Figure 02 vs Rainbow Robotics RB-Y1: Price, Specs & Use Case Comparison (2026)

The core tension is clear: a cutting‑edge US lab humanoid (Figure 02) that isn’t for sale versus a South Korean wheeled humanoid (RB‑Y1) you can order right now — both hovering around $120,000–$130,000. This article maps every technical and practical dimension that matters for robotics engineers, researchers, and operations managers who are cross‑shopping the two platforms. You’ll find a side‑by‑side spec table, a verdict by use case, deep dives into software, durability, and real‑world deployments, and a final recommendation.

Figure 02 vs Rainbow Robotics RB-Y1: Specs Compared

Winner in raw payload & dexterity: Figure 02 — 20 kg payload and 28 DOF crush the RB‑Y1’s 3 kg limit, but the RB‑Y1’s wheeled stability and actual purchase path make it a pragmatic winner for most 2026 buyers.

Figure 02: Bipedal Dexterity with Helix AI

Launched in 2024 by Figure AI (Sunnyvale, California), the Figure 02 brings a Step‑Change in humanoid capability through its proprietary Helix multimodal AI model. Standing 1.68 m and weighing 70 kg, it carries a 20 kg arm payload and runs for 5 hours on a single charge. With 28 DOF, it executes complex manipulation tasks — the kind you’d expect from a human worker — and has been running industrial pilots at BMW’s Spartanburg plant since 2024, handling sheet metal and assembly steps. Pricing is an estimated $130,000 for early access; general availability has not been announced.

Image: Figure 02 side view Figure 02 in its pilot configuration. Source: humanoid.guide

Rainbow Robotics RB-Y1: Wheeled Stability, Immediate Availability

The Rainbow Robotics RB-Y1 is a wheeled humanoid from South Korea’s Rainbow Robotics, now available for order at $120,000. Originally introduced in 2011 as a research platform, the current RB‑Y1 iteration is purpose‑built for AI‑driven mobile manipulation in factories, warehouses, and academic labs. It stands 161 cm tall (with the wheeled base) and weighs a hefty 131 kg, which delivers exceptional stability for arm movements. The two‑wheeled differential‑drive base enables speeds of 5.4 km/h. The robot features 24 DOF, a 3 kg arm payload, and a 3‑hour battery life, making it suited for repetitive logistics tasks like kitting, picking, and light assembly rather than heavy‑payload industrial work.

Image: Rainbow Robotics RB-Y1 Rainbow Robotics RB-Y1 wheeled humanoid. Source: humanoid.guide

Price & Total Cost of Ownership

Base price: $120,000 (RB‑Y1) vs. $130,000 (Figure 02, estimated). The sticker‑price delta is only $10,000, but the real difference lies in availability: the RB‑Y1 can be ordered today (likely a pre‑order with defined lead time), while the Figure 02 is locked behind pilot programs with no public purchase channel. For a research lab or integrator that needs a robot in the next quarter, the RB‑Y1 is the only viable option.

Total cost of ownership (TCO) remains opaque for both. Neither manufacturer publishes annual software subscription fees, spare‑part costs, or service‑level agreements. Given the RB‑Y1’s simpler wheeled base and lower battery capacity, replacement batteries will likely be cheaper than the Figure 02’s bipedal‑specific packs, but this is speculative. Both robots carry roughly the same upfront investment, but the RB‑Y1 wins on immediate project‑start capability.

> Need pricing for your project? Request a quote for the Rainbow Robotics RB-Y1 from our partner network.

Winner: Rainbow Robotics RB-Y1 — you can actually buy it.

Performance & Specs

Payload

The Figure 02’s 20 kg arm payload is a league above the RB‑Y1’s 3 kg. For bin picking, machine tending, or tool handling, the Figure 02 can manipulate a substantial range of industrial parts. The RB‑Y1’s payload suits light object transport, tablet kitting, or sensor manipulation — typical for logistics research rather than heavy manufacturing.

Winner: Figure 02.

Speed

At 10.8 km/h, the Figure 02 walks swiftly; the RB‑Y1 rolls at 5.4 km/h — drastically slower. However, the wheeled base trades speed for stability. Over a warehouse floor, a walking humanoid may stumble on uneven surfaces, while the RB‑Y1 glides smoothly. For long‑distance autonomous navigation, the RB‑Y1’s wheeled efficiency might yield better uptime despite lower peak speed.

Winner: Figure 02 (raw speed), but RB‑Y1 for floor‑proven stability.

Battery life

5 hours (Figure 02) vs. 3 hours (RB‑Y1). The Figure 02’s longer runtime supports continuous shifts without immediate battery swapping, while the RB‑Y1’s 3‑hour window demands more frequent charging — manageable in a controlled lab but a hindrance in 24/7 logistics.

Winner: Figure 02.

Degrees of Freedom

28 DOF (Figure 02) vs. 24 DOF (RB‑Y1). The extra joints in the Figure 02’s hands, arms, and torso give it superior dexterity for complex assembly. The RB‑Y1’s 24 DOF are carefully allocated to its dual arms and waist, providing competent bimanual tasks but without the same finger‑level precision.

Winner: Figure 02.

