Pepper is the world's most recognized social robot, a semi-humanoid standing 120 cm tall and weighing 28 kg, originally developed by Aldebaran Robotics and manufactured by Foxconn for SoftBank. First unveiled in June 2014, Pepper was designed to read human emotions through facial expression and voice tone analysis, interact through natural speech in 20+ languages, and serve as a customer-facing assistant in retail, hospitality, healthcare, and education. Approximately 27,000 Pepper units were manufactured before production was paused in June 2021 due to weak demand. In February 2025, manufacturer Aldebaran filed for bankruptcy, and in July 2025, Shenzhen-based Maxvision Technology Corp. acquired Aldebaran's core assets including all Pepper and NAO intellectual property. As of April 2026, no new Pepper units are being manufactured. Used units sell for approximately $3,000–$15,000 on the secondary market.
Price and Availability
Pepper is no longer in production. No new units are available for purchase.
| Pricing History | Price |
|---|---|
| Original Japan consumer price (2015) | ¥198,000 (~$1,900) + monthly subscription |
| Original enterprise/business price | ~$25,000–$32,000 |
| Used units (secondary market, 2026) | $3,000–$15,000 depending on condition |
| Academic surplus units | $3,000–$8,000 |
| Milestone | Date | Status |
|---|---|---|
| Pepper unveiled (Tokyo) | June 5, 2014 | Completed |
| Japan consumer sales begin | June 2015 | First 1,000 sold in 60 seconds |
| UK launch | 2016 | Completed |
| ~12,000 units sold in Europe | By May 2018 | Cumulative |
| ~27,000 total units manufactured | By 2021 | Cumulative |
| Production paused | June 2021 | Weak demand |
| SoftBank Robotics Europe → Aldebaran rebrand | 2022 | United Robotics Group acquisition |
| Aldebaran files for bankruptcy | February 2025 | — |
| Aldebaran enters receivership | June 2025 | ~106 employees laid off |
| Maxvision Technology acquires Aldebaran assets | July 2025 | IP for Pepper + NAO acquired |
| Current status | April 2026 | No new production; Maxvision plans TBD |
Full Specifications
Chassis and Build
| Spec | Value |
|---|---|
| Height | 120 cm (3 ft 11 in / 47 in) |
| Weight | 28 kg (62 lbs) |
| Mobility | Omnidirectional wheeled base (3 wheels) — no legs |
| Speed | Up to 3 km/h |
| DOF | 20 (Head: 2, Shoulders: 2×2, Elbows: 2×2, Wrists: 1×2, Hands: 1×2, Hip: 2, Knee: 1, Base: 2) |
| Display | 10.1-inch touchscreen tablet on chest |
| Battery | Lithium-ion, ~12 hours of operation |
| Stair climbing | No — flat surfaces only |
Sensors
| Sensor | Details |
|---|---|
| Head cameras | 2× HD 5 MP cameras (mouth and forehead) |
| 3D depth sensor | Behind eyes |
| Microphones | 4 omnidirectional |
| Touch sensors | 3 on head, 2 on hands |
| Sonar | 2 (on base) |
| Lasers | 6 (on base) |
| Bumper sensors | 3 (on base) |
| Gyroscope | Torso + base |
| IMU | Yes |
Computing and AI
| Feature | Value |
|---|---|
| OS | NAOqi (proprietary) |
| Programming | Python, C++, Java, Choregraphe (visual programming) |
| SDK | Full development toolkit with multiple language support |
| Languages supported | 20+ (including Japanese, English, French, Spanish, Chinese) |
| Emotion recognition | Facial expression analysis + voice tone detection |
| Voice assistant | Built-in speech recognition + compatible with Alexa skills |
| Cloud AI | Original design included cloud-based learning from collective interactions |
| LLM integration | Community-developed integrations with ChatGPT, Claude, Gemini via Python API |
| App store | Pepper/NAO dedicated app store |
Key Capabilities
Pepper was designed as a social interaction platform, not a physical labor robot. Its core capabilities include greeting customers and visitors, providing information through speech and touchscreen, facial recognition and person tracking, emotion detection and adaptive responses, video calling, media playback, smart home integration, guided tours and wayfinding, educational programming via Choregraphe, light object delivery via small tray (limited), and serving as a research platform for human-robot interaction studies.
