Android's Creator Is Building a Robot Operating System in Tokyo

Andy Rubin wants to do for robots what he did for phones. The man who co-created Android, led Google's robotics acquisitions in 2013, and later founded hardware startup Essential and venture firm Playground Global has been quietly building a humanoid robotics company in Tokyo since late 2025. In an interview with Nikkei Asia published July 29, Rubin confirmed that Genki Robotics will launch its first product before the end of this year and open a developer marketplace for robot behaviors.

The plan is blunt. Rubin intends to build the horizontal software layer that sits between humanoid hardware and the tasks people want robots to perform. Developers will write behaviors, package them, and sell them through a storefront that works like a mobile app store. Customers will download capabilities the way smartphone owners download apps. The robot itself becomes a platform rather than a finished product.

The Android Playbook Applied to Physical Machines

Rubin's track record makes the comparison unavoidable. Android launched in 2008 as an open-source mobile operating system. Within five years it ran on more than a billion devices. The key insight was that no single company could write all the software the world's phone users would need. An open ecosystem solved the problem by distributing development across millions of third parties while Google captured value at the platform layer.

Genki is attempting the same architecture in robotics. The company is designing both the hardware and the foundational software stack, but it plans to open the behavior layer to outside contributors. In a related Nikkei Japan piece published July 22, Rubin described the vision directly. Customers would purchase and download robot behaviors the same way they purchase smartphone applications. The robot would arrive with basic locomotion and manipulation. Everything else would come from the marketplace.

This is a deliberate contrast to the vertical integration model that dominates the humanoid industry today. Companies like Figure, Unitree, and Tesla are building robots where the hardware, the foundation model, and the task-specific behaviors are all developed in-house. Rubin is betting that the vertical approach will hit the same wall that Nokia and BlackBerry hit in phones. No single engineering team can cover every use case. The company that opens the platform wins.

A Billion-Dollar Valuation Before a Single Product Ships

Genki Robotics raised approximately US$50 million in a Series A round disclosed in April 2026, according to Axios. The round valued the company near US$1 billion. Playground Global, Rubin's own venture fund, participated alongside undisclosed institutional investors. The startup was incorporated in late 2025, meaning it reached unicorn territory in roughly six months without shipping hardware.

That timeline is aggressive even by the standards of the current humanoid funding boom. In China, Morphi reached a similar valuation in under a year with backing from Alibaba and Tencent. In the United States, Figure raised US$675 million at a US$2.6 billion valuation in early 2024 before delivering commercial units. The difference is that Rubin is not primarily selling robots. He is selling the idea that robots need an operating system, and that the operating system is where the value will concentrate.

The Tokyo headquarters is itself a strategic choice. Japan has the world's oldest industrial robotics ecosystem, the deepest pool of manipulation and locomotion researchers, and a demographic crisis that makes labor-replacing automation a political priority rather than a political risk. Rubin can recruit from the same talent pool that built ASIMO, NEXTAGE, and the HRP series while operating in a regulatory environment that actively welcomes humanoid deployment.

What the Marketplace Model Means for the Industry

If Genki ships a working product and opens its marketplace, the implications run in several directions. First, it creates a new category of robotics company. Developers who lack the capital to build hardware can still participate in the humanoid economy by writing and selling behaviors. Second, it introduces network effects. The more behaviors available on the platform, the more useful the robot becomes, which attracts more customers, which attracts more developers. This is the flywheel that made iOS and Android dominant.

Third, it challenges the assumption that foundation models trained on massive robotic datasets are the only path to general-purpose capability. Rubin's model suggests an alternative. Instead of one giant model that does everything, you assemble capability from a library of specialized modules. The robot becomes general-purpose not because its brain is general-purpose but because its app store is.

The skeptical counterpoint is obvious. Robots are not phones. A smartphone app that crashes causes frustration. A robot behavior that fails causes physical damage, injury, or worse. The safety certification burden for downloadable robot behaviors in unstructured environments is orders of magnitude harder than anything the mobile app ecosystem ever faced. Rubin has not publicly addressed how Genki will handle safety validation for third-party behaviors, and the Nikkei Asia interview did not explore the question.

