Humanoid Global Bets on Defense and China as Robotics Industry Nears Scale

After a 2,950% surge, Humanoid Global loses 21% amid robotics sector news. The firm pivots to defense, eyes China supply chains, as Bank of America forecasts 86% CAGR for humanoids.

Humanoid Global Holdings is navigating a sharp turnaround. The Vancouver-based investment company, fresh off a staggering 2,950% annual share price surge, has lost roughly 21% over the past month — including a 2.17% slide last week. With a beta of 2.42, the stock’s extreme sensitivity to sector news means every development in the robotics ecosystem hits hard. And this week brings a trio of them: the Automate trade fair in Chicago, fresh signals from Tesla’s production timeline, and the looming IPO of Chinese rival Unitree.

Behind the near-term noise lies a broader strategy shift. Humanoid Global is actively hunting for an acquisition or joint venture in the defence and security space, moving beyond its original remit of broad humanoid robotics and embodied AI. The logic is straightforward: global military spending is projected to hit $6.6 trillion by 2035, creating a nascent market for robots that can replace personnel in hazardous environments and protect critical infrastructure. The company’s CEO, Shahab Samimi, has also set his sights on China, where state-backed ambitions are reshaping the competitive landscape.

“China is world-leading in scalable and cost-efficient production of humanoid robots,” Samimi said in a statement, pointing to a national fund of $138 billion dedicated to AI and robotics. The country already ships 30 times more robotics units than the United States. That production depth is why Humanoid Global’s portfolio is increasingly tilted toward Chinese and Asia-based supply chains — even as its investments remain modest in size.

In January, the firm put CAD 150,000 into MBody AI, a startup building a hardware-agnostic software platform that gives robots perception, decision-making and coordination. A month later, it chipped in $75,000 for HowToRobot’s Series A round. The sums are small for a microcap story, but they signal a bet on the software layer that separates laboratory curiosities from industrial tools.

That industrial shift is already underway elsewhere. At BMW’s Spartanburg plant, robots from Figure have helped assemble more than 30,000 vehicles, and the pilot is now extending to Leipzig. Toyota Canada uses Agility’s units for material transport. Japan Airlines signed a three-year contract in May to deploy humanoids at Tokyo Haneda Airport. According to IDTechEx, the amortisation period for these machines can shrink to roughly six months at high utilisation rates.

Bank of America Global Research provided the long-range scaffolding for the sector in March 2026. It forecasts annual deliveries rising from 20,000 units in 2025 to 10 million by 2035 — a compound annual growth rate of 86%. By 2026, it expects 90,000 units shipped; by 2030, 1.2 million. The bank’s even longer-term vision sees a global population of 3 billion humanoid robots by 2060, exceeding the current fleet of 1.5 billion cars, with 62% working in private households.

The immediate catalysts, however, are more tangible. Tesla is converting Model S and Model X assembly lines at its Fremont factory for humanoid production, with first output slated for late summer 2026. Meanwhile, Unitree — which delivered more than 5,500 units last year, outshipping all other humanoid providers combined — is preparing a $610 million IPO after a 335% revenue jump. It began selling its G1 model on Amazon US for $17,990 in February.

Both rivals underscore the competitive pressure on Humanoid Global’s investment thesis. The company’s portfolio is a collection of stakes, not operational assets; its value hinges on whether the broader sector narrative translates into real industrial momentum. Bessemer Venture Partners captured the mood in May by calling the moment “the GPT-2.5 moment for robotics” — capabilities are real, scaling laws are emerging, but the distance to 99.9% production reliability remains vast.

Bottlenecks persist. Actuators, force and tactile sensors, and dexterous hands still limit mass adoption. China dominates manufacturing cost and volume, while the US and Europe take the lead in AI and safety-certified applications. The European Union will require mandatory safety certification for commercial robots by the third quarter of 2027, setting a regulatory wedge that could separate winners from also-rans.

Against this backdrop, Humanoid Global’s dual play — defence M&A and deeper China exposure — aims to place the company in the fastest-growing niches of a sector that is still writing its playbook. The highly volatile stock has already priced in a great deal of optimism. The upcoming Automate forum, running June 23–24 in Chicago with more than 50,000 attendees and 1,000 exhibitors, will test whether the market’s faith is justified. Speakers from Boston Dynamics, Neura Robotics, NVIDIA and the Toyota Research Institute are expected to offer new data points on commercial viability. Behind the scenes, world-model developers raised roughly $6 billion in the first quarter alone, ensuring that capital is not the constraint.

For Humanoid Global, the next few weeks are less about quarterly results and more about the story holding together. If Tesla hits its ramp-up, Unitree prices its IPO successfully, and the defence deal materialises, the stock’s recent pullback could be a pause before another leg. If not, the 2,950% annual gain leaves a long way to fall.