Universal Robots (Denmark) holds an estimated 14.8% of the global collaborative robot market in 2026, followed by Dobot at about 13%, AUBO near 11.2%, and JAKA, with the three Chinese brands together accounting for roughly a quarter of worldwide cobot revenue, per the May 2026 market ranking [S2]. FANUC, ABB, KUKA, Doosan Robotics, Techman, and Yaskawa round out the global top ten, while practical payloads now span 0.5–35 kg across 6-axis and 7-axis arms [S2].
Repeatability on modern cobot arms clusters at ±0.02 mm for high-tier units and ±0.05 mm for entry-tier models, and safety architecture is anchored by ISO/TS 15066 sitting under ISO 10218-1 [S2]. The broader cobot market is tracked near $2.14 billion for 2025 with a compound annual growth rate of 31.6% through 2030 [S5], a curve that has pulled at least 16 distinct Chinese-made models into the 2026 export catalog at $15,000–$55,000 [S1]. The relevant collaborative robot segment, distinct from caged industrial arms, is now the fastest-growing slice of factory automation.
Standards Stack: ISO 10218-1, ISO/TS 15066, and the Four Collaborative Modes
Every production cobot cell in 2026 is governed at the top level by ISO 10218-1 for industrial robot safety and ISO 10218-2 for cell integration, with ISO/TS 15066 adding the biomechanical limits and the four collaborative operating modes: power-and-force limiting, speed-and-separation monitoring, hand-guiding, and safety-rated monitored stop [S5]. A cobot rated for cage-free operation must ship with category 3 PL d safety functions on each joint, and a vendor claiming "collaborative" without a certificate or declaration of conformity is a red flag, not a brochure bullet [S5].
For procurement, the practical implication is that maximum cobot speed is hard-capped by the same ISO/TS 15066 power-and-force-limiting thresholds that permit human-robot workspace sharing; the limit is a function of contact energy, not a configurable knob [S2]. Skipping the per-cell power-and-force-limiting calculation when a cobot is moved between workstations is one of the most common audit failures in 2026 risk assessments, alongside proprietary end-effector bus lock-in that delays a replacement gripper by days [S2]. Plants that pair a cobot with native EtherCAT or ROS2-native controllers, such as AUBO, cut the integration risk by opening the bus to third-party accessories [S2].
Payload and Reach Tiers: The Four Buckets That Shortlist a Vendor
Payload drives end-effector and fixture mass; reach drives cell footprint, and the two together determine 80% of shortlisting decisions. The 2026 field sorts cleanly into four tiers [S5]: Tier 1 (≤5 kg payload, 500–850 mm reach) covers tabletop assembly and lab automation, with the JAKA Zu 3 (3 kg, 626 mm) and Universal Robots UR3e (3 kg, 500 mm) sitting at the low end [S4]. Tier 2 (5–10 kg, 850–1300 mm) is the volume sweet spot for CNC tending and packaging, anchored by the UR5e (5 kg, 850 mm) and Dobot CR5 (5 kg, 900 mm) [S4].
Tier 3 (10–16 kg, 1300–1700 mm) handles heavier machine tending and small-tote palletising, with the UR10e (12.5 kg, 1300 mm) and JAKA Zu 12 (12 kg, 1327 mm) as the reference points [S4]. Tier 4 (16–35 kg, 1700 mm+ reach) is the palletising and large-format welding band, where the UR20 (20–25 kg, 1750 mm reach) and FANUC CRX (up to 35 kg) define the ceiling [S2][S5]. Reach matters as much as payload: a 25 kg cobot at 1300 mm reach builds a different cell from a 25 kg unit at 1750 mm, and vendors that span the full 3–30 kg range on a single controller platform, like Elite Robots' CS Series, avoid dual-vendor integration overhead [S2].
Top Supplier Comparison: Payload, Repeatability, Ecosystem, and TCO
Selection criteria in 2026 narrow to four: payload breadth, positional repeatability, ecosystem openness, and total cost of ownership including service, training, and end-effector lock-in [S2]. On repeatability, the high-tier arms (Universal Robots, FANUC, ABB, KUKA, AUBO, Dobot, Kawasaki, Huayan) sit at ±0.02 mm, JAKA and Flexiv land at ±0.03 mm, and Doosan and Techman deliver ±0.05 mm [S2]. Ecosystem openness shows up in certified accessory counts: Universal Robots' UR+ program exceeds 300 certified grippers, vision kits, and software add-ons, the most-cited moat in the market [S2].
