FANUC Is Known for Robots. FA Generated 63% of Its Order Growth.

The fastest-growing orders were not for finished robots

Mention FANUC and many people picture yellow industrial robots: welding and assembling vehicles, or automating repetitive tasks across factories and logistics operations. That image is justified. But its October 23, 2026 results presentation contains a less familiar story.

In July–September, total orders rose 49.0% year on year. Robot orders increased 12.6%. Yet FA—the business supplying CNC and servo technology to control machine tools—grew much faster, with orders up 130.1%.

A year earlier, robot orders had been almost twice the size of FA orders. This quarter, FA was larger. That does not mean FANUC is abandoning robots, or that growth in the robot industry has stopped. It shows how investment in AI data centers, semiconductors and humanoid-related manufacturing can create faster demand for production equipment and its controls than for finished industrial robots.

There is a second development. On October 22, the day before its earnings announcement, FANUC disclosed initiatives to make detailed machine data easier to collect with machine-tool builders and to combine factory information with Microsoft Copilot.

Read alongside the CNC and servo order surge, these separate announcements raise a question: will FANUC’s customer value end with making a machine move, or extend to understanding why it stopped and how its operation could improve?

1. Inside the 49% increase, a different demand mix emerges

Page 12 of FANUC’s official presentation reports quarterly orders by business. The following compares July–September 2025 with July–September 2026. Amounts below are in billions of yen, converted from the Japanese presentation’s ¥100 million units; growth rates are those reported by the company.

FA: ¥48.8bn → ¥112.2bn; +130.1%.

Robot: ¥91.8bn → ¥103.4bn; +12.6%.

Robomachine: ¥29.7bn → ¥51.2bn; +72.3%.

Service: ¥35.1bn → ¥39.3bn; +12.0%.

Total: ¥205.3bn → ¥306.0bn; +49.0%.

The ranking changed. In the earlier quarter, robot orders of ¥91.8 billion comfortably exceeded FA’s ¥48.8 billion. In the latest quarter, FA reached ¥112.2 billion against ¥103.4 billion for robots.

FA’s reported share of total new orders rose from 23.7% to 36.7%, while the robot share fell from 44.7% to 33.8%. Robot orders grew in absolute terms; other businesses simply expanded faster.

One quarter does not establish that FANUC has become FA-led for the first time in its history, or that its long-term business structure has permanently reversed. It does reveal where recent incremental orders are concentrated. Displayed amounts are rounded, so the sum of business figures can differ from the reported total.

2. FA supplied approximately 63% of the increase

The total 49.0% order increase is significant. Breaking down the increment is more revealing. Total orders rose from ¥205.3 billion to ¥306.0 billion, an increase of approximately ¥100.7 billion. FA increased from ¥48.8 billion to ¥112.2 billion, adding ¥63.4 billion.

¥63.4bn ÷ ¥100.7bn = approximately 63%.

In other words, FA generated roughly 63% of FANUC’s total year-on-year order increase. Robots added ¥11.6 billion, approximately 12% of the total increment.

Robomachine also matters. Its orders rose from ¥29.7 billion to ¥51.2 billion, adding ¥21.5 billion. Combined, the FA and Robomachine increments were ¥84.9 billion, or approximately 84% of the total increase.

These are not interchangeable businesses. FA covers CNC and servo controls and drives. Robomachine includes ROBODRILL machining centers, ROBOSHOT injection-molding machines and ROBOCUT wire electrical-discharge machines. Both, however, are distinct from finished industrial robots.

The recent increase in FANUC orders was therefore driven much more by equipment and controls that manufacture components and products than by robots themselves. That is a different observation from simply saying that robot demand is rising.

The 63% and 84% figures are Banseog calculations from rounded amounts shown in the presentation. They are not contribution percentages disclosed by FANUC, and they do not measure the contribution of individual end markets such as humanoids.

3. What does FANUC’s FA business actually sell?

FA stands for factory automation, but treating this segment as everything that automates a factory would be misleading. FANUC’s official product classification centers on CNC and servo systems.

CNC—computer numerical control—directs a machine tool’s machining movements using numerical instructions. Servo systems precisely control motor position, speed and motion. Consider a metal-cutting machine: its cutting tool must follow a particular path, at a particular speed and with sufficient positional accuracy. CNC and servo technology can provide those capabilities.

Robot manufacturers must make bodies, joints, drive components and other precision parts. Semiconductor-equipment manufacturers also need high-precision mechanical components. AI data-center equipment uses manufactured parts associated with cooling, power, connectivity and mechanical assemblies.

Increasing production of those components and systems can require machine tools and other manufacturing equipment. Demand for a final product can therefore translate into investment further upstream in production capacity. FA participates at a different point from an order for a finished industrial robot.

Sometimes the capacity to manufacture a product must expand before the product can be supplied at scale. That is not a universal sequence: spare capacity in existing equipment and customer investment schedules can change when new machinery is needed.

4. FANUC itself named data centers and humanoid-component machining

There is no need to attribute the increase to a vague AI boom. FANUC identifies specific areas of demand in its official results presentation.

