FANUC In-Depth Research

In public primary materials, FANUC emphasizes its origin in Fujitsu's NC team more than individual heroism. For investors, this is an important signal: the company's DNA is long-term, engineering-oriented, quality-oriented, process-oriented, and B2B-oriented, rather than the fast product narrative of consumer electronics. To this day, it still uses "Genmitsu" (precision) and "Tomei" (transparency) as basic principles and treats "industrial automation" as the straight road it should focus on over the long term. This explains why FANUC is rarely noisy, and why even in the Physical AI era, as it opens up partnerships, it still emphasizes the control layer, reliability, long-term maintenance, and supply responsibility.

The issue price, fundraising size, and listing valuation of the IPO year were not found in sufficiently solid original public evidence that could be directly checked this time, so I will not write a number that looks complete but has not been verified. What can be confirmed is that its listing path was early and its capital-market history is long. That means FANUC's stock price has long been a mirror of Japanese manufacturing cycles and global automation investment sentiment, not the valuation of a "new concept company."

Development Stages

FANUC's first stage was the foundation-building period "from NC to CNC." The core of this stage was technical direction, not revenue scale: the company treated machine-tool control, servo control, and industrial automation as a base layer that had to be deepened over the long term, not as replaceable accessories. Looking back today, CNC is more like the core of FANUC's moat than the industrial robot body because once a control system is written into customers' production lines, machine-tool ecosystems, engineers' habits, and after-sales maintenance systems, the switching cost is far higher than that of a new robot product line. The company's cumulative shipment of 5 million CNC units by 2022 is the most direct result of this technical route.

The second stage was an expansion period "from the base layer to the application layer." FANUC migrated CNC and servo capabilities into robots, ROBODRILL, ROBOSHOT, and ROBOCUT, gradually forming today's three major businesses of FA, ROBOT, and ROBOMACHINE, then tying them together with Service. It did not follow the highly customized systems-engineering-company route taken by some peers. Instead, it standardized underlying control, hardware modules, and reliability as much as possible. The cost is that in some turnkey scenarios it is less flexible than integrators. The benefit is fast replication, large scale, and globally deployable after-sales service. By 2023, the company's cumulative robot shipments had exceeded 1 million units, showing that this platform expansion model has been validated by global manufacturing over a long period.

The third stage was the formation of global capacity and service networks. The company provides support in more than 100 countries, has disclosed more than 280 official service locations, and retains two core service centers in Japan. It has also made multi-site production and BCP a long-term strategy: in recent years it built the Mibu Factory in Tochigi Prefecture as a second CNC production base and established robot assembly systems at headquarters and Tsukuba. The most important impact of this stage for shareholders is that the risk model changed, not that the revenue curve looked better. For capital-goods companies with long supply chains and customers that suffer large losses from downtime, the service network and backup capacity themselves are competitive advantages. Customers buy FANUC in large part to buy "less line stoppage," not "cheap hardware."

The fourth stage was a cyclical period in which post-pandemic expansion and inventory correction intertwined. In FY2022 and FY2023, global automation, semiconductor, and EV-chain capex drove FANUC sales from JPY 551.3 billion to JPY 852.0 billion, while operating margins remained high at 22% to 25%. The company used this period to increase investment in factories, service, research, and human capital. Then in FY2024, the robot inventory correction appeared, full-year revenue slid to JPY 795.3 billion, and operating margin fell to 17.8%. This was capex-cycle earnings elasticity showing in reverse, not a broken moat. In July 2024, management publicly said China's robot inventory had "almost returned to an appropriate level," while inventory adjustments in Europe and the United States were also progressing.

The fifth stage is a new period in which recovery and openness run in parallel. FY2025 revenue recovered to JPY 857.8 billion, a company record, and operating margin rebounded to 21.4%. At the same time, the company's tone on technology route changed. It began to cooperate more actively with external platforms such as NVIDIA and Google. This change matters because it means FANUC has begun to acknowledge, rather than try to keep all upper-layer software and AI capabilities inside its own system, that in the Physical AI era, underlying control and execution still have moats, but simulation, models, agents, and data tools need an open ecosystem. For capital markets, this is also the key inflection point that lifted the valuation narrative from "traditional automation cyclical stock" toward "industrial AI gateway asset."

Key Milestones

One milestone worth highlighting separately is the company's long-term investment rhythm in production capacity. During one boom period, FANUC did not treat single-product ROBODRILL demand as a permanent fact. It actively reduced its bet on some single-hotspot capacity and shifted more resources toward ROBOT and CNC, while building multi-site capabilities. The company's own retrospective describes this directly: in the 2010s, ROBODRILL demand from some IT industries was very large, but the company did not believe this could last, so it chose to tilt resources toward ROBOT and CNC, which had longer-term futures. This is one of the few moments when one can directly see "counter-procyclical capital allocation" in the company's own wording.

The second milestone was the 2024 European ROBOCUT EMC compliance testing issue. In 2024, the company disclosed that European ROBOCUT products might not have completed testing under EMC harmonized standards and established a special investigation committee. For an industrial automation company with a brand image of "high reliability, compliance, and lifetime maintenance," such an incident may not immediately change the income statement, but it directly touches the most sensitive layer of the moat: customers' trust in the quality system and governance culture. The company later made this issue one of the key investor-communication topics, which also shows management knew it was a source of governance discount, not a "technical detail."