What Was Pepper Used For?
Retail and Hospitality
Pepper's most visible deployment. SoftBank placed Pepper units in its own mobile stores across Japan, and the robot expanded to retail locations, banks, hotels, airports, and cruise ships worldwide. Pepper greeted customers, answered questions, provided product information, and directed visitors.
Healthcare and Elder Care
The EU-funded CARESSES project researched Pepper as a culturally competent care companion for elderly residents. Long-term studies showed that care home residents willingly interacted with Pepper and staff were open to integrating robots into daily care routines. Pepper handled medication reminders, social engagement, and telepresence calls with family.
Education
One of Pepper's strongest proven use cases. The Choregraphe visual programming tool made Pepper accessible for teaching robotics concepts. Studies showed increased student engagement in STEM subjects and foreign language learning. A massive library of academic resources, lesson plans, and research papers exists around Pepper.
Research
Pepper became one of the most widely studied robots in human-robot interaction (HRI) research. Thousands of academic papers reference Pepper, and it remains a standard platform for HRI studies in universities worldwide.
Buyer's Guide: What to Know in 2026
Pepper Is a Legacy Product
No new units are being manufactured. The original manufacturer (Aldebaran) went bankrupt in 2025. Maxvision Technology acquired the IP but has not announced new production plans as of April 2026. You are buying a discontinued product with uncertain long-term support.
Used Market Is Your Only Option
Used Pepper units are available from academic institutions cycling out equipment, businesses that decommissioned their deployments, and secondary market sellers. Prices range from $3,000 for academic surplus to $15,000 for well-maintained enterprise units.
Software Can Be Updated
While official SoftBank support has ended, the developer community has kept Pepper's capabilities alive. Custom integrations with modern LLMs (ChatGPT, Claude, Gemini) dramatically improve conversational ability beyond Pepper's original AI. This requires Python programming skills.
Physical Limitations Are Real
Pepper rolls on wheels — no legs, no stair climbing. Its hands have 1 DOF each — they gesture but cannot grasp, carry, or manipulate objects meaningfully. The tray on its chest can hold a phone or small object (up to ~3 kg) but Pepper is not a delivery robot.
Consider Modern Alternatives
For customer interaction and greeting: Temi Go ($3,950+) offers a larger screen, autonomous navigation, video calling, and an open SDK. For education: Unitree R1 ($4,900) and Noetix Bumi ($1,400) offer bipedal locomotion and modern AI at comparable or lower prices. For research: newer platforms offer more sensors, better AI, and active manufacturer support.
SoftBank Pepper vs Similar Robots
- Pepper vs Temi Go: Temi is actively manufactured, has a larger 13.3" screen, autonomous delivery capability (3 kg tray), 8-hour battery, open SDK, and fleet management. Pepper has more expressive arms and hands, emotion recognition, and a decade of HRI research behind it. Temi is the practical modern choice; Pepper is the legacy icon.
- Pepper vs NAO: NAO is Pepper's smaller sibling, also from Aldebaran. At ~58 cm tall with 25 DOF, NAO is more physically capable (walking, grasping) but smaller and less suitable for public-facing roles. Both are now owned by Maxvision.
- Pepper vs Unitree R1: The R1 ($4,900) is a bipedal humanoid with 20+ DOF, modern AI, and active manufacturer support. It walks, runs, and flips — things Pepper cannot do. For education and research, the R1 is the forward-looking choice.
- Pepper vs Noetix Bumi: Bumi ($1,400) is a child-sized bipedal humanoid targeting education. At one-tenth the cost of a used enterprise Pepper, it offers true bipedal locomotion and modern programming interfaces. For new classroom deployments, Bumi is dramatically cheaper and more current.
