AI humanoid robots inch their way toward the workforce

AI-based humanoid robots could soon start becoming a workforce reality, thanks to parallel advances in generative AI and electromechanical components and mechatronics, enabling robots to have reasoning skills as well as physical abilities that exceed their predecessors.

“We’re still very early in the journey, but it’s only going to get better,” says Dwight Klappich, a research vice president at Gartner. “AI is really the brain driving humanoid robots like Agility, Tesla Optimus, and Boston Dynamics’ Atlas. The intelligence is actually going to evolve faster than the mechatronics.”

Mechatronics combines mechanics, electronics, and computers to create intelligent medical devices and robots, for instance. 

Boston Dynamics’ well known robotic dog ‘Spot’ was among the first advanced robots, and most use machine learning (ML) pattern recognition models. Today, key vendors — xAI, Meta, IBM, Boston Dynamics, Agility Robotics, Apptronik, Figure.ai, FourierIntelligence, and Sanctuary.ai — have plans to develop AI humanoid robots that can reason and adapt.

Analysts expect such robots to be commercially available for manufacturers, supply chain and logistics giants, and retail industries within two years.

Logistics giant GXO, for instance, has signed a multi-year agreement with Agility Robotics, and has deployed Agility’s Digit humanoid robot in a Spanx warehouse facility. It is also conducting pilot tests with Reflex Robotics and Apptronik humanoid robots. Figure.ai has signed two commercial contracts for its humanoid robots, according to the company. And xAI founder Elon Musk pledged that the company’s Optimus humanoid robots based on newly introduced Grok-3 foundation model — which was trained on xAI’s Colossus supercomputer  — will be operational in months, not years.

According to Gartner, by 2027 10% of all smart robots sold will be next-generation humanoid working robots. Hundreds of thousands of industrial robots have been in operation since the 1960s and many thousands of newer ‘smart robots’ that handle repetitive tasks are in production today. 

The AI robot proposition

One projected use case for humanoid robots will be taking over hazardous human duties. But while there is significant investment and innovation going on, Gartner’s Klappich sees “still fairly limited use cases, mostly proof of concepts.”

Within weeks, Rapid Robotics will unveil an industrial humanoid robot designed to “pick, pack, and palletize,” according to a company spokesman. The robot, called Rapid 3PRO, has been under development since September.

Meanwhile, Meta plans to make investments in humanoid robots through its Reality Labs’ hardware division to first target the consumer market, according to a report from Bloomberg. 

At a recent UAE conference in Abu Dubai, Musk claimed xAI’s next Grok-3-fueled Optimus robot will leapfrog his rivals. 

“With Tesla building the most advanced humanoid robot, it can be directed by deep intelligence at the data set level,” Musk said, noting that Grok-3 was trained with synthetic data and far more Nvidia horsepower than any other model. “You can [then] produce any product, provide any service. There’s really no limit to the economy at that point.”

There are more than a few naysayers who believe AI humanoid robots — in the factory or in the home — are still a ways off.  

“I do believe that humanoid robots or robots that have more human-like dexterity have potential within industrial operations. The idea of furthering human-robotic collaboration is easier if they both can operate the same set of tools. However, I don’t think that Musk’s claims of 2025 deployment are realistic,” says Carlos M. González, research manager of industrial IoT and intelligence strategies at IDC.

“Many car manufacturers launched pilot programs in 2024, but we are still in the research phase, primarily since it was determined that the robots Musk showcased were being piloted by human operators and not acting independently,” he adds. 

Meanwhile, the cost of humanoid robots is high, and the ability for mechatronics makers to achieve human-like sensory capabilities and mobile flexibility remains challenging. One vendor source who asked not to be named said AI safety issues still need to be addressed.

Outlook on deployments

Despite the ongoing hurdles, CIOs and consultants see promise for AI humanoid robots in manufacturing, warehousing, retail, hospitality, healthcare, and construction.

“Humanoid robots are advancing in dexterity, perception, and mobility at a rapid pace, but they are not yet at the reliability and cost-efficiency level of specialized automation like AMRs [autonomous mobile robots] or robotic arms,” says Chris Nardecchia, CIO of Rockwell Automation. “With Grok-3 and other generative AI models, these robots will improve in situational awareness and decision-making. Since these AI models are advancing rapidly, I would expect that this will translate into all robots advancing quicker than ever.”