For FA in China, the company cites AI data centers, semiconductor-related demand, humanoid-robot component machining and new-energy vehicles. In Japan, South Korea and Taiwan, it describes strong demand from AI data centers, semiconductors and aerospace. Automotive and infrastructure demand support India; aerospace supports the Americas; Europe shows signs of gradual recovery.

This range matters. The FA surge cannot be attributed exclusively to humanoid components. At the same time, the company explicitly refers to the machine-tool demand involved in making humanoid parts, rather than only to orders for complete humanoid robots.

A growing robot market can create opportunities for component makers and the suppliers of their production equipment, not just the companies assembling robots. Public materials do not show what percentage of the FA increment came from humanoid-related orders, so assigning the entire increase to that industry would be inaccurate.

The point is not to invent an end-market contribution ratio. It is to identify how the spread of AI and robotics can create demand in established machine-tool and manufacturing-control businesses.

5. Geography tells a different story in each market

FANUC’s total orders by region also moved at different speeds. Year-on-year changes for July–September 2026 were: Asia excluding China, +111.9%; China, +79.8%; Japan, +54.3%; the Americas, +15.0%; and Europe, +5.3%.

These are regional figures for all businesses combined, not FA-only regional growth rates. Asia excluding China’s 111.9% increase cannot be presented as a 111.9% increase in that region’s FA orders.

Nevertheless, total orders clearly increased quickly in China and the rest of Asia. FANUC describes strong machine-tool demand in China, while robots and automotive-related automation investment are important in the Americas.

The same company therefore has different growth centers across regions. CNC, servo and machine-tool demand can be particularly relevant to manufacturing investment in China and Asia; general-industry automation and automotive robot investment matter considerably in the Americas.

A company-wide growth rate cannot explain every regional industrial market in the same way. Japanese component and equipment suppliers using FANUC as a demand indicator should separate customer industries and geographies instead of extrapolating from the headline total.

6. Robots still lead FA in recognized revenue

The reversal concerns new orders—not sales recognized in the quarter. Robot revenue remains larger. FANUC’s July–September 2026 figures show the distinction clearly.

FA: orders ¥112.2bn; sales ¥62.5bn.

Robot: orders ¥103.4bn; sales ¥98.2bn.

Robomachine: orders ¥51.2bn; sales ¥41.6bn.

Service: orders ¥39.3bn; sales ¥40.3bn.

Robot sales were still substantially above FA sales. Both grew year on year: robots by 7.2% and FA by 15.4%. It would be inaccurate to say that FANUC’s main business has already switched completely from robots to FA.

Orders rising faster than revenue may signal more business to be recognized later. But subtracting sales from orders does not, on its own, establish the actual change in backlog, shipment timing, cancellation risk or profitability. Those assessments require backlog data, shipping schedules and product-specific capacity information.

What is supported is narrower: robots remain a large revenue business, while recent incremental customer orders were stronger in FA and Robomachine. Keeping the measures separate reveals a possible direction of change without overstating it.

7. From knowing that a machine stopped to understanding why

On October 22, FANUC announced that cooperation with machine-tool builders would expand the data capabilities of FIELD system Basic Package, or FsBP. The existing platform collects information from machine tools and industrial robots to visualize equipment conditions and support production improvement. It is designed to collect data from equipment made by other suppliers as well as FANUC.

A basic operating signal can tell a factory whether a machine is running or stopped. Understanding why it stopped, which process took longer or whether a tool was replaced at the right time requires more detailed information. Different machines can use different signal addresses and meanings, leaving customers to investigate and configure each connection separately.

The announced initiative would pre-register machine-specific signals defined by participating machine-tool builders in FsBP. Customers could select the machine they use and more easily access the detailed machine-condition data provided by its manufacturer.

FANUC lists production results, good- and defective-part information, tool changes, tool life, robot waiting time and pallet-change waiting time among the data to be used. This is more than adding another sensor: producing a signal and understanding what that signal means are different problems.

Imagine a machine stopping for 20 minutes. An uptime record only establishes the stoppage. Tool-change information, robot waiting time and pallet-change delays can help distinguish its cause. The appropriate response might be a faster machine cycle, a different tool-replacement plan or a revised sequence between the robot and machine tool.

FANUC’s initiative aims to reduce the investigation and individual configuration needed for this kind of analysis. It does not establish that every machine-tool manufacturer is already integrated. Coverage by manufacturer, actual customer adoption and demonstrated cost savings still require verification.

8. Microsoft Copilot addresses a different information gap

The other October 22 announcement concerns combining shop-floor data with enterprise work information using Microsoft Copilot and Microsoft’s cloud and AI services. FsBP handles equipment data, but a company also has maintenance histories, meeting materials, work reports, previous fault-response documents and employee communications.

The knowledge needed to resolve a machine problem does not live exclusively in sensor readings. Records may explain what happened during a similar fault, which part was replaced and why an employee chose a particular response. FANUC is exploring how to connect these two forms of information.