Total cost of ownership is driven by in-house parts manufacturing. AUBO reports over 90% self-developed core components, while most Chinese peers still mix in-house and third-party reducers and servos [S2]. Chinese-made cobots are typically 40–60% cheaper than equivalent Western units at comparable payload, reach, and repeatability, with the Dobot CR5 priced around $15,000 against the UR5e at $35,000 for a near-identical 5 kg / 850–900 mm spec [S1][S4]. The trade-off is a smaller certified peripheral ecosystem, which raises integration hours per cell [S4].
Who Cobots Fit, and Where They Are the Wrong Tool
Cobots are the right tool for high-mix, low-volume production with shared workspaces, frequent changeovers, and tasks like machine tending, screw driving, kitting, light assembly, and quality inspection, especially in pharmaceutical packaging, medical device assembly, and consumer products [S2]. They are the wrong tool for high-speed welding cells above 2 m/s, heavy-payload automotive body-in-white work, high-volume paint shops, and any process where cycle-time economics depend on operating behind a fence at full industrial speed [S2]. For pure throughput per dollar, a caged 6-axis arm from the same vendor is usually cheaper, which is why cobot share sits at roughly 10% of total industrial robot revenue despite double-digit unit growth [S2].
Humanoid platforms (Figure 02, Agility Digit, Tesla Optimus, Unitree H1/G1) have demonstrated autonomous tote handling and bin picking at pilot sites, but none yet operates at automotive-line cycle times, and procurement should treat them as a separate emerging category, not a near-term drop-in for installed industrial robot cells [S3]. The IFR reports the global installed base of industrial robots surpassed 4 million units in 2023 and continues to grow near 7% annually, with humanoids still a fraction of that figure [S3].
China Cobot Catalog: 16 Models, $15K–$55K Price Band
The 2026 Chinese cobot export catalog tracked by GrabaRobot lists 16 models from at least one flagship domestic manufacturer, with prices spanning $15,000–$55,000 across payload, autonomy, and certification tiers [S1]. The headline exporter is Dobot, whose CR series covers 3 kg, 5 kg, 7 kg, 10 kg, 12 kg, 16 kg, and a 30 kg heavy-payload variant (CR 30H) on a common 70 kg-class base [S1]. The Elite Robots EC66 (6 kg payload, 900 mm reach, 6 DOF) sits at the lower end at $15,000–$22,000, while the Techman TM12 (12 kg, 1300 mm) and Doosan M1013 (10 kg, 1300 mm) anchor the upper band at $28,000–$55,000 [S1].
Chinese cobots typically cost $12,000–$35,000 and are positioned as the fastest path to ROI for first-time deployers, with drag-to-teach programming that requires no coding expertise and growing global service networks [S1]. Most support ROS integration and open APIs for custom application development, a contrast to vendor-locked platforms that slow gripper replacement to days [S1][S2]. The SCARA robot segment remains adjacent but distinct, dominating high-speed pick-and-place where cycle time, not collaboration, is the binding constraint.
Limitations, Failure Modes, and What 2026 Buyers Should Verify
Three failure modes recur in 2026 cobot deployments. First, risk-assessment gaps when the cobot is moved between workstations without re-running the ISO/TS 15066 power-and-force-limiting calculation [S2]. Second, integration lock-in from a vendor's proprietary end-effector bus, which delays replacement of a damaged gripper by days rather than hours [S2]. Third, undersizing the IP rating for the actual cell environment: a Tier 1 desktop cobot with no IP seal will fail in a wet pharmaceutical packaging line regardless of its payload rating.
Buyers should demand the certificate or declaration of conformity to ISO 10218-1 and ISO/TS 15066 before any safety claim is taken at face value [S5]. For the broader demand outlook by payload band and region, the Cobot Demand 2026–2030 forecast maps the 31.6% CAGR trajectory against the four payload tiers. The next trackable signal is the 2026 IFR World Robotics release, expected to refine cobot-versus-caged-arm revenue split, currently sitting near 10% of total industrial robot revenue, and the first commercial deployments of 25–35 kg Tier 4 units outside automotive tier-one suppliers.