The Robotics Procurement Map 2026: Where the $87B Goes, and Who Builds It

A global robotics market growing from $55B to $87B by 2030, a supply base of roughly 900 robots from 90 makers concentrated in Japan and Europe, $4.2B of fresh capital chasing humanoids and defense, and a Southeast Asian robot-density gap that Thailand sits squarely inside.

The robotics market is two markets wearing one label. One is mature: industrial articulated arms, a $16.5 billion segment growing at a steady 6.2 percent, dominated by a handful of Japanese and European names that have shipped the same kinematic families for thirty years. The other is explosive: humanoids, autonomous mobile robots, and surgical systems compounding at 24 to 45 percent a year on the back of AI foundation models and a wall of venture capital. A buyer sourcing either one is sourcing from the same narrow supply base, and that base is not in Southeast Asia. This is the procurement map that the numbers draw.

A $55 billion base growing to $87 billion

The aggregate market grows from $42.5 billion in 2020 to $55.1 billion in 2023 and on to a projected $87.0 billion by 2030. Robot unit shipments climb from 384,000 a year to roughly 750,000, and the global installed base doubles from 3.0 million to 6.3 million units. That is the calm surface. Underneath it, the segment-level divergence is where procurement risk and opportunity actually live.

Humanoids and mobile robots are pulling away

For a procurement function this matters more than it looks. A buyer standardizing on industrial arms is buying into a slow, stable, deeply second-sourced market. A buyer reaching for cobots, AMRs, or humanoids is buying into a market where this year’s lead supplier may be acquired, pivoted, or out-shipped before the purchase order ages a year. The spec language, the controller, and the fieldbus protocol all change underneath you. That volatility is a sourcing problem, not just a technology one.

The supply base is concentrated, and it is not in Southeast Asia

Across roughly 900 mapped robots from about 90 manufacturers, the concentration is stark. FANUC alone accounts for 102 distinct models; ABB, Mitsubishi, Codian, Yamaha, and KUKA each field 44 to 53. Industrial articulated arms make up 282 of the mapped models, SCARA 169, collaborative 116, and delta 116. The named manufacturers cluster in Japan, Germany, Switzerland, and increasingly South Korea. None of the volume sits in the Thailand corridor. The question for a Thai-based buyer is not who makes the best robot; it is who can be qualified as a proximate, serviceable second source for one.

This is the gap and the opening at once. China installs 276,288 robots a year, more than the rest of the world’s top economies combined, and runs a density of 392. South Korea and Singapore have saturated. Southeast Asia outside Singapore has barely started: Vietnam at 18, the broader region in the low double digits, and Thailand absent from the dataset. As China-plus-one reshoring pushes assembly into Thailand and Vietnam, that near-zero installed base is exactly where the next decade of robot procurement gets written. The buyers exist before the supplier network does.

Capital is racing ahead of revenue

Across 21 tracked rounds, roughly $4.2 billion of disclosed capital is concentrated in categories that barely register on the revenue charts yet: defense autonomy at $1.5 billion, humanoids at just over $1.0 billion, AI foundation models for robotics at $0.7 billion. The innovation base tells the same story from the patent side, where FANUC (5,847 filings), ABB (4,523), and Yaskawa (3,892) hold the stock while South Korean filers, Samsung and LG, grow filings fastest at 22 to 25 percent. Capital and patents are pointing at the high-growth, low-revenue segments, which means the supplier landscape there will keep reshaping for years.

“The buyers arrive before the supplier network does. In Southeast Asian robotics, that gap between demand and qualified local supply is the entire opportunity.”

What this means for Thailand-based procurement

A robot is not a catalog line. It is a payload rating, a reach envelope, a repeatability tolerance, a controller, and a fieldbus protocol that has to interoperate with everything already on the floor. The database maps 65 distinct controllers and 51 fieldbus protocols, from EtherCAT and PROFINET to the safety variants FSoE and PROFIsafe. A contract that names a robot but stays silent on protocol, safety rating, and end-effector compatibility is a wrong-quote waiting to happen. The same spec-discipline problem that governs PCB and semiconductor sourcing governs robotics, only with more moving interfaces.

• Demand leads supply in the region. Thailand and its neighbors are buyers long before they are a qualified local robot-supply base, so second-source qualification against the Japanese and European incumbents is the highest-leverage sourcing task.
• Segment choice is a risk choice. Industrial arms are deeply second-sourced and stable; cobots, AMRs, and humanoids are fast-moving markets where supplier continuity is not guaranteed across a single purchase cycle.
• Spec the interfaces, not just the robot. Controller, fieldbus protocol, safety rating, and end-effector compatibility belong in the contract; 65 controllers and 51 protocols mean silence equals drift.
• Origin data is the missing layer. The supply base is concentrated in a few countries, and country-of-origin must be made explicit for corridor, tariff, and CBAM-adjacent decisions before a robot enters a build.

The robotics market will roughly double to $87 billion by 2030, but the value for a Thailand-hub procurement platform is not in the headline number. It is in the structural mismatch underneath it: concentrated, mostly Japanese and European supply; surging, capital-heavy new segments; and a regional installed base so thin that Thailand is not yet on the chart. Closing that gap, robot by robot and second source by second source, is the work.