When Robots Become Coworkers
For most of the digital revolution, artificial intelligence lived largely behind a screen. Algorithms recommended what we watched, detected fraud, translated languages, analyzed data and increasingly generated words and images. The machine might have been extraordinarily capable, but it remained largely invisible. The next chapter may have legs.
On September 15, 2026, Oregon-based Agility Robotics introduced Digit 5, the newest generation of its humanoid robot designed for industrial work. The announcement is significant not simply because another humanoid robot can walk, lift or carry. Digit represents a larger technological ambition: to give artificial intelligence a physical presence inside workplaces designed for human beings.
Factories and warehouses have used robots for decades. Traditional industrial robots can perform repetitive tasks with extraordinary speed and precision, but many operate in controlled environments, often separated from people by physical barriers. The humanoid proposition is different. Instead of rebuilding the workplace around the machine, companies are attempting to build machines capable of navigating environments already designed around us.
Doors, shelves, aisles, stairs, containers and workstations all reflect human dimensions. A machine with a roughly human form can theoretically operate within that existing infrastructure without requiring an entire facility to be redesigned. That is part of the economic logic behind Digit.
Agility says Digit 5 builds on more than 65,000 hours of operation and commercial deployments and is engineered to work in closer proximity to people without the physical safety barriers associated with much conventional industrial automation. The company says it has received more than $300 million in multi-year orders, with customers across manufacturing, warehousing and logistics. The idea is no longer confined to a laboratory.
In 2024, logistics company GXO entered a multi-year agreement with Agility following a 2023 proof-of-concept program. Digit robots were deployed in regular operations at a GXO facility near Atlanta serving SPANX, moving totes as part of a workflow integrated with other automation. GXO and Agility described it as the first formal commercial deployment of humanoid robots and the first humanoid Robots-as-a-Service agreement. That last phrase may ultimately prove as important as the robot itself.
Robots-as-a-Service changes the economics of automation. Instead of treating an advanced robot solely as a large capital purchase, companies can potentially acquire robotic capability as an ongoing service. It resembles a transformation that has already occurred in software, where businesses moved from purchasing programs outright to subscribing to cloud-based services.
If that model expands, the physical-AI economy could encompass much more than robot manufacturers. There will be software platforms for managing fleets, sensors, semiconductors, maintenance, cybersecurity, insurance, training, systems integration and perhaps entirely new categories of businesses built around robotic labor.
In other words, the humanoid robot may be the most visible part of a much larger commercial ecosystem.
The attraction for employers is understandable. Many industrial environments contain jobs that are repetitive, physically demanding or difficult to staff consistently. A robot does not need to reproduce every capability of a human being to become economically useful. It needs to perform particular tasks reliably enough, safely enough and at a cost that makes deployment worthwhile. That is an important distinction because the popular imagination tends to frame humanoid robotics in dramatic terms: Will robots replace people? The economic reality is likely to be considerably more complicated.
Some tasks may be automated. Some jobs may change rather than disappear. New occupations may emerge around programming, supervising, maintaining and coordinating machines. Companies may redesign workflows so humans and robots perform different parts of the same process. Productivity gains could create new investment and economic activity, while particular workers or occupations could simultaneously experience disruption.
Technology has produced versions of this tension before. Mechanization transformed agriculture. Industrial machinery transformed manufacturing. Computers changed offices. The internet eliminated certain businesses while creating industries that barely existed a generation earlier. Physical AI introduces a new variable because software intelligence can now interact directly with the material world.
The challenge is not simply making a robot walk. A commercially useful humanoid must perceive its surroundings, understand tasks, manipulate objects, respond to unexpected conditions and operate safely around people. It must also be dependable enough that a business can build actual operations around it. A spectacular demonstration is one thing. Thousands of hours of uneventful work are another.
Safety may therefore become one of the industry's most important competitive advantages. Digit 5's development emphasizes what Agility calls “cooperatively safe” operation; an attempt to make robots capable of working near people rather than only within segregated robotic zones. Whether humanoids can consistently achieve the safety, reliability and economics required for broad deployment will help determine how quickly this market develops. There is also a broader question of human value. If machines become increasingly capable of physical work as well as cognitive tasks, what becomes more valuable for people?
Perhaps the answer will not be the ability to compete with machines at repetition. Human advantage may increasingly lie in judgment, creativity, relationships, adaptability, leadership, craftsmanship and forms of dexterity or perception that remain difficult to automate economically.
That would make the robotics revolution not simply a story about machines becoming more humanlike, but about businesses becoming more precise about what human beings are uniquely good at.
For now, Digit 5 should not be mistaken for evidence that humanoid robots are about to appear in every office, hotel or home. Industrial deployment remains an emerging market, and technical demonstrations should not be confused with universal commercial viability.
But something important has changed. Humanoid robotics is moving from the world of prototypes toward the world of contracts, customers, facilities and workflows. Artificial intelligence is beginning to acquire a body.
The next great technology company may therefore sell more than software. The next productivity revolution may involve machines walking through workplaces alongside us. And the defining question may not be whether robots can become more like humans. It may be how humans choose to work with them.
Comments