Author: Rinat Mirzaitov

Morgan Stanley has sharply raised its 2026 forecast for China’s humanoid robot shipments, giving the market a stronger near-term signal that Chinese manufacturers are moving faster than earlier expected. The development matters because it suggests accelerating production, policy support, and supply-chain activity in China’s humanoid robotics sector. It does not, however, prove that humanoid robots are already delivering reliable, paid, repeatable work in customer environments. CNBC reported that Morgan Stanley now expects 50,000 humanoid robots to ship in China this year, up from a previous forecast of 28,000 units and an initial January estimate of 14,000 units. South China Morning…

Kepler, Unitree and Fourier are showing a different side of China’s humanoid robotics push: productized platforms are becoming as important as headline-grabbing demos. Each company is making humanoids available, or at least commercially positioned, for developers, researchers, industrial partners and early adopters. That matters because broader access can accelerate experimentation, data collection and ecosystem formation, even if it does not yet prove that the robots are ready for repeat customer deployment. The confirmed signal is that these companies are packaging humanoid robots as products rather than only research projects. Unitree lists low-cost humanoids with published pricing and open control interfaces.…

Alibaba’s Qwen team has introduced Qwen-Robot Suite, a set of three robotics foundation models aimed at navigation, manipulation, and physical world prediction. The launch matters because it shows a major AI platform supplier moving from language and multimodal models into systems that could eventually support robots operating in physical environments. What it does not yet show is customer-verified deployment, paid usage, humanoid robot readiness, or reliable task performance in production settings. The linked X post focuses on Qwen-RobotWorld, which Alibaba describes as a world model for physical agents using a single model across more than 20 embodiments, a natural-language action…

Boston Dynamics’ electric Atlas is entering the phase where technical credibility must turn into commercial proof. The company has unveiled a product version of Atlas, scheduled 2026 shipments to Hyundai and Google DeepMind, and tied the robot’s future to Hyundai manufacturing and Google’s robot AI models. The open question is whether Atlas can move from one of robotics’ most impressive engineering programs into useful, repeatable factory workflows. The confirmed signal is meaningful. Boston Dynamics says it will begin manufacturing the product version of Atlas at its Boston headquarters, with 2026 fleets scheduled for Hyundai’s Robotics Metaplant Application Center and Google…

Apptronik has raised more than $935 million across an unusually large Series A, giving the Austin-based robotics company substantial resources to manufacture, train, and deploy its Apollo humanoid robot. The financing, supported by Google, Mercedes-Benz and several industrial investors, strengthens Apptronik’s competitive position, but it also raises the standard of evidence the company must now meet. The funding is confirmed. Apollo’s broader commercial readiness is not. Apptronik has named enterprise relationships with Mercedes-Benz, GXO Logistics and Jabil, and Mercedes-Benz has publicly confirmed that it is testing a small number of Apollo robots in production environments. However, there is still no…

Figure AI’s most interesting signal from Brett Adcock’s recent podcast appearance was not the package-sorting livestream, the Figure 4 teaser, or the company’s confidence about future home robots. It was a brief story about a fridge task, in which Adcock said a Figure robot’s success rate improved sharply after the model was trained on additional data from unrelated tasks, a claim that speaks directly to the central commercial question in humanoid robotics: whether Physical AI can generalize across the messy real world. The claim should be treated carefully. Adcock described an internal Figure evaluation, not an independently verified benchmark, customer…