The Rise of Humanoid Robots: Where We Are Now and What Comes Next
By Tecstasy Contributor
Published on 11 January 2026
Humanoid robots have long occupied a special place in our collective imagination. From early science fiction to polished marketing videos, the idea of a machine that walks, sees and works like a person has been a powerful symbol of technological progress. Until recently, this was more fantasy than factory reality. Over the past few years, however, a convergence of advances in robotics, sensors and artificial intelligence has pushed humanoid robots out of the lab and into warehouses, car plants and retail environments. Major firms are now racing to turn these systems into scalable products that can address labour shortages, support ageing populations and unlock new forms of automation.
This article explores what makes a robot “humanoid”, why industry is suddenly paying serious attention, the leading projects around the world, and the likely social, economic and ethical implications of deploying machines that work alongside us in human spaces.
What Exactly Is a Humanoid Robot?
In technical terms, a humanoid robot is a machine whose body plan and capabilities are designed around the human form. Typically this means a bipedal platform with a torso, a head housing sensors, arms with articulated hands and a control system capable of balancing, walking, grasping and manipulating objects in environments built for people. Classic research platforms such as Honda’s ASIMO and the early generation of bipedal robots were highly controlled prototypes that demonstrated impressive motion, but they were not truly general purpose workers.
Today’s humanoids combine this physical architecture with far more powerful computing and machine learning. Modern systems use advanced sensors such as depth cameras, LiDAR and tactile arrays, along with on board or edge computing, to localise themselves, map their environment and plan movements in real time. They also draw heavily on breakthroughs in reinforcement learning, imitation learning and large scale vision and language models. This shift from rigid programming to data driven learning is central to the current wave of interest, since it raises the prospect of robots that can be re trained quickly for new tasks instead of being hard coded for narrow roles.
The appeal of the humanoid form is largely pragmatic. Factories, warehouses, hospitals and homes are all designed around human bodies. Doorways, stairs, shelving and tools follow human dimensions. A robot that can walk where people walk and use the tools people use can, in principle, take over a wide variety of jobs without requiring a complete redesign of the physical workspace.
Why Humanoids Are Emerging Now
Several forces have combined to make humanoid robots commercially interesting rather than purely speculative. On the demand side, ageing populations and persistent labour shortages in logistics, manufacturing and care work mean that businesses and governments are actively searching for ways to do more with fewer workers. Analysts expect the industrial and service robotics markets to grow strongly through the late 2020s, driven by the need to automate physically demanding and repetitive tasks in warehouses, factories and hospitals.
On the supply side, core technologies have matured. High performance electric actuators are now lighter and more efficient. Battery energy density has improved, making multi hour operation feasible. Cloud connectivity allows constant software updates. Crucially, advances in AI are starting to give robots the ability to understand scenes, reason about tasks and adapt to new conditions rather than simply replaying a fixed script. The integration of models like Google’s Gemini into robotic platforms is an early example of this trend, offering robots richer perception and decision making capabilities.
As a result, companies are beginning to treat humanoids as a plausible way to automate some of the dull, dirty and dangerous work that remains stubbornly difficult to mechanise with traditional industrial robots or specialised machines.
Leading Humanoid Platforms and Their Progress
Tesla’s Optimus: General Purpose Ambitions
Tesla’s Optimus project is one of the highest profile attempts to create a general purpose humanoid. According to Tesla, the goal is a bi pedal autonomous robot that can perform “unsafe, repetitive or boring tasks”, with long term aspirations that stretch from factory work to domestic chores. The company’s AI and Robotics division describes Optimus as a platform that leverages the same software stacks used for autonomous driving, adapted to balance, navigation and manipulation in the physical world.
Public demos and technical updates through 2024 and 2025 showed Optimus moving from basic walking to more complex pick and place operations, sorting objects and performing simple assembly tasks. Reports suggest that Tesla aims for mass production in the mid 2020s, targeting a unit cost that would be closer to an expensive car than a one off research prototype. While timelines in robotics are famously slippery, the Optimus program illustrates how automotive style manufacturing and vertically integrated AI development can be used to pursue humanoids as scalable products rather than exotic experiments.
Boston Dynamics, Hyundai and the Atlas Production Push
Boston Dynamics, famous for its viral videos of robots performing parkour, has long been a benchmark for legged locomotion. After being acquired by Hyundai, the company has pivoted towards industrial applications. At CES 2026, Hyundai and Boston Dynamics unveiled an all electric Atlas humanoid in its first live demonstration and confirmed plans to deploy production versions on Hyundai’s factory floors.
