The Apptronik Apollo is a full-size general-purpose humanoid robot standing 173 cm (5 ft 8 in) tall and weighing 73 kg, designed from the ground up for industrial logistics and manufacturing work. It can carry payloads up to 25 kg (55 lbs), runs for 4 hours on a hot-swappable battery, and features 71 degrees of freedom for whole-body dexterity. Apollo is not currently available for general purchase, but it is actively deployed in pilot programs at Mercedes-Benz factories, GXO Logistics warehouses, and Jabil manufacturing facilities. Backed by over $935 million in Series A funding from Google, Mercedes-Benz, and ARK Invest at a $5.5 billion valuation, Apptronik is targeting commercial-scale production in 2026 with a long-term price target below $50,000.
Price and Availability
Apollo is not available for general purchase as of April 2026. It is in active pilot deployment with select enterprise customers. Apptronik has not published a retail price.
| Milestone | Status | Timeline |
|---|---|---|
| Apollo unveiled | Completed | August 2023 |
| Mercedes-Benz pilot (Hungary, Berlin-Marienfelde) | Active | March 2024–present |
| GXO Logistics warehouse pilot (US) | Active | 2024–present |
| Jabil manufacturing partnership | Active | February 2025–present |
| Series A total funding | $935 million ($5.5B valuation) | February 2026 |
| Google DeepMind AI partnership (Gemini Robotics) | Active | 2025–present |
| Commercial-scale production target | Planned | 2026 |
| Long-term target price per unit | Below $50,000 | At scale |
| Current estimated manufacturing cost | $150,000 per unit (industry estimate, down from ~$250,000 in 2023) | Early 2026 |
| Production capacity | <1,000 units to date | Scaling via Jabil partnership |
Important context on availability: Apollo is an enterprise product being deployed through direct partnerships, not a retail product you can order online. If you are a manufacturer, logistics provider, or enterprise buyer interested in piloting Apollo, contact Apptronik directly through their website. There are no consumer sales planned for the near term.
Full Specifications
Chassis and Build
| Spec | Value |
|---|---|
| Height | 173 cm (5 ft 8 in) |
| Weight | ~73 kg (160 lbs) |
| Total degrees of freedom | 71 |
| Frame | Industrial-grade materials, modular design |
| Form factor | Human-scale — fits through standard doorways, navigates human workspaces |
| Modularity | Upper body can be mounted on legs, wheeled base, or stationary pedestal |
| LEDs | Head, mouth, and chest — communicate operational status |
| Design philosophy | "Friendly interaction, mass manufacturability, high payloads, and safety" |
Performance
| Spec | Value |
|---|---|
| Payload capacity | 25 kg (55 lbs) |
| Walking speed | Moderate (exact speed not officially published — estimated ~1.5 m/s based on demo footage) |
| Mobility modes | Bipedal legs, optional wheeled base, or stationary mount |
| Manipulation | Dexterous hands with force control for grasping and object handling |
| Demonstrated tasks | Tote handling, parts transport, assembly kit delivery, component inspection, sorting, packing |
Battery and Power
| Spec | Value |
|---|---|
| Battery type | Hot-swappable lithium packs |
| Runtime per battery | ~4 hours |
| Battery swap time | Under 5 minutes |
| Daily operational capacity | Up to ~22 hours (with battery rotation) |
| Tethered operation | Supported — can plug in for continuous operation without battery limits |
| Charging | Standard plug-in charging between shifts |
Sensors and Perception
| Feature | Value |
|---|---|
| Vision system | Multiple cameras (details not fully public) |
| Force/torque sensing | In all joints — enables safe force-controlled movements around humans |
| Foot sensors | Force/torque with human-inspired geometry for dynamic balance |
| Safety zones | Configurable outer detection zone — robot pauses when human crosses boundary |
| Emergency stop | Yes |
Computing and AI
| Feature | Value |
|---|---|
| Core compute modules | NVIDIA Jetson AGX Orin + Jetson Orin NX |
| AI platform | NVIDIA GR00T (Project GR00T foundational model) |
| AI models | Google DeepMind Gemini Robotics 1.5 + Gemini Robotics Extended Reasoning |
| Learning approach | Imitation learning from demonstrations + natural language task instructions |
| Fleet learning | Data collected across deployed units feeds back into central AI training |
| Demonstrated AI capabilities | Handling unfamiliar objects, following verbal commands, adapting to changing conditions, multi-step task planning |
| Software | Point-and-click control suite for warehouse and manufacturing integration |
Safety Architecture
| Feature | Value |
|---|---|
| Control approach | Force control (not position control) — robot senses and adapts to contact |
| Collision response | Joint compliance reduces impact forces — similar to cobot vs industrial robot safety |
| Safety zones | External sensors and "light curtain" boundaries — robot pauses if human enters zone |
| Future goal | Full "collaborative safety" — slowing, stopping, and maneuvering alongside humans without barriers |
| Design speed | Limited top speed — humans can easily move away |
| Weight | 73 kg — manageable by a person in emergency |
What Makes Apollo Different?
