Agile Robots Turns Factory Revenue Into a Physical AI Signal

The News

Agile Robots CEO Zhaopeng Chen told the Wall Street Journal that the company is on track to double revenue this year, from about EUR 300 million in 2025, and aims to reach profitability within the next two to three years. Robotics & Automation News summarized the interview on August 4, placing the target in the context of rising industrial automation demand.

That matters because the humanoid robotics sector is full of companies with impressive technical claims but thin commercial disclosure. Agile Robots is still developing humanoid systems, including Agile ONE, but the company is not waiting for humanoids to carry the business. Its present scale comes from industrial robotics, machine tending, lab automation, production equipment, and factory integration work.

The company has also been buying distribution and integration capacity. In April, Agile Robots acquired thyssenkrupp Automation Engineering, a move positioned as a way to expand market reach, strengthen North American access, and deepen its physical AI strategy. The result is a robotics company that looks less like a pure foundation-model startup and more like a manufacturing systems integrator with AI ambitions.

A Different Kind of Humanoid Bet

Most public attention in humanoid robotics goes to the body. How fast can it walk? Can it recover from a shove? Can it fold laundry, pick parts, sort parcels, or load totes? Agile Robots is making a quieter argument. The body is only one part of the problem. The company also needs fixtures, tooling, sensors, safety envelopes, process engineering, service teams, and an installation playbook.

That view fits Chen's reported comments about factory reality. Industrial customers do not buy autonomy in a vacuum. They buy uptime, cycle time, safety, maintainability, and line compatibility. A robot that works in a lab but fails on uneven flooring, awkward lighting, messy cable routing, or legacy workstations is still a research system from a factory manager's point of view.

Agile Robots' revenue target suggests the company is winning near-term work where the value proposition is already clear. It can install automation around constrained processes, then use those deployments to collect operational knowledge. That is a more practical road to physical AI than asking customers to accept a general-purpose humanoid before reliability, service economics, and safety evidence are ready.

AI-generated image

Physical AI depends on perception, calibration, tooling, and control loops, not just a humanoid shell. Source: AI-generated editorial illustration.

Google DeepMind Gives the Strategy a Software Layer

Agile Robots' March partnership with Google DeepMind is the other half of the story. The companies said they would combine Gemini Robotics foundation models with Agile Robots' scalable industrial robotics platform. The stated goal was adaptable, reasoning robots for industrial environments.

That phrase can sound vague, but the practical target is specific. Factories need robots that can handle variation without endless hand-tuning. A machine tending cell may see parts arrive slightly off position. A logistics station may face packaging that shifts by supplier. A robot may need to reason about sequence, grip choice, obstacle avoidance, and when to stop for a human operator.

The DeepMind tie-up gives Agile Robots access to one of the most visible robot foundation model programs in the market. Recent Gemini Robotics updates have emphasized whole-body control, embodied reasoning, dexterous manipulation, on-device adaptation, and safety-aware operation. Agile Robots brings the factory customers, system integration, and hardware surfaces where those models can be tested against real operating constraints.

Why Revenue Is the Signal

A EUR 300 million revenue base is not proof that humanoid robots are ready for broad deployment. It is proof that Agile Robots has found demand for robotics work in places where customers already measure return on investment. That distinction is important.

Humanoid startups often ask investors to underwrite a future market. Agile Robots can point to existing automation revenue while making the case that physical AI will expand what those systems can do. In practical terms, the company has a bridge from current revenue to future autonomy. Many humanoid-first companies still need to build that bridge.

The company's M&A strategy also changes the risk profile. Buying integration-heavy businesses is not glamorous, but it can solve a major robotics bottleneck. The bottleneck is not only whether a robot can perform a task once. It is whether a vendor can scope, install, commission, maintain, and support dozens or hundreds of systems across different facilities.

The Humanoid Implication

Agile ONE should be understood as part of a broader industrial platform, not the whole company story. The humanoid form factor is attractive because factories, warehouses, and service sites are built around human reach, human tools, and human pathways. But it is also expensive, mechanically complex, and harder to certify than a fixed arm or mobile base.

That is why Agile Robots' industrial base matters. A company already selling automation into factories can introduce humanoids where the form factor solves a narrow integration problem. It does not need to start with a universal worker. It can start with supervised material handling, tool use, inspection support, or awkward tasks where traditional automation is too rigid.

The risk is that revenue from conventional automation does not automatically translate into humanoid leadership. Industrial arms, custom cells, and lab systems are different products from bipedal general-purpose robots. Agile Robots will still need to prove humanoid reliability, battery life, safety behavior, serviceability, and the basic unit economics of placing a mobile robot near people.

AI-generated image

The commercial race in humanoids increasingly runs through edge compute, industrial data, and integration capacity. Source: AI-generated editorial illustration.

What This Means for the Market

Agile Robots is now a useful counterweight to the humanoid market's loudest narratives. One narrative says the winning company will be the one with the best foundation model. Another says the winner will be the one with the cheapest robot body. Agile's story points to a third requirement: the winner may need the boring machinery of deployment.

That machinery includes application engineers, safety documentation, customer support, spare parts, plant audits, local service partners, and the ability to make a robot work around processes that were never designed for robots. It is expensive and operationally messy. It is also where many automation projects succeed or fail.

If Chen's revenue target holds, Agile Robots will have more than a technical claim. It will have a scaling business inside the same factories where physical AI is supposed to prove itself. That does not make Agile the automatic humanoid winner. It does make the company harder to dismiss as another lab-stage robotics bet.

FAQ

The Bottom Line

Agile Robots' revenue target is not a humanoid deployment victory lap. It is a reminder that physical AI may scale through the industrial automation companies already close to customers, equipment, and factory data.

That makes Agile Robots one of the more interesting companies to watch in the next phase of humanoid robotics. The company has AI partnerships, humanoid ambitions, and a growing automation business. The hard question is whether those pieces combine into deployable general-purpose robots, or simply make Agile a stronger industrial automation platform.

For now, the signal is clear enough: in robotics, revenue still counts. The companies that can sell useful systems today may have the best chance of turning physical AI into something customers trust tomorrow.