Software & SDK Ecosystem

Both companies have yet to disclose a full SDK or ROS2 stack publicly. Figure AI’s Helix model is a multimodal “vision‑language‑action” AI trained on human demonstrations, promising zero‑shot task generalization, but developers cannot yet access it directly — all current deployments are tightly managed pilot programs. Rainbow Robotics has historically offered Python/C++ APIs and simulation support for its collaborative robots; it is likely that the RB‑Y1 will ship with a similar SDK. Until an open developer portal emerges, neither robot offers a plug‑and‑play ecosystem for third‑party apps. For researchers who need to write custom controllers, the RB‑Y1 may gain an edge if Rainbow publishes its SDK first; for now, both are opaque.

Winner: Too early to call. RB‑Y1 leads slightly based on company track record.

Autonomy & Sensors

Figure 02 leverages Helix AI for vision‑based task learning and autonomous navigation. In BMW pilots, it reportedly navigates the factory floor and picks parts without external markers — impressive, but the sensor suite (cameras, lidar, etc.) remains undisclosed. The RB‑Y1 relies on AI‑driven control for mobile manipulation; its base likely uses 2D lidar and cameras for localization, though exact specs are again unpublished. Neither robot has a published IP rating, which limits outdoor or wash‑down use.

Winner: Figure 02 for demonstrated autonomous dexterity; RB‑Y1 for wheeled navigation stability.

Build Quality & Durability

The RB‑Y1’s 131 kg steel‑and‑aluminum frame and wheeled base promise robust, low‑maintenance operation indoors. The Figure 02, at 70 kg, uses lightweight composites and actuators to achieve agility but may be less tolerant of rough handling. Both lack published IP ratings, so neither is suited for wet, dusty, or outdoor environments. The RB‑Y1’s dual‑wheel design eliminates leg‑joint stress but introduces tire wear — a manageable consumable. The Figure 02’s bipedal system will require regular actuator and joint maintenance, though no data is available on MTBF (mean time between failures).

Winner: RB‑Y1 for mechanical simplicity and likely longer uptime in controlled environments.

Which Should You Buy? Figure 02 vs Rainbow Robotics RB-Y1 by Use Case

• Choose the Figure 02 if…
• You are an automotive or heavy‑manufacturing integrator willing to join a closed pilot program (BMW‑style).
• You need 20‑kg payload dexterity for assembly, machine tending, or tool use.
• You can wait 12–24 months for general availability and are comfortable with evolving AI (Helix).

• Choose the Rainbow Robotics RB‑Y1 if…
• You need a stable mobile manipulation platform for logistics, warehouse picking, or research today.
• Your payload needs are ≤3 kg (light kitting, sensor scanning, lab automation).
• You prioritise a straightforward wheeled base over legged agility and can accept a slower walking speed.

• Choose neither if…
• You need a proven, off‑the‑shelf quadruped for payloads over 20 kg and outdoor autonomy (consider Boston Dynamics Spot or Unitree B2).
• Your budget is under $50,000 — both robots start above $100,000.

Real-World Deployments & Track Record

• Figure 02: Active industrial pilots at BMW’s Spartanburg plant since 2024 (source: Figure AI press release, 2024). The robot has been shown handling sheet metal and performing autonomous kitting. No public number of units shipped; estimates suggest fewer than 10 pilot units. Operating hours are undisclosed.
• Rainbow Robotics RB‑Y1: Rainbow Robotics has a strong track record in collaborative robots, but the RB‑Y1 is a new wheeled humanoid. As of mid‑2026, no large‑scale commercial deployments have been publicly announced. The robot is available for order, suggesting at least a few early adopter installations in South Korean research labs and logistics firms (source: Rainbow Robotics product page, 2026).

Both robots are at the early‑stage deployment end of the spectrum. The Figure 02’s BMW partnership lends more credibility for industrial use, but the RB‑Y1’s purchase‑ready status makes it the more attainable platform for immediate testing.

Winner (deployment readiness): Figure 02 by reputation; RB‑Y1 by actual orderability.

SDK, Software & Developer Ecosystem

At this stage, no developer‑friendly platform exists for either robot. Figure AI has hinted at a “developer program” but hasn’t launched it. Rainbow’s existing line of humanoids and cobots has some ROS2 support, and the company might release an RB‑Y1 ROS package in 2026. For researchers who must code from scratch, the RB‑Y1’s lower‑level access (typical for Rainbow’s DNA) could be faster to exploit than the Figure 02’s closed Helix stack. If AI‑driven task learning is your focus, the Figure 02’s on‑board Helix gives a unique advantage — once access opens.

Winner: Too early to call, but RB‑Y1’s hardware is more likely to be hackable first.

Support, Warranty & Availability

No formal warranty terms or enterprise support SLAs are published for either robot. Rainbow Robotics, as an established manufacturer, likely offers standard 1–2‑year warranty and after‑sales service; however, the RB‑Y1’s classification as a “pre‑order” model suggests limited immediate support bandwidth. Figure AI provides direct engineering support to its pilot partners (BMW), but no commercial service model exists for the Figure 02. Lead time for the RB‑Y1 is unspecified but presumably weeks to months once ordered; the Figure 02 remains invite‑only.

Winner: RB‑Y1 for having a purchase button.