Hyundai has announced an ambitious plan to manufacture around 30,000 humanoid robots per year by 2028 to support vehicle assembly, initially at its Savannah electric vehicle plant in the United States. These Atlas units are designed to handle parts sequencing, material handling and other physically demanding tasks, with features such as high payload capacity, operation in extreme temperatures and the ability to autonomously swap their own batteries. The integration of Google DeepMind’s Gemini model is intended to give Atlas richer contextual awareness, so it can reason about instructions, understand scenes and adapt to dynamic factory conditions.
Agility Robotics’ Digit: Humanoids in Warehouses
Oregon based Agility Robotics has taken a slightly different approach with its Digit robot. Digit is a legged, torsoed machine with arms, designed specifically for bulk material handling in warehouses. It can walk through facilities, pick up and move totes, and work in spaces built around human workflows. Agility’s partner program promises early customers deliveries from 2024, with general availability planned for 2025.
Amazon has been one of the most visible early adopters. The company began testing Digit in its fulfilment centres in 2024 as part of a broader automation strategy focused on worker safety and throughput. In 2024, logistics provider GXO also started deploying a fleet of Digits in a Spanx warehouse under a robots as a service arrangement. These deployments suggest a near term future in which humanoid robots are used as flexible “legs and arms” to complement fixed conveyor systems and traditional robotic arms, particularly for tasks like trailer unloading and tote movement where environments are cluttered and layouts change frequently.
Figure 01 and the BMW Collaboration
California startup Figure is pursuing a fully autonomous general purpose humanoid called Figure 01. In 2024 the company announced a commercial agreement with BMW Manufacturing to deploy humanoid robots in the car maker’s Spartanburg plant, in what was widely described as a first of its kind deal for automotive production.
By mid 2024 and into 2025, Figure released videos showing its robot carrying out tasks in an assembly environment, and media coverage highlighted that a Figure robot had started work on a BMW assembly line, powered by OpenAI based control systems. The partnership uses a staged, milestone based approach, reflecting the reality that complex, safety critical manufacturing processes are unlikely to be handed over to robots in a single leap. Instead, humanoids are being trialled on clearly scoped tasks, under close supervision, with human workers still central to the line.
Sanctuary AI’s Phoenix: Learning New Tasks in Hours
Vancouver based Sanctuary AI positions its Phoenix robot as a general purpose humanoid worker intended to address global labour challenges. The company’s mission statement emphasises a vision of “millions of industrial grade humanoid robots” deployed across sectors, from logistics to retail. In 2023 Sanctuary reported its first commercial deployment in a retail environment, and in 2024 it unveiled a seventh generation Phoenix that had completed more than 110 distinct retail tasks during a one week pilot, including shelf restocking and cleaning activities.
Perhaps most striking are Sanctuary’s claims that Phoenix can automate new tasks in less than 24 hours through a combination of human teleoperation, imitation learning and data driven generalisation. This approach blends human expertise with machine learning in a tight loop. People demonstrate tasks via remote control, the system generalises those demonstrations into reusable policies and the robot then executes them autonomously. If such methods scale, they could dramatically reduce the cost and time needed to “teach” humanoids new jobs.
Apptronik’s Apollo and Other Industrial Players
Another prominent entrant is Apptronik’s Apollo, a human sized humanoid built with mass manufacturability and human friendly interaction in mind. Apollo stands around 1.7 metres tall, can handle payloads of roughly 25 kilograms and runs for several hours on a swappable battery pack. Apptronik markets Apollo for use in logistics, manufacturing and retail, promoting applications such as palletising, case picking and line replenishment.
The broader industrial ecosystem is also heating up. Sanctuary’s progress has attracted significant investment. Automotive supplier ecosystems are forming around companies like Hyundai and Tesla. Even firms known for automotive perception and driver assistance, such as Mobileye, are joining the race. In early 2026, Mobileye announced a plan to acquire humanoid startup Mentee Robotics for around 900 million US dollars, with the explicit aim of repurposing its self driving sensing and perception stack for embodied robots that can be trained from a single human demonstration.
From Demos to Deployment: Likely Paths Over the Next Decade
Despite the flurry of announcements and polished videos, humanoid robotics is still in its early commercial stages. Over the next five to ten years, development is likely to follow a staged path rather than an overnight transition to fully autonomous robot co workers.
In the near term, most deployments will be in relatively structured industrial settings: warehouses, logistics hubs and automotive plants. These are environments where tasks such as moving totes, loading pallets, sequencing parts and performing repetitive assembly steps can be carefully scoped, while safety systems and processes are refined. Robots will often be deployed under robots as a service contracts, where customers pay for hours of work rather than buying hardware outright. That model, already used for Digit and Apollo in logistics, allows companies to experiment without massive upfront capital expenditure.
