Humanoid robots are no longer just a figment of science fiction. Imagine a world where robots not only assist in factories but also greet us in stores, aid in surgeries, and provide care for our loved ones. With Tesla planning to deploy thousands of its Optimus robots by 2026, the age of humanoid robots is rapidly approaching.
This vision is becoming increasingly tangible as companies and research teams unveil groundbreaking innovations in the field of robotics. At the 2025 Consumer Electronics Show (CES), for instance, products like ADAM, the robot bartender from Richtech Robotics, and puppy dogs from Tombot Inc., designed to comfort elderly adults with dementia, showcased the progress made in both functionality and human-centric design. While these robots may still be in the early stages of deployment, they illustrate the vast potential for humanoid robots to enhance various aspects of human life.
Even with these promising advances, experts caution that achieving fully human-like capabilities is still a distant goal. Renowned AI researcher Yann LeCun, one of the “Godfathers of AI,” has pointed out that current AI systems are still far from reaching the level of intelligence required to understand the physical world, plan, or reason in a human-like manner. LeCun argues that smart robots, as we envision them today, may not yet be achievable due to these fundamental gaps in AI capabilities.
Despite this, the future of humanoid robots seems bright. Elon Musk recently stated that Tesla plans to produce thousands of Optimus units by 2025, with an expected 50,000 to 100,000 units to be shipped in 2026. This marks a dramatic leap from the small number of units currently in operation, most of which are designed for specific functions. While Musk has a history of setting ambitious timelines—like his 2016 prediction that fully autonomous driving would be achieved within two years—his comments reflect the optimism surrounding humanoid robot development.
Tesla is not alone in this race. Companies such as Agility Robotics, Boston Dynamics, and Figure AI are all leaders in the humanoid robotics field. Recently, Agility Robotics’ CEO Peggy Johnson shared that it will soon be “very normal” for humanoid robots to become coworkers across various industries. Figure AI, with backing from major investors like Microsoft and Nvidia, announced that its F.02 humanoid robots are already hard at work for commercial clients.
However, despite these advancements, many challenges remain before humanoid robots can function seamlessly in dynamic, real-world environments. The current generation of robots faces three key hurdles: processing visual information quickly enough for real-time reactions, understanding subtle human behavioral cues, and adapting to unexpected changes in their surroundings. Many humanoid robots rely heavily on cloud computing, which often results in latency and limits their ability to perform tasks, such as picking up objects, efficiently.
To address these limitations, one innovative company is World Labs, founded by “AI Godmother” Fei Fei Li. Li emphasizes that current robotic systems struggle to understand and interact with the physical world due to their reliance on basic visual processing. World Labs is tackling this challenge by developing spatial intelligence—a technology that helps robots map their surroundings in real time and predict how objects might move. Spatial intelligence mimics human cognition, allowing robots to interact with dynamic environments more effectively and with greater adaptability.
One notable project at World Labs is Generative World Explorer (GenEx), a system that uses AI to infer the structure of the physical world from a single image, much like how humans make inferences about their surroundings. Although still in the research phase, this capability would help robots navigate new and unfamiliar spaces, making quick decisions based on limited data. For instance, it could allow humanoid robots to adapt to environments they’ve never encountered before, significantly improving their performance in situations like search-and-rescue missions.
As companies push the boundaries of AI and robotics, Nvidia is making significant strides with technologies like Cosmosand GR00T. Cosmos, a family of AI “world foundation models,” helps robots understand physics and spatial relationships, while GR00T (Generalist Robot 00 Technology) allows robots to learn by observing humans, just as apprentices learn from masters. These advancements help humanoid robots not only understand what to do but also how to do it naturally, further enhancing their potential for real-world applications.
Nvidia’s work in this space is already showing promise for sectors like healthcare and disaster response. For example, GR00T could enable robots to assist medical professionals by learning from their actions, while GenEx might help robots navigate hazardous environments with minimal sensory input. Nvidia’s CEO, Jensen Huang, recently predicted that the “ChatGPT moment for robotics” is fast approaching, referring to a moment when AI’s ability to reason and adapt transforms the robotics landscape just as it has revolutionized natural language processing.
Google DeepMind is also advancing the field of physical AI, with plans to develop large-scale generative models that simulate the physical world. These models will be crucial for improving robots’ abilities to predict, plan, and learn from their experiences, all essential for humanoid robots to operate autonomously and efficiently.
The next decade promises to be a pivotal period for humanoid robots. With ongoing innovations in AI, spatial intelligence, and real-time learning, the dream of robots that can adapt to complex environments and interact meaningfully with humans is becoming more achievable. While challenges remain, including the need for faster processing, better sensory integration, and improved human-robot interaction, the progress being made suggests that humanoid robots will soon play an increasingly prominent role in everyday life.
As the robotics industry continues to evolve, humanoid robots equipped with advanced cognitive and physical adaptability will not just work alongside us—they will transform the way we live, work, and care for each other. The age of humanoid robots is upon us, and the future is looking more collaborative and intelligent than ever before.
By Impact Lab