Built for Work, Not for Research
Most humanoid robots on the market today are research platforms (Unitree G1, H1) or development programs (Tesla Optimus, Figure). Apollo was designed from the start as a commercial workhorse for logistics and manufacturing. Every design decision — hot-swappable batteries, modular mounting, force-controlled safety, 25 kg payload — reflects real-world operational requirements rather than lab demonstrations.
NASA Heritage
Apptronik's founding team worked on NASA's Valkyrie (R5) humanoid robot at the University of Texas Austin. This background in building rugged, reliable systems for extreme environments directly shaped Apollo's emphasis on durability, safety, and practical functionality over flashy athleticism.
Force Control Architecture
Unlike most humanoid robots that use position control (move precisely to a specific point), Apollo uses force control — it senses and responds to contact forces in real time. This makes collisions less dangerous, enables safe human collaboration, and allows the robot to adapt to unpredictable environments. It's the same principle behind collaborative industrial robots (cobots), applied to a humanoid form factor.
Hot-Swappable Batteries
Each battery pack provides 4 hours of operation and can be swapped in under 5 minutes without shutting down the robot. With battery rotation, Apollo can operate up to 22 hours per day. This is a genuine operational advantage over robots that require plug-in charging and sit idle during charge cycles.
Modular Base
Apollo's upper body can be mounted on bipedal legs, a wheeled platform, or a fixed pedestal. This means the same robot can walk through a warehouse, roll through a factory floor, or sit at a workstation — adapting to whatever the deployment requires.
71 Degrees of Freedom
Apollo has significantly more DOF than most competitors: Tesla Optimus has ~28 body DOF (plus hands), Figure 02 has ~41, and Unitree H2 has 31. Apollo's 71 DOF enables more complex whole-body movements — reaching, bending, rotating, and manipulating with coordinated arm-torso-leg motion.
Who Is Apollo For?
Manufacturing and Automotive
Mercedes-Benz is testing Apollo at its Berlin-Marienfelde facility and a plant in Hungary for delivering assembly kits, transporting components between workstations, and inspecting parts on the production line. Jabil, the global electronics manufacturer, is both a manufacturing partner (helping scale Apollo production) and a deployment customer (integrating Apollo into its own factories).
Warehousing and Logistics
GXO Logistics, the world's largest pure-play contract logistics provider, is running a multi-phase pilot with Apollo in a US distribution center. The initial focus is case picking — moving standard containers (cases and totes) through warehouse operations. This is the same use case Agility Robotics targets with Digit.
Future Applications
Apptronik envisions longer-term expansion into construction, oil and gas, electronics production, retail, home delivery, and elder care. However, near-term deployment is firmly focused on manufacturing and logistics where the business case is clearest.
Buyer's Guide: What Enterprise Buyers Need to Know
You Cannot Buy Apollo Off the Shelf
Apollo is deployed through direct enterprise partnerships, not retail channels. If you are interested, contact Apptronik directly. Expect a pilot program structure — evaluation phase, task identification, staged deployment — rather than an off-the-shelf purchase.
Current Price Point Is Enterprise-Grade
Current manufacturing costs are estimated at roughly $150,000 per unit (down from ~$250,000 in 2023). Apptronik's long-term target is below $50,000 at scale, but that requires production volumes the company has not yet achieved. Early customers should budget at enterprise robot pricing levels.
