Robotis Humanoid Software Pivot Backed by 88.7 Billion Won

Robotis is committing 88.7 billion won, about $65 million, to a humanoid software pivot, moving the Korean company away from the actuators and hardware platforms it has sold for more than two decades and toward the layer where robot fleets are managed, trained and deployed. The commitment, disclosed on September 8, 2026, draws in part on the sale of the chief executive's personal stake in the company. Part of the proceeds is earmarked for Robotis AI, the subsidiary that runs the group's autonomous mobile robot business.

The scale is small next to the money moving through the sector. Humanoid robotics startups have drawn $8.7 billion in venture funding in 2026 through July, already double the full-year record set in 2025, according to Dealroom. Figure AI alone raised $1 billion to fund training-data collection and production capacity. Robotis is financing its shift largely through an insider transaction rather than a venture round, which caps how much it can spend but avoids the dilution that comes with outside capital.

Why a hardware maker is buying into software

Robotis built its reputation on precision components: actuators, servo modules and the humanoid platforms that researchers use as testbeds. Hardware of that kind has carried most of the industry's value. Sector analysis drawing on Grand View Research data puts hardware at 69.7% of the humanoid market in 2024, with software taking the remainder.

That split explains both the risk and the logic of the move. Hardware margins compress as Chinese suppliers scale production, and the Chinese market is crowded with vendors pushing toward mass adoption. The contracts that recur are in fleet management, simulation, model training and on-site deployment support. Those are the categories Robotis is now chasing, and they are the same ones Korean and Chinese vendors are contesting.

OpenMind chief executive Jan Liphardt has argued that the hardware problem is largely solved, with low-cost, reliable humanoid platforms now available. If that reading holds, differentiation moves to whoever can keep a fleet working reliably across many sites. Robotis is betting that its hardware heritage gives it an edge in knowing what those software contracts have to control.

Where Robotis sits against Korea's giants

Robotis is not the only Korean company moving on this ground, and it is by far the smallest. Samsung has placed its humanoid hardware and AI work under a single software chief ahead of a CES reveal, with group investment in robotics and physical AI estimated at roughly 60 trillion won, or about $44.6 billion. Hyundai Motor Group showed an electric Atlas humanoid earlier in 2026, and Samsung later hired that program's former strategy lead, Lee Dongkun. LG Group has its CLOi'd platform.

Public money reaches further than headline capital. The government program funds multi-articulated, tactile-sensing hands for industrial assembly, homegrown neural processing units for on-device perception, and sensor arrays built for low onboard power budgets, with a 2031 milestone for unifying hardware and compute. Machines from the program will be placed with universities and government-funded research institutes for testing. For Robotis, that is an early, accessible market for both components and the fleet software it is now assembling.

Chinese competition sets the price ceiling. Unitree is filing for an initial public offering at a $7 billion valuation carrying 60% gross margins, which shows how fast a hardware maker can reach scale once component costs fall. Robotis cannot match that volume, and the software layer is attractive partly because it is not priced per unit. Component suppliers at home are scaling too: Wonik Robotics has pushed a 350 billion won capital raise.

Deployment evidence is arriving in parallel. BMW validated Figure 02 across 1,250 operational hours on an active automotive line and is expanding to Leipzig. Japan Airlines is putting robots to work at Haneda Airport, and Amazon is building a consumer physical AI platform through Fauna. Those installations create demand for the fleet software Robotis wants to sell, and they also give larger competitors a proving ground that a mid-cap supplier lacks.

The trade-offs in the humanoid software pivot

The case for Robotis is focus. A company that has shipped actuators to robotics labs for years understands the failure modes that software teams discover late: joint wear, calibration drift, thermal limits on repeated cycles. Selling software that anticipates those problems is a credible differentiator, and recurring licence revenue would be worth more than component sales.

The case against is arithmetic and timing. At roughly $65 million, Robotis has less than one-fiftieth of Samsung's stated robotics commitment. Software revenue takes years to build, while the hardware business that funds it may shrink as attention shifts. Hiring machine-learning engineers in Seoul means bidding against Samsung, Hyundai and LG for the same people.

Capital is arriving at a pace that leaves little room for a slow transition. The sector recorded 25 disclosed equity rounds worth $6.20 billion across 20 companies in the eleven months to June 2026, an average of 2.27 rounds per month, and roughly $34 billion flowed into robotics in 2025. Humanoid, a robotics startup, raised $152 million in a Series A that valued it at $1.35 billion. Robotis is competing for partners and engineers against companies that raise more in one round than its entire pivot budget.

A structural risk sits underneath. Foundation-model developers are pushing toward general-purpose robot policies, and the spread of market forecasts reflects how unsettled the endpoint is. Morgan Stanley projects more than one billion humanoids by 2050 in a $5 trillion market, while ABI Research sees $6.5 billion by 2030 and IDTechEx $30 billion by 2035. If generalist models absorb the integration work, the middleware layer Robotis is entering could be squeezed between cheap hardware and widely licensed models.

That range is itself the finding. A market sized anywhere between $6.5 billion and $5 trillion is one where no supplier has settled the business model, and where a focused player can still take a defensible position.

What to watch

Timing tightens the test. Samsung's CES reveal and Korea's 2030 mass-production targets both fall inside the window in which Robotis must prove software traction. Should those larger programs settle on in-house stacks, third-party fleet software loses the neutral position it needs to sell across customers.

Three signals will show whether the humanoid software pivot holds. The first is whether Robotis AI names paying deployment customers rather than pilot projects. The second is whether the chief executive's stake sale is followed by an external raise, which would show that the software roadmap costs more than insider proceeds can cover. The third is Samsung's CES presentation, where a unified hardware and software stack would set the price expectations for everything Robotis sells.

Until those land, the honest read is that Robotis has chosen the right layer of the stack at the wrong scale, and is relying on domain knowledge to close a capital gap it cannot match head-on.

Why this matters

Robotis shows how the humanoid race is separating into two contests: one over hardware volume, dominated by conglomerates and Chinese manufacturers, and one over the software that keeps machines productive after installation. Buyers evaluating robotics vendors should expect the second contest to decide which suppliers survive the first wave of deployments. For Korean robotics, the risk is that its largest players win the hardware story while the software value accrues elsewhere.

Sources

Scaling the humanoid robotics supply chain into billion-dollar wins | McKinsey

Photo by Enchanted Tools on Unsplash

Related Articles

• How the Hyundai Boston Dynamics Acquisition Reshapes Physical AI Strategy
• Humanoid Becomes Europe's First Humanoid Robotics Unicorn with $152M Series A
• Hyundai's Atlas Humanoid Robots: The Unit Economics of a 25,000-Unit Factory Fleet

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.