Humanoid Robots and eVTOL Aircraft Push Automation From Factory Floors to City Skies

Humanoid robots and electric vertical takeoff and landing (eVTOL) aircraft are moving beyond laboratory demonstrations into increasingly practical environments, from automotive production lines to flight-testing programs. Figure 03 and Boston Dynamics’ electric Atlas are being developed for real industrial work, while companies such as Archer, Joby and BETA are testing electric aircraft for advanced air mobility. The shift is significant because both technologies are now being judged not only by what they can demonstrate, but by how reliably they can operate in real-world conditions.

Humanoid Robots Are Moving Into Real Workplaces

For years, humanoid robots were primarily associated with research laboratories and carefully controlled demonstrations. That picture is changing as manufacturers begin testing machines in factories, warehouses and other environments designed for people.

The appeal of a humanoid form is relatively straightforward: factories and logistics facilities are already built around human-sized doors, shelves, tools, workstations and material flows. A robot capable of navigating the same environment could potentially be introduced without completely redesigning the workplace.

But the harder problem is making that robot reliable enough to work repeatedly, safely and with limited supervision.

Figure 03 from Figure AI is one example of that transition.

Figure introduced the third-generation Figure 03 in October 2025, describing it as a general-purpose humanoid designed for both the workforce and home environments. The robot combines an updated sensor system and hands with Figure’s Helix vision-language-action system, which allows the machine to interpret its surroundings and translate instructions into physical actions.

The company lists Figure 03 at 5 feet 8 inches tall, 61 kilograms in weight, with a 20-kilogram payload and a stated runtime of up to five hours.

Figure 03 Takes Physical AI Into Automotive Logistics

The more important development is not the robot’s specifications but where it is being tested.

At BMW Group Plant Spartanburg in South Carolina, Figure 03 is being used for a logistics-sequencing application. Components arrive unsorted in larger containers, and the robot is designed to pick them up and place them into sequencing trolleys so they can reach assembly workers in the required order.

This follows BMW’s earlier experience with Figure 02.

During a roughly 10-month deployment, Figure 02 supported production of more than 30,000 BMW X3 vehicles, handling the removal and positioning of sheet-metal components for welding. BMW said the robot moved more than 90,000 components and accumulated approximately 1,250 operating hours.

Figure 03 therefore represents an important change in the type of task being explored. Instead of simply repeating a fixed motion at one station, the robot is being tested in a logistics environment requiring movement, object recognition, manipulation and sequencing.

BMW describes this broader approach as Physical AI—the combination of digital AI systems with machines that can perceive and act in the physical world.

Boston Dynamics’ Atlas Takes a Different Route

Boston Dynamics is approaching humanoid robotics from a more industrial direction with its new electric Atlas.

The company unveiled the product version of electric Atlas in January 2026, replacing the earlier hydraulic research platform with a fully electric machine designed for commercial industrial applications. Initial deployments were planned with Hyundai and Google DeepMind.

The current Atlas platform is designed around material handling and industrial workflows. Boston Dynamics says it can autonomously navigate to a charging station and swap its own battery, while supporting applications such as part sequencing, machine tending and order building.

The robot is rated for a sustained payload of up to 30 kilograms and an instantaneous payload of up to 50 kilograms, according to the company’s current specifications.

The significance of Atlas is therefore less about humanoid appearance and more about industrial integration. A useful factory robot must interact with production-management systems, understand task requirements and perform consistently rather than simply complete an impressive demonstration.

Why Autonomy Matters More Than Movement

Modern humanoid robots have become remarkably capable at walking, balancing and manipulating objects. But physical agility alone does not make a useful autonomous worker.

A commercially viable system has to combine several layers:

• Perception: understanding objects, people and surroundings.
• Planning: determining what action should happen next.
• Manipulation: grasping and moving objects reliably.
• Mobility: navigating changing physical environments.
• Safety: operating around people and machinery.
• Integration: communicating with existing factory and logistics systems.
• Learning: adapting skills without requiring engineers to manually program every movement.

This is why current deployments are so important. They expose robots to the variability that laboratory demonstrations often avoid.

A component may be positioned differently. A box may be partially obstructed. A person may enter the robot’s path. Lighting and floor conditions can change.

The real test is whether the system can recover from those situations consistently.

eVTOL Aircraft Are Taking Automation Into the Sky

A parallel transformation is occurring in aviation.

Electric vertical takeoff and landing aircraft—commonly known as eVTOLs—combine electric propulsion with vertical takeoff and landing. The broader category falls within what the FAA calls Advanced Air Mobility, which includes highly automated, electrically powered aircraft capable of vertical takeoff and landing.

Unlike conventional helicopters, many eVTOL designs use multiple electric motors and distributed propulsion systems. This can provide redundancy and different approaches to controlling lift and forward flight.

The technology is being tested for potential air-taxi services, cargo transportation, medical logistics and other short-distance missions.

The U.S. Federal Aviation Administration has moved beyond purely theoretical discussions through its eVTOL Integration Pilot Program (eIPP), which is intended to test advanced air-mobility operations in real environments while coordinating aviation safety, infrastructure and local authorities.

