safe bipedal locomotion.
The AI Safety Stack Question
Agility co-founder and chairman Damion Shelton offered one of the more technically astute framings of the generative AI question I've seen from a robotics executive. He draws a hard line between what AI should and shouldn't control.
""When you think about self-driving cars, as a non-humanoid example, you really don't want the anti-lock brake controller under AI control. The analog with humanoids is all the safety stuff needs to go through a path that's not generative AI. You don't want to get creative with your safety stack."
That's not AI skepticism — it's systems engineering. Safety-critical paths need deterministic behavior. What generative AI does unlock, Shelton argues, is the application layer: the near-infinite variety of tasks you'd want a physical robot to do, for which there will never be enough engineers to hand-code solutions.
Board member Bruce Leak — the QuickTime inventor — apparently pushed the founding team early on how they'd handle programming Digit for diverse tasks. Shelton admitted they didn't have a great answer at the time. Generative AI, he says, now answers that question definitively.
Agility is also moving toward the public markets. CEO Johnson is currently leading a reverse-merger that would make Agility the first pure-play humanoid robot company listed on public markets, expected to close later this year.
Meanwhile, One Startup Is Building Robot Soldiers
While Agility is optimizing for warehouse compliance checklists and IT infrastructure handshakes, a very different humanoid story is unfolding in Florida. Wired reported on July 17th that Foundation Future Industries, a startup founded in 2024, is explicitly developing its humanoid robot — the Phantom MK1 — for lethal military applications.
"We have some kinetic things we're exploring," CEO Sankaet Pathak told Wired, using careful language for what are plainly weapons systems. "We'll probably unveil something in the next couple of months."
The company has high-profile political backing. Eric Trump — the president's son — is both an investor and the company's chief strategy adviser. The connection has opened doors: a Fox Business segment in April cited a "$24 million contract with the Pentagon," though Wired's reporting reveals that figure appears to include contracts inherited from Boardwalk Robotics, a company Foundation acquired, plus contracts that originated with the Institute for Human and Machine Cognition (IHMC) in Florida, a nonprofit research institute with deep humanoid expertise. Foundation does not appear to have independently secured new government contracts.
That's a meaningful distinction — but the underlying market thesis may still be sound. An anonymous roboticist familiar with the company told Wired: "If you put a military hat on, it makes a lot of sense, because it's where soldiers still die — that first entry through a door." The same source added: "I think it is so close to feasible that I'm surprised they're not already fielded."
Foundation says its robots have already been tested with Ukrainian forces — a detail that underscores how the war in Ukraine has accelerated the development and real-world validation of autonomous and semi-autonomous military systems globally.
The US military's interest in humanoids has deep institutional roots. DARPA funded major humanoid competitions between 2012 and 2015, and the Army's current xTechHumanoids program specifically funds technologies for "militarized humanoid capabilities." Foundation is positioning itself at the intersection of that long-standing interest and the current wave of capable hardware.
What This Week Tells Us About the Industry
These two stories, arriving within hours of each other, illuminate something important about where humanoid robotics actually stands in mid-2026. The technology has matured enough that the market segmentation conversation has begun in earnest. Warehouse logistics. Military applications. Healthcare. In-home assistance — notably, something Agility is explicitly not pursuing, a position that aligns with most independent robotics experts who caution that today's robots aren't ready for unstructured domestic environments.
Agility's commercial traction — real customers, real deployments, a public market debut on the horizon — represents the cleaner, more verifiable story of the two. Foundation's military pivot is more speculative, more politically entangled, and involves claims that deserve scrutiny. But the underlying demand signal from defense customers is real, and the logic of using bipedal robots for dangerous entry operations is militarily coherent.
Watch for Agility's reverse-merger timeline in the second half of 2026 — its debut on public markets will be a genuine signal moment for investor confidence in commercial humanoid deployment. And watch for Foundation's promised "kinetic" unveiling in the coming months. How the Pentagon and defense community respond will tell us a great deal about how seriously the military is treating this generation of hardware.
The robots are here. The question now is simply: what do we send them to do?