Boston Dynamics has opened the Robotics Metaplant Application Center (RMAC) inside Hyundai's Georgia factory to train its electric Atlas humanoid robots for manufacturing work, but no robots are actually deployed on production lines yet. The facility marks the first operating site tied to Hyundai and Boston Dynamics' strategy announced at CES 2026, giving the robot developer a structured path from prototype testing to fleet operations. However, the much-publicized commitment to deploy more than 25,000 Atlas robots across Hyundai and Kia factories remains a 2028 target, not a current reality.

What Is Boston Dynamics Actually Doing at the Georgia Factory?

The RMAC facility is designed as a "data factory," according to Boston Dynamics CEO Robert Playter, where each training run improves both the robot and the company's proprietary manufacturing dataset. Engineers capture how Atlas lifts, turns, grasps, recovers from errors and works around people under real factory conditions. The initial task is parts sequencing for the IONIQ 5 and IONIQ 9 vehicles, where the robot must retrieve components from bins and arrange them in the order required by the assembly line.

The facility combines physical training runs with simulation and operational data from Hyundai's "Software-Defined Factory," a connected production system that collects data across the plant. This allows Boston Dynamics engineers to use those records to retrain robot control policies, test updates in simulation, and return validated versions to the factory floor. The first phase is operating now, with plans to move into a building roughly 10 times as large next year.

When Will Atlas Actually Start Working on Assembly Lines?

The deployment timeline is measured in years, not months. Atlas is scheduled to begin production deployment at Hyundai Motor Group Metaplant America in 2028, with expansion to Kia's Georgia plant in 2029. By 2030, the task set will expand from parts sequencing to component assembly and heavier repetitive work. These are planned milestones, not confirmed achievements.

The delay reflects a fundamental economics problem. Atlas currently costs an estimated $130,000 to $140,000 per unit to build, but Hyundai expects that cost to fall to around $30,000 once cumulative production passes 50,000 units. That kind of manufacturing scale requires significant infrastructure investment. Hyundai Mobis, the group's parts division, is building a U.S. actuator plant designed to reach 350,000 units of annual capacity starting in 2028.

How Does Boston Dynamics' Vertical Stack Actually Work?

Hyundai has structured the Atlas program across three separate operations, creating what amounts to an integrated supply chain:

• Boston Dynamics: Develops the Atlas hardware platform, control software, and training system
• Hyundai Mobis: Manufactures the actuators that drive the robot's joints
• Hyundai Motor Group: Supplies the factories, production workflows, and operational data

This arrangement gives Hyundai direct control over the robot developer, actuator production, and a planned fleet large enough to influence hardware design. The structure can shorten the cycle between observing a failure, updating the software, and testing the change on the same workflow. All 2026 Atlas production is committed to RMAC and Google DeepMind, with additional customers planned for early 2027.

What's the Reliability Problem Nobody's Talking About?

While shipment numbers for humanoid robots are climbing, actual factory deployments remain limited. The real obstacle isn't the timeline; it's whether the robots can perform reliably enough to replace existing automation. Industry veterans are skeptical about the current state of the technology.

"80% success is failure in production," stated Mikell Taylor, General Motors executive.

Mikell Taylor, General Motors

A robot that works four times out of five in a lab demo doesn't survive contact with a production line that has to run the same way every shift without an engineer standing by to intervene. Jim Brown of Teradyne Robotics, which makes industrial automation systems, noted that robot makers aren't always "thinking about what problems we are actually trying to solve." He points to a cost most demos skip past: standing still isn't free. Today's humanoid robots draw meaningful power just holding a pose, a real cost on a factory floor that already runs wheeled robots and fixed arms that solve the same tasks more efficiently.

How Does Atlas Compare to Other Warehouse Robots?

Atlas isn't the only humanoid robot targeting warehouse and manufacturing work. Agility Robotics unveiled Digit 5, its fifth-generation bipedal warehouse robot, which emphasizes charging speed over payload capacity. Digit 5 runs for about 90 minutes on a charge and returns to full power in roughly 9 minutes, a 10-to-1 run-to-charge ratio that Agility frames as enabling more than 20 productive hours per day.

On paper, Atlas currently claims an advantage in raw capability. Atlas is rated to carry up to 66 pounds and reach 7.5 feet high, while Digit 5 can repeatedly lift loads of up to 50 pounds and reach items up to 7.2 feet high. Atlas also runs for about four hours on a hot-swappable battery pack that can be physically pulled and replaced rather than recharged in place. Digit 5's advantage is charging speed and, by extension, uptime: a fixed nine-minute recharge cycle is a different operational model than swapping battery packs.

However, neither robot has demonstrated sustained, reliable operation at the scale Hyundai is promising. BMW's plant in Leipzig, Germany began testing AEON, a humanoid built by Hexagon Robotics, in December 2025, and Figure AI's Figure 02 robot assisted with more than 30,000 vehicle builds at BMW's Spartanburg, South Carolina plant in 2025. These are pilots, not fleets, and neither program comes close to the volume Hyundai has committed to for 2028.

Steps to Understanding the Real Atlas Timeline

• Separate announcements from deployments: Hyundai announced its 25,000-robot commitment at a JPMorgan investor conference in May 2026, not from a production floor. No Atlas units are deployed in any Hyundai factory today.
• Track infrastructure investment, not just robot orders: Hyundai Mobis is building a U.S. actuator plant designed to reach 350,000 units of annual capacity starting in 2028. That's an infrastructure bet that signals confidence in the timeline, but it's not a deployment.
• Watch for reliability metrics, not just shipment numbers: The real test will be task performance data: pick success rates, cycle time, recovery rates, uptime, and human interventions per shift. These numbers will determine whether 2028 is an early estimate or an optimistic one.
• Monitor cost reduction milestones: Atlas costs $130,000 to $140,000 per unit today but needs to reach $30,000 to make the economics work at scale. Watch for announcements about cumulative production hitting 50,000 units, the threshold Hyundai has set for cost reduction.

The humanoid robot boom is real in terms of shipments. Chinese vendors now account for more than 97 percent of global humanoid robot shipments, with Agibot shipping somewhere between 8,400 and 9,700 units in the first half of 2026 alone. But the factory takeover story is still navigating a reliability gap that no company has fully solved.

Boston Dynamics' RMAC facility represents a genuine step forward: on-site training, a large internal order, and direct access to Hyundai's factory systems give the company a structured path from prototype testing to fleet operations. But the gap between announcing a 25,000-robot deployment and actually running those robots on a production line remains measured in years, not months. The real question isn't whether Atlas can be built and trained; it's whether it can run a Western factory floor reliably enough to replace the wheeled robots and fixed arms already doing the job.