The Big Four at 38%: Thirty Years of Rule, Repriced

Walk into a Chinese automotive body shop around 2012 and the robots on the line came in four colours: Fanuc yellow, ABB white, Yaskawa cobalt blue, KUKA orange. Buying an industrial robot was effectively a choice of four, because for three decades it effectively was one — at its peak the quartet held more than 60 percent of the global industrial robot market, and more still in automotive and semiconductor applications.

By 2025 that is no longer the market. Mordor Intelligence puts the four companies’ combined global share at about 38 percent — below 40 percent for the first time. In China, the world’s largest industrial robot market with over 336,000 units sold in 2025, the shift is sharper: the four families’ combined share has fallen from above 60 percent around 2020 to roughly 45 percent, while domestic brands climbed from under 29 percent to over 55. Estun shipped about 33,400 robots in China last year — roughly 10 percent — and became the number-one brand by unit shipments for the first time.

“The Big Four are declining” is too crude to be useful. What actually happened is four different stories, and reading them separately is what makes the numbers actionable.

Fanuc: the fortress holds globally and leaks in China

Fanuc’s full-year results for the year ending March 2026 show revenue of ¥857.8 billion (about $5.7 billion), up 7.6 percent and a record, on a 21.4 percent operating margin. Its cumulative installed base has passed one million units — no competitor is close, and that base is itself a moat: a line that has run for ten years is not replaced over a 10 percent price difference, because the switching cost sits in the control system, the programs and the maintenance culture, not in the arm.

Yet Fanuc’s share in China has fallen from over 20 percent at its peak to around 9 percent. Record global results and a halved China share in the same reporting period is the single most informative fact in this story: the fortress is intact, and the siege is regional.

Yaskawa: volume survives, margin does not

Yaskawa’s robot segment posted ¥247 billion (about $1.65 billion), up 4 percent, with operating profit down — the classic signature of a vendor holding share in a market that has entered a price war. Its roots are the deepest in the industry: servo motors and motion control remain the reference technology. But in mid-range general-purpose machines, the segment where Chinese volume is concentrated, there is little room left to defend both price and position. Yaskawa’s China share is now around 4 percent.

ABB: the rational exit

In October 2025 ABB agreed to sell its entire robotics division to SoftBank at an enterprise value of $5.375 billion, abandoning a planned spin-off. The arithmetic behind it is unemotional: robotics was $2.3 billion of revenue in 2024 — 7 percent of the group — on a 12.1 percent operational EBITA margin, against 16–18 percent in ABB’s core electrical and automation businesses. Leaving was rational for ABB. For the industry it means something larger: the “Big Four” no longer exists as a coherent strategic bloc. One of the four has already left the table.

KUKA: the only one growing in China, for a different reason

KUKA shipped more than 32,000 robots in China in 2025, up over 30 percent, for a domestic share near 9.6 percent; China now accounts for about 30 percent of group revenue. The structural advantage is specific: in robots above 300 kg payload, KUKA holds 47.4 percent of Chinese sales — a position built on decades of automotive and heavy-payload engineering that cannot be replicated quickly. And as part of Midea, KUKA has what the other three never had: local manufacturing (about half of its Chinese output is produced in China), a localized core-component supply chain above 85 percent, and dense channel coverage. KUKA is the first of the four to complete the “in China, for China” turn — which is precisely why it grows while the others bleed.

What domestic substitution actually replaced

The share shift is usually told as a price story. The mechanism is a component story, and it is further along than most buyers assume.

• RV reducers: domestic share of the Chinese market rose from 41.6 percent in 2022 to 52.9 percent in 2023 and 60.8 percent in 2024 (GGII).
• Harmonic reducers: from 16.4 percent in 2017 to roughly 75 percent in 2024.
• Servo systems: Inovance holds about 28.3 percent of the Chinese general servo market — first place — against roughly 7 percent for Yaskawa, 7 percent for Panasonic and 10 percent for Siemens.