Production Is Scaling, Not Scaled
Fewer than 1,000 Apollo units have been produced to date. The Jabil manufacturing partnership is designed to ramp this up significantly in 2026 and beyond, but Apollo is not yet at mass production volume. Lead times and availability will depend on your deployment scale and timing.
AI Is Advancing Rapidly
The Google DeepMind partnership (Gemini Robotics) and NVIDIA GR00T integration mean Apollo's AI capabilities are improving continuously. Recent demonstrations show the robot handling unfamiliar objects, following verbal instructions, and adapting to changing conditions — but these remain controlled demo environments, not fully autonomous production deployments.
Safety Is a Core Design Principle
Apollo's force control architecture, configurable safety zones, and emergency stop systems are designed for human-robot collaboration. However, current deployments use designated zones with light curtain boundaries — humans and Apollo work in separated areas. Full barrier-free collaborative operation is a future goal, not a current capability.
Apptronik Apollo vs Similar Robots
- Apollo vs Tesla Optimus: Apollo is further along in commercial deployment — active pilots with Mercedes-Benz, GXO, and Jabil versus Optimus's internal-only Tesla factory deployment. Apollo has 71 DOF (vs Optimus's ~28 body + hands), higher payload (25 kg vs 20 kg), and hot-swappable batteries. Optimus has Tesla's massive manufacturing scale and a much lower target price ($20,000–$30,000 vs Apollo's sub-$50,000 target). Neither is available for general purchase.
- Apollo vs Figure 02/03: Figure has raised even more capital ($2.6B+ at $39B valuation) and demonstrated strong AI reasoning via its Helix platform. Both target similar industrial use cases. Apollo has higher payload (25 kg vs 20 kg) and more DOF (71 vs ~41). Figure has BMW as a factory partner. Neither is available for retail purchase.
- Apollo vs Agility Digit: Digit is the most deployed humanoid in commercial logistics, with active Amazon and GXO warehouse programs. Digit has a purpose-built design for case handling with an 8-hour battery. Apollo is more general-purpose with higher payload and more versatile mounting options. Digit costs roughly $250,000; Apollo's current cost is estimated at $150,000.
- Apollo vs Unitree H2: The H2 is available now at $29,900 — dramatically cheaper than Apollo's estimated $150,000. The H2 has 31 DOF (vs Apollo's 71), lower payload (7 kg vs 25 kg), and is targeting a broader market including education and commercial service. Apollo is a purpose-built industrial platform; the H2 is a general-purpose humanoid accessible to a wider buyer base.
- Apollo vs Unitree G1: The G1 ($13,500–$73,900) is a compact research humanoid you can buy and deploy today. Apollo is an enterprise industrial platform you cannot yet purchase. Different categories entirely — the G1 serves researchers, the Apollo serves Fortune 500 factory floors.
- Apollo vs Boston Dynamics Atlas: Atlas is the most mechanically advanced humanoid ever built but is in limited enterprise pilot programs at estimated $320,000–$420,000. Apollo offers a more accessible price trajectory and is further along in commercial partnership deployment.
The Bottom Line
The Apptronik Apollo is one of the most commercially mature humanoid robots in the world as of 2026. Real deployments at Mercedes-Benz, GXO, and Jabil — backed by $935 million in funding and partnerships with Google DeepMind and NVIDIA — give it credibility that most competitors lack. The 71 DOF, 25 kg payload, hot-swappable battery system, and force-controlled safety architecture are all designed around practical industrial requirements rather than research flexibility or consumer appeal.
But Apollo is not a product you can buy today. It is an enterprise deployment accessed through direct partnership with Apptronik. Current costs are estimated at $150,000 per unit, production is still ramping, and the AI capabilities — while impressive in demonstrations — are not yet fully autonomous in production environments.
For enterprise buyers evaluating humanoid robots for factory or warehouse pilots: Apollo should be on your shortlist. Contact Apptronik directly to explore a pilot program.
For researchers, educators, or developers who need a humanoid robot they can actually purchase and deploy now: look at Unitree's lineup (G1, H2, R1) which offers genuine capability at accessible price points and is shipping today.