Illustrative questions include: Why has this machine stopped more often than last week? Has a similar abnormality occurred before? What action was taken to resolve it? Answering them may require both current equipment conditions and the company’s accumulated operating knowledge. These questions illustrate the possible workflow, not independently verified product performance.

FANUC brings manufacturing-equipment and control-data expertise. Microsoft provides technologies for enterprise documents, collaboration and AI services. Connecting them could support a broader workflow than monitoring a dashboard alone.

The release says concrete use cases and architecture are being examined. It does not show that the Copilot functionality is available to every FsBP customer, or that savings have been demonstrated across customer factories. Nor does it announce that AI will autonomously take control of machine tools. The verified stage is an initiative to combine information and assist work.

9. Together, the announcements change the strategic question

Three developments can be considered together: the October 23 report of 130.1% FA order growth; the October 22 plan to make machine-specific signals easier to use; and the same day’s proposal to connect factory data with enterprise knowledge through Microsoft Copilot.

They do not constitute an already proven, unified business model. In particular, the October 22 announcements cannot be cited as the demonstrated cause of orders booked during the preceding July–September quarter. Current hardware demand and the future data strategy must remain distinct.

FANUC’s established strengths include the CNC and servo technology that makes machine tools perform precise movements. Machine-tool builders and factory operators use those controls to produce real products. The existing customer value is rooted in making the equipment execute the required motion reliably and accurately.

If FsBP adoption expands, customers may also analyze production, quality and maintenance information across machines. Linking enterprise documents could make past problem-solving knowledge accessible alongside those signals. The potential progression is from controlling a machine, to understanding its condition and output, to connecting multiple assets and operating knowledge to support improvement.

The key is not the AI model’s name. It is the structure that collects data from different manufacturers and makes its meaning usable. Reading a signal address is not the same as explaining a production or quality problem; reducing that gap could lower the effort of adopting a data service.

The effect on FANUC’s profitability, software subscriptions or long-term contracts remains unverified. FsBP already exists, but that fact does not prove the commercial results of the newly announced signal expansion or Copilot integration. A possible future value proposition must be distinguished from realized business performance.

10. An opportunity—and a choice—for Japanese machine-tool builders

Machine-tool builders understand the detailed data of their own equipment: which signals relate to machining problems, and which condition information helps quality control and maintenance. Making that knowledge usable through a common platform could make it easier for customers to compare and manage equipment from multiple suppliers.

But it also raises questions about responsibilities. Which data will the builder provide? How far will signal meanings be standardized? How will the factory customer control access and permitted uses? Who will turn the data into useful recommendations—the builder, platform supplier or an outside software company? Which actual customer costs will the cooperation reduce?

FANUC’s announcement does not establish that ownership of customer data transfers to FANUC, or that participating builders lose their customer relationships. Equally, a shared platform does not automatically deliver the same economic value to every participant.

Customers need demonstrated operating improvement. Builders and platform providers must consider how the costs and responsibilities of integration are shared. For specialized Japanese manufacturers, the decision is what to share to expand the market, and what to retain to preserve distinct competitive value.

That question may also apply to measurement equipment, inspection systems and semiconductor-production equipment. Whether the same structure holds in a particular precision industry must be tested separately, not assumed from the FANUC case.

11. The capabilities involved extend beyond robot development

The results establish growing orders for manufacturing-related equipment. Relevant capabilities may therefore extend beyond engineers who develop robot bodies: CNC and motion control, servo systems, mechanical design, machine-tool applications, manufacturing quality, factory automation and industrial data integration can also matter.

Connecting factory data with AI requires several types of expertise. Someone must understand the physical machine signals. Someone must know what those signals mean for production, quality and maintenance. Others must collect data across manufacturers and organize it into a usable common structure, then connect the result to a system that factory employees can actually use.

These are not all the same occupation. Value depends on their coordination. Yet neither the order increase nor the technology announcements prove that FANUC or its partners are currently adding headcount in those roles.

Evidence of a hiring shift would require actual vacancies, organizational changes or expanding customer projects. Investment direction can suggest how future capability requirements may change, but it is not a substitute for that evidence.

Focusing only on finished-robot development risks overlooking the capabilities needed to manufacture robots and advanced components, and to operate the systems producing them.

12. What remains unconfirmed

First, the 130.1% FA order increase covers July–September 2026. There is no verified causal link from the software-related announcements made on October 22 to those earlier orders.

Second, the company identifies multiple demand sources, including AI data centers, semiconductors, humanoid parts and new-energy vehicles. It does not disclose each end market’s precise order contribution.

Third, FA overtook robots in new orders, while robots remained larger in recognized quarterly sales.

Fourth, the practical coverage of pre-registered machine signals, the participating manufacturers and customer operating results require further confirmation.

Fifth, Copilot integration is an initiative under development, not evidence of completed large-scale commercialization.

Sixth, future FsBP revenue, profitability and incremental data-service income have not been established.

The conclusion is not that FANUC has stopped being a robotics company and become an AI software company. New demand is accelerating in its established machine-tool control business while it seeks to expand the usefulness of factory data. The importance lies in considering current results and future strategy together without treating them as the same thing.