As hardware reliability improves and AI systems gain more robust perception and reasoning, humanoids are likely to tackle more varied tasks within those same environments. For example, instead of handling only empty totes, a robot might be trusted to manage mixed, fragile items or to assist directly on flexible production lines. The integration of multimodal AI models such as Gemini is likely to accelerate this, since robots will be better able to interpret verbal instructions, process documentation and adapt to novel configurations.
Beyond factories and warehouses, there is strong interest in deploying humanoids in healthcare, aged care and hospitality. These settings are more challenging from both a technical and ethical perspective. Perception must be extremely robust, safety standards are higher and people are in vulnerable situations. It is therefore likely that humanoids will first appear in tightly defined support roles, such as out of hours logistics in hospitals, basic cleaning in aged care homes or back of house tasks in hotels, with human staff retaining responsibility for direct patient and customer interactions for some time.
Economic, Social and Labour Market Impacts
Whenever automation touches physical work, debates about jobs and inequality follow. Humanoid robots intensify these debates because they are designed to step into roles traditionally filled by people, using the same doorways, tools and workflows. It is reasonable to expect that widespread adoption in sectors such as warehousing and manufacturing will displace some manual roles, particularly those involving repetitive lifting and moving of objects.
At the same time, many of the companies investing in humanoids present them explicitly as tools to enhance safety and productivity rather than direct replacements. Amazon’s robotics program, for instance, has emphasised that systems like Sequoia and Digit help reduce strain injuries and allow employees to move into higher value roles supervising and maintaining robots. GXO describes its partnerships with Apptronik and Agility as part of a strategy to address labour shortages while scaling operations.
History suggests that both narratives can be true. Some categories of jobs will shrink or vanish, while new roles emerge in robot fleet management, maintenance, simulation, data annotation and process engineering. Regions that prepare workers through accessible retraining, vocational education and pathways into robotics aligned trades are likely to see better outcomes than those that simply hope the market will adjust on its own. For countries like Australia that face demographic ageing and high labour costs in certain sectors, humanoid robots may become one of several tools used to maintain service levels without relying solely on migration or increased working hours.
Ethical, Legal and Cultural Questions
The deployment of humanoid robots raises a cluster of ethical and regulatory questions that go beyond the purely technical. Safety is the most immediate concern. Large, mobile machines operating in close proximity to people must meet stringent standards around collision avoidance, emergency stops and fail safe behaviour. The field of “robot safety engineering” is evolving quickly, blending traditional industrial safety principles with AI specific concerns about unpredictable behaviour, dataset bias and model brittleness.
Liability is another open issue. If a humanoid robot controlled by a cloud based AI model causes harm, is responsibility borne by the hardware manufacturer, the AI provider, the integrator or the operator who configured the task? Existing product liability and workplace safety laws provide some guidance, but regulators are increasingly exploring robot specific frameworks. Organisations deploying humanoids at scale will need to treat legal compliance and transparent incident reporting as core governance functions, not afterthoughts.
Culturally, humanoid robots will force societies to make choices about how much human likeness is desirable. There is a spectrum between highly functional but clearly mechanical devices and machines that mimic human appearance and social cues. In some contexts, a friendly humanlike presence may be welcomed, particularly if robots are assisting in care settings. In others, an overtly human facade could trigger discomfort or unrealistic expectations. Designers will need to consider not only what is technically possible, but what is psychologically and socially appropriate for specific communities and use cases.
Conclusion: Humanoids as a New Layer of Infrastructure
Viewed through a sober lens, humanoid robots are unlikely to become universal, general purpose “mechanical people” in the short term. Their strengths and weaknesses will be shaped by physics, economics and the messy realities of real world environments. Yet the progress of the past few years suggests that humanoids are poised to become an important new layer of infrastructure across logistics, manufacturing and specific service sectors.
As companies like Tesla, Hyundai, Agility Robotics, Figure, Sanctuary AI, Apptronik and Mobileye push the boundaries of what these machines can do, policymakers, educators and communities have a narrow but important window in which to shape how the technology is deployed. The key questions are not only “what can humanoid robots do” but “what should we ask them to do”, under what safeguards, and in service of which social and economic goals.
For technologists, investors and informed citizens, the emergence of humanoid robots is therefore not just a robotics story. It is a preview of a world in which embodied AI will share our workspaces and, eventually, parts of our everyday lives. The choices made over the next decade will determine whether that world feels empowering, equitable and humane, or whether it deepens existing divides. The technology is arriving. The real work now is to decide how we will live with it.