The race for humanoid robots has entered its final stretch: Reports from China describe an unusually large order of linear actuators, hinting that Tesla is approaching the completion of its "Optimus" humanoid robot. Meanwhile, a recently circulated video showed the robot performing kung fu exercises in real time, responding to pushes and making small jumps to maintain balance. The combination of large-scale procurement and real-time performance upgrades suggests that significant production may begin as early as next year. Competitors are responding intensely, demonstrating robots for manufacturing plants, warehouses, and demonstration kitchens, turning the field from a theoretical competition into a fight over real production lines.
A large order of components from the Chinese industry provides the first sign that Optimus’ industrial design is taking shape, turning the futuristic scenarios seen so far only in science fiction films into reality. According to reports, the robot components order indicates preparation for a production line, not just continued production of prototypes previously used for demonstrations. Tesla maintains communication ambiguity, noting no official information is available for release, but the fact that this involves an existing supplier and timing that coincides with hardware and software updates strengthens the assessment that the transition from interim models to a version approaching initial production is moving forward.
Behind the scenes, Tesla has faced this year an engineering bottleneck familiar to almost anyone working with humanoid robotics: The hands. Fine dexterity, wrist joint thermal resilience, and lifespan under varying loads over long shifts all dictate the pace of progress. According to industry sources, in recent months Tesla has converged on a new hand configuration with multiple degrees of freedom, more efficient motion, and improved force and touch sensors. If this configuration proves durable and precise, the order for linear actuators may also reflect a leap from experimental shelving to production shelving.
Alongside hardware preparation, Tesla shows software progress. The video in which Optimus practices kung fu demonstrates continuous real-time motion, entering combat stance, responding to pushes, and recovering without falling. However, the combat demonstrations are a testing stage rather than a goal, as they examine critical capabilities such as touch management, weight transfer, slip prevention, and precise timing of grip points on the floor. Compared to earlier videos where motion speed was accelerated in editing or remotely controlled, there is now a clear effort to demonstrate autonomous control that reads sensors, processes vision, and generates joint commands in real time.
The perceptual core of Optimus relies on an array of cameras and internal sensors, enabling it to build a three-dimensional model of the environment, recognize humans and objects, and track them. Learning modules map short verbal instructions to sequences of skills, such as approaching a shelf, assessing weight, selecting a grip, lifting, moving, and precisely placing an object. The control layer executes this sequence into thousands of micro-corrections per second to maintain stability.
Regarding its mechanical body, Optimus is built to integrate into a world designed for humans. Its chassis is lighter than previous models, allowing it to carry medium loads without fatigue, and the dense electric joints provide stability on both smooth and rough surfaces. The engineering behind the joints is designed to withstand repeated loads over long work hours without overheating. But the next big question is not just how strong the robot is, but how fast, delicate, and durable it will be—and, most importantly, how much maintenance will actually cost compared to a human worker.
Around Tesla, a dense field of competitors is emerging, moving their models from lab to field. Figure AI, for example, has already deployed robots on automotive production lines and focuses on integrating motion with language comprehension. U.S.-based Agility Robotics specializes in bipedal robots for logistics and operates a new production line in Oregon. In Texas, Apptronik is developing Apollo, currently tested by large industrial and logistics companies. Veteran Boston Dynamics upgraded Atlas to a more flexible electric version but still presents it as a demonstration rather than a shelf product. In China, Jack Ma’s Ant Group showcased R1, a robot that cooks and serves food at tech events, though at a pace that illustrates how far we are from a chef robot to a real work robot. Unlike them, Tesla benefits from data depth, software based on its autonomous car experience, and the ability to develop everything in-house, yet competitors are already closing commercial deals and presenting more focused solutions.
Elon Musk presented Optimus as a combination of a grand vision and an almost stubborn work methodology, as a solution for repetitive, dangerous, or simply boring work—or essentially, a robot that replaces humans where possible. The next phase is just around the corner. To move from experimental to serial production, Tesla will need a stable supply chain, extended durability tests, and proof that the robot can work a full day without interruption. Then comes the big question: The cost. Musk has already hinted that the target price is lower than a car, around $20,000 to $30,000 per unit. Initially, large industrial companies will receive the robots. Home versions, which could help with folding laundry or cleaning, will arrive later, once stability is proven outside the lab. Nevertheless, the humanoid robot is no longer a futuristic dream but one step away from becoming a reality integrated into all our lives.