Timing and Competitive Context

Rubin's announcement lands in a week when the boundaries of the humanoid market are being redrawn. Washington banned imports of Chinese humanoid robots on July 28. Berlin's IFA trade show confirmed a record humanoid exhibitor roster for September. Teradyne reported a 33 percent year-over-year jump in its robotics segment. The industry is simultaneously growing and fragmenting along geopolitical lines.

A Tokyo-based platform company sits in an interesting position within that fragmentation. Japan is allied with the United States but maintains deep commercial ties with Chinese robotics suppliers. A robot OS built in Tokyo could theoretically serve hardware from both blocs, provided export controls do not close that door. Rubin has not commented on whether Genki's platform will be hardware-agnostic across geographies, but the horizontal model implies it.

The first product launch later this year will be the real test. A marketplace is only as good as the hardware it runs on. If Genki's robot cannot perform basic locomotion and manipulation reliably, no amount of platform strategy will matter. Rubin built Android on top of hardware that already worked. He is now trying to build the platform and the hardware simultaneously. That is a harder problem, and the market will judge him on the hardware first.

The Founder's History With Robotics

Rubin's interest in robots predates Android. He worked at Carl Zeiss on robotic manufacturing systems in the early 1990s before moving into mobile software. At Google, he led the acquisition of eight robotics companies in 2013, including Boston Dynamics, Schaft, and Industrial Perception. The robotics division was later dismantled after Rubin left Google in 2014 amid internal controversy. Boston Dynamics was sold to SoftBank and eventually to Hyundai.

The Genki venture represents Rubin's return to the field he tried to build inside Google a decade ago. The difference is that the enabling technologies have matured. Large language models, vision-language-action architectures, and sim-to-real transfer did not exist in practical form in 2013. The hardware has also improved. Actuator costs have fallen, battery density has risen, and Chinese manufacturers have driven component prices down by an order of magnitude. The conditions that made Android possible in 2008, cheap hardware plus expensive software, may now apply to humanoid robots.

Whether Rubin can execute at 62 years old, from a standing start, against companies that have been shipping robots for years, is the open question. The billion-dollar valuation says the market believes he can. The lack of a shipped product says the market has not yet been proven right.

Japan as the Neutral Ground for a Platform Play

There is a second layer to the Tokyo decision that deserves attention. Platform businesses need adoption from hardware makers, and hardware makers are wary of platforms controlled by their competitors. Google could spread Android because Google did not make phones when the OS launched. A robot OS headquartered in San Francisco would face suspicion from Chinese hardware makers. One headquartered in Shenzhen would face bans in Washington. Tokyo sits outside both gravity wells.

Japan also brings assets no other neutral venue can match. The country's component makers, Harmonic Drive, Nabtesco, THK, Nidec, supply precision parts to nearly every humanoid program on earth, including the American and Chinese ones now being walled off from each other. A platform company embedded in that supply chain sees the whole industry's hardware roadmap. And Japan's government has spent three years assembling subsidies for robotics software, an area officials openly describe as the country's weakest link. A foreign-founded robot OS company in Tokyo is precisely the kind of venture those programs were designed to attract.

The local ecosystem is already responding. Japanese robotics engineers who once had two career paths, the large machinery groups or academia, now have a third option with Silicon Valley compensation attached. Recruiters in Tokyo describe Genki as one of the most sought-after destinations for young researchers this year, competing directly with Preferred Networks and the corporate labs of Toyota and Honda for talent. Whatever the product turns out to be, the company has already shifted the labor market for robotics software in Japan.

The first product reveal, promised before the end of the year, will settle the most basic question. Rubin told Nikkei the device will not be a humanoid. That single detail rules out the industry's default form factor and suggests the company has found a wedge application where a simpler machine can ship sooner. Android launched on one phone before it ran on thousands. The pattern may be repeating.

Disclaimer: This article is for informational purposes only and does not constitute investment advice, endorsement, or verification of any claims made by the companies mentioned.