Archer’s Midnight Moves Toward Air-Taxi Operations

Archer Aviation is among the companies pushing toward operational testing.

Its all-electric Midnight aircraft completed a piloted round trip between Salinas Municipal Airport and Monterey Regional Airport in California in July 2026. Archer reported that each flight leg took roughly nine minutes, compared with more than 35 minutes by car for the same journey. The flights were conducted in coordination with the FAA as the company prepared for operations under the eIPP.

Archer is pursuing certification and infrastructure development alongside its aircraft program. The company has also been working with regulators in the United States and United Arab Emirates on pathways toward limited commercial operations.

The important distinction is that successful flight testing does not automatically mean a full-scale air-taxi network is ready.

Aircraft still have to satisfy certification requirements, operating procedures, maintenance standards, pilot or autonomy requirements, charging infrastructure and local airspace rules.

Joby Demonstrates a Different Kind of Autonomy

Joby Aviation has been developing both eVTOL aircraft and autonomous flight technology.

In September 2026, Joby announced that an aircraft equipped with its autonomy technology completed a 3,199-mile transcontinental flight across the United States without control inputs from its onboard safety pilot. The aircraft performed autonomous taxiing, takeoff, navigation and landing while being remotely supervised.

The aircraft used for that demonstration was a converted Cessna Caravan rather than Joby’s passenger eVTOL itself. That distinction matters: the flight demonstrated Joby’s autonomous-flight technology, not the completion of an autonomous commercial eVTOL service.

Joby has also been selected for multiple applications under the U.S. eIPP, creating a pathway toward early operations in participating states.

From Demonstrations to Infrastructure

The biggest change across both robotics and advanced aviation is the emergence of supporting infrastructure.

Humanoid robots need charging systems, fleet-management software, safety protocols, replacement parts, data systems and integration with factory workflows.

eVTOL aircraft need vertiports or suitable landing locations, charging infrastructure, maintenance networks, airspace procedures and regulatory frameworks.

In both cases, the machine itself is only one component of a much larger system.

That is why the current phase of development is less about proving that a robot can walk or an aircraft can fly and more about determining whether these technologies can become dependable parts of existing operations.

The Common Technology Behind the Two Fields

Humanoid robotics and eVTOL aviation appear to be very different industries, but they share a technological direction.

Both increasingly depend on physical AI—AI systems that interpret real-world environments and make decisions that result in physical actions.

For a humanoid robot, that could mean identifying a component, deciding how to grasp it and navigating it to another location.

For an autonomous aircraft, it could mean interpreting sensor data, determining a safe flight path and controlling the aircraft through changing environmental conditions.

This makes perception, simulation, real-time decision-making and safety validation just as important as mechanical engineering.

The challenge is also similar: failures in the physical world are fundamentally different from incorrect answers produced by conventional software. A wrong decision by a robot can damage equipment or injure someone. A wrong decision by an aircraft can have far more serious consequences.

That is why commercial deployment is progressing through staged testing rather than simply switching from human control to full autonomy overnight.

What Comes Next

Humanoid robotics is moving toward more flexible industrial work, while eVTOL development is moving toward limited real-world advanced air-mobility operations.

Figure 03 and Atlas show how humanoids are being evaluated for logistics, material handling and manufacturing. Meanwhile, eVTOL programs from companies such as Archer and Joby are increasingly combining aircraft development with certification, infrastructure and operational testing.

Neither field has reached the fully autonomous future often portrayed in science fiction.

But the technological threshold is changing.

The most meaningful demonstrations are increasingly happening where robots have to work and aircraft have to operate—not simply where engineers can control the conditions.

That shift from laboratory capability to real-world reliability could determine whether humanoid robots become a major industrial technology and whether electric air taxis become a practical new layer of urban transportation.

Most Searched 5 FAQs

1. What is Figure 03?

Figure 03 is the third-generation humanoid robot developed by Figure AI. It is designed for general-purpose tasks and has been tested in real-world environments including automotive logistics.

2. What is Boston Dynamics Atlas used for?

The electric Atlas is being developed primarily for industrial applications such as material handling, part sequencing, machine tending and order building. Boston Dynamics says Atlas can autonomously navigate to charging stations and swap its battery.

3. What does eVTOL mean?

eVTOL stands for electric vertical takeoff and landing. These aircraft use electric propulsion and vertical takeoff and landing capabilities and form an important part of the broader Advanced Air Mobility sector.

4. Are eVTOL air taxis already operating commercially?

Limited testing and pilot programs are underway, but widespread commercial air-taxi networks are not yet established. U.S. programs such as the FAA’s eIPP are designed to test and integrate these aircraft into real operating environments while certification and infrastructure requirements continue to develop.

5. Are humanoid robots fully autonomous?

Some humanoid robots can perform specific tasks autonomously, but autonomy remains task- and environment-dependent. Current industrial deployments are generally designed around defined workflows, safety systems and varying levels of human oversight rather than unrestricted human-equivalent autonomy.