Once the core components crossed from “usable” to “good”, the cost structure and iteration cadence of Chinese robot makers stopped being set abroad. The leading domestic OEMs are now reproducing the vertical integration that built the Big Four’s advantage in the first place: Estun spans motion controllers, servos, robot controllers, arms and workstations; Inovance is weaving servos, controls, SCARA, six-axis robots and machine vision into one stack. In light-payload small six-axis, welding, palletizing and collaborative segments, domestic share has reached roughly 58, 62, 70-plus and 85 percent respectively, and top domestic light six-axis arms now claim static repeatability of ±0.02–0.03 mm — within range of mid-tier Japanese and European products on paper.

Paper is the operative word. In automotive body-in-white lines, semiconductor front-end equipment and high-precision assembly — where takt consistency, ten-year reliability and mature process packages decide the award — the incumbents’ position remains hard to displace. A robot that hits its numbers in a lab and one that holds them on a line running twenty hours a day for a decade are different products. That trust is accumulated in decades, and the four families have thirty to forty years of it.

The dimension that actually changed

The deeper shift is not share. It is the scoring system.

For thirty years the criteria were precision, speed and reliability — mechanical engineering, accumulated through installed base and time. The emerging criterion is whether the robot can decide: understand a task goal, adapt to variation, and complete complex operations without an engineer programming every move. Fanuc has connected Google’s Gemini as a robot AI assistant; ABB paired RobotStudio with NVIDIA Omniverse; Fanuc and Yaskawa are both in NVIDIA’s Cosmos physical-AI ecosystem; KUKA is trying to codify engineers’ field experience into reusable knowledge through its KAI platform. SoftBank’s stated rationale for buying ABB’s robotics business points the same way: physical AI as the next strategic frontier.

That reframes the threat. The four families’ real problem is not losing a few points of share in the old game — their engineering depth remains formidable. It is that the new game’s rules are set by perception, decision and autonomy stacks, and value there accrues to whoever owns the data loop. A million-unit installed base is suddenly a training corpus, which is an argument for Fanuc’s resilience that a share chart cannot show.

What a buyer should do with this

• Retire the default spec. If your robot specification still says “Fanuc, ABB, KUKA or Yaskawa or equivalent”, the first two words no longer describe the market. Rewrite it around process requirements and validated performance, and let Estun, Inovance and peers bid where the process allows.
• Split your fleet by risk, not by brand loyalty. High-takt, ten-year-life processes still justify the incumbents’ premium. Palletizing, handling, welding and cobot applications are segments where domestic share above 58–85 percent means the reference installations already exist.
• Verify repeatability claims on your parts. ±0.02–0.03 mm static repeatability is a lab number. Ask for it measured across a shift on your part, with the test protocol attached.
• If you supply components, expect consolidation on the demand side. The domestic customers winning share are the ones verticalizing — your reducer, servo and controller volumes are consolidating into fewer, larger, more demanding OEMs. That is pricing power moving away from you; contract accordingly.
• Price the localization, not the logo. KUKA’s 85-percent localized supply chain and Estun’s 10 percent share mean “imported” and “domestic” no longer map to cost, lead time or service depth. Map them again for your own supply base.

Related reading

• Humanoid robot market H1 2026: 36,000 units — where the next share battle is being set up.
• Robot cell three sums: what to check before you buy — the procurement discipline this market shift rewards.
• Hollow-cup motor winding: the wire is the part — how component process depth decides who can actually scale.
• Cable-driven humanoids: a tendon rewrites the BOM — the same story one layer down, happening now.

Sources: market share figures, financial results, the ABB–SoftBank transaction, component localization data (GGII) and the four companies’ AI platform moves follow 做减速机的老欧 (19 October 2026), cross-checked against Mordor Intelligence and company disclosures where cited. The buyer checklist, the consolidation argument for component suppliers and the data-loop reading of physical AI are Latest Robot News editorial.