Nardecchia adds that widespread adoption will depend on several factors — cost, reliability, safety, regulatory compliance, and enterprise integration — noting that AI robotics will face the same barriers to adoption as other emerging technologies, but he does not dismiss the possibility.

“Gartner’s prediction may be a bit optimistic and ambitious but not unattainable given the rapid advancements in AI, robotics, and edge computing,” Nardecchia says, adding that technical feasibility and economic viability will determine the technology’s viability.

For its Spanx implementation, GXO has deployed Agility Solutions’ multi-purpose, human-centric Digit robots and Agility Arc, its cloud automation platform for Digit fleets, at a warehouse for the womenswear brand, integrating the fleet with existing automation such as AMRs. Currently, Digit robots assist with repetitive tasks, such as moving totes and placing them onto conveyors.

Anders Brown, CEO of Tompkins Solutions, a systems integrator that has a partnership with Agility, says Agility, which has the manufacturing capacity to produce up to 10,000 robots a year, has deployed multiple Digit robots at various customer locations for production use as part of proof of concept or robot-as-a-service (RaaS) arrangements.

“The key to truly scaling beyond initial deployments is having a collaboratively safe robot, which Agility has made a pledge to release in 2026,” Brown says.

Tom Richer, CEO of AI consultancy Intelagen and a former CIO, pointed to Figure.AI robot, which has garnered significant attention and secured funding and a partnership with BMW for manufacturing applications. Along with xAI’s Optimus, Richer also noted AgiBot, a Chinese company that claims to have produced almost 1,000 units, and Boston Dynamics, whose humanoid robot Atlas has demonstrated impressive feats of movement and task execution, “but widespread deployment remains limited.”

“AI humanoid robots are still largely in early stages of development but making progress,” Richer says. “These examples, while not yet ubiquitous, demonstrate the accelerating progress in AI humanoid robotics, driven by advancements in machine learning and computer vision, and suggest that the 2027 [Gartner] prediction may well be within reach.”

Boston Dynamics will soon start testing its Atlas humanoid robot in Hyundai facilities this year, according to a company spokesman, who added that Boston Dynamics believes Atlas will have opportunities to move parts with complex sizes, shapes, and weights in other industries as well.

“Humanoids will be most effective if they are deployed with in-depth models of a facility and lots and lots of data about how it operates, and the increased focus on AI only helps advance the research and development of humanoid robots,” the spokesman said. “Many companies are jumping into the humanoid race, but not all of them will be successful and we’re excited for the competition, it means the industry is advancing and innovating.”

In terms of that increased focus on AI, Figure.AI this week announced Helix, what the company describes as “a generalist vision-language-action (VLA) model that unifies perception, language understanding, and learned control to overcome multiple longstanding challenges in robotics.”

Rockwell CIO Nardecchia says large-scale deployments will likely take longer than 2027, but pilot programs and niche use cases will increase within that timeframe.

“I would anticipate some industries, which are particularly impacted by labor shortages, partnering with robotic technology companies and making investments, or even using their own manufacturing capabilities to jump the line to leverage these technologies sooner than their competitors,” the CIO says.

The batch of AI humanoids coming within the next few years — if technically achieved — will be a significant advance for all of human society, Gartner’s Klappich notes.

“Humanoid robots imitate the human body with a head with sensors and cameras for sensing its environment; a body that houses the power and mechanicals; arms and hands/grippers for grasping, manipulating, and carrying items; and legs for dynamic locomotion all powered by an AI-enabled brain,” Klappich wrote in a Gartner supply chain trends report published last March.

“While there are many examples of highly productive mobile robots that use wheels, there have been few successful examples of robots that offered mobile manipulation, such as moving around and picking up and carrying things, and that could handle varied or unpredictable tasks and terrain,” Klappich wrote. “The next generation of humanoid working robots will combine sensory awareness with mobile manipulation and dynamic locomotion to perform productive work that was previously relegated to biological human.”