The third milestone was the Physical AI partnership in spring 2026. On May 13, the company announced cooperation with Google to apply Gemini Enterprise-driven AI agents to robot systems. On May 15, it announced a further strengthening of its cooperation with NVIDIA, including coupling Isaac Sim with ROBOGUIDE, using Isaac GR00T N for imitation learning, and adopting the Jetson Thor platform. This changed the capital-market narrative because it answered a long-standing market concern, not because it immediately added a lot of revenue to the current income statement: if industrial AI is truly going to land, will traditional robot majors be marginalized by "upper-layer AI platforms + new hardware startups"? At least for now, FANUC has chosen active integration rather than passive observation.

The fourth milestone was the stock price reaching an all-time high in May 2026 on the Google partnership news. The market reaction to this type of move was direct: on 2026-05-14, the stock briefly reached an intraday record of JPY 8,880. In hindsight, part of that reaction was reasonable and part was overheated. The reasonable part is that FANUC is indeed one of the few companies qualified to connect Physical AI from "laboratory demo" to "factory production line." The overheated part is that the company itself has not separately accounted for related orders, and the Q&A clearly warned that Q4 orders cannot simply be extrapolated linearly. For today's investment judgment, this milestone tells you that FANUC's valuation center is no longer just an order cycle, rather than "proving that the AI narrative can keep pushing the stock higher."

Longitudinal Financial Review

Looking at FY2021 to FY2025 together, FANUC's revenue, profit, ROE, and capex all show a clear capital-goods cycle imprint. FY2021 revenue was JPY 551.3 billion, rising to JPY 733.0 billion in FY2022 and JPY 852.0 billion in FY2023, then returning to JPY 795.3 billion in FY2024 and recovering to JPY 797.1 billion in FY2025. Operating margin rose from 20.4% in FY2021 to 25.0% in FY2022, then fell to 17.8% in FY2024 and repaired to 19.9% in FY2025. The latest FY2026 further lifted revenue to JPY 857.8 billion and operating margin to 21.4%. This trajectory shows two things: first, FANUC is not a compounding company whose revenue rises smoothly every year; second, when the upcycle comes, its margins have meaningful operating leverage.

Profit quality has been good in the two most recent directly verifiable years. FY2025 operating cash flow was JPY 255.3 billion versus net profit attributable to owners of parent of JPY 147.6 billion. FY2026 operating cash flow was JPY 250.9 billion versus net profit attributable to owners of parent of JPY 166.5 billion. Both were clearly above net profit. This shows the company's accounting profit is not "paper profit," and collections and cash conversion have not significantly deteriorated. More importantly, during the FY2026 recovery, operating cash flow did not become distorted as profit recovered, showing the current growth is at least not being built through large-scale stretching of receivables or inventory accumulation.

The balance sheet is where FANUC is easiest for the market to underestimate and also easiest to criticize. Its strength lies in extreme conservatism: as of 2026-03-31, cash and bank deposits were JPY 718.1 billion, short-term marketable securities were JPY 35.8 billion, investment securities were JPY 223.3 billion, and cash and cash equivalents were JPY 615.1 billion. Total liabilities were only JPY 207.8 billion, and liability items were mainly accounts payable, taxes payable, warranty provisions, and other operating items, with no short-term or long-term borrowing line item shown. Its weakness is capital efficiency: the latest FY2026 ROE was only 9.3%, clearly above FY2025's 8.6%, but still far from aggressive. This is exactly where it is most likely to be questioned under the TSE capital-cost reform context: is this much cash being held to get through cycles, or because there are not enough high-return reinvestment destinations?

Capex and R&D also show the company's long-termism. From FY2021 to FY2025, R&D expenses roughly stayed in the JPY 46.9 billion to JPY 51.9 billion range, while capex, after a major expansion, had fallen from the FY2019 high of JPY 133.1 billion to JPY 40.1 billion in FY2025. The company has not clearly cut R&D for short-term profit, but it has also not kept capex in an aggressive state. This combination matches its business model: underlying control and product reliability require continuous investment, while capacity should expand slowly only when demand trends are clear.

Stock Price and Valuation History

FANUC's capital-market history can be roughly divided into four phases. The first was 2016 to 2018, when the market treated it as a beneficiary of "China automation + electronics manufacturing capex + Japanese high-end manufacturing," and revenue, profit, and valuation all rose together. The second was 2019 to 2020, when macro, trade, and industry cooling combined to pull revenue and profit down, reminding the market that even the best automation leader is still a capex cyclical stock. The third was 2021 to 2023, when post-pandemic industrial rebuilding, EVs, semiconductors, and labor shortages drove automation and revenue set new highs again. The fourth is 2024 to 2026: first a robot inventory correction, then order recovery and a Physical AI narrative re-rating. After the stock was pushed to a new high by the Google partnership in May 2026, it quickly pulled back because valuation had run up too fast.

Looking again at the shift in valuation center, what really changed is that the market added an option, not that the company suddenly changed from a cyclical stock into a growth stock. In the past, the market's labels for FANUC were more often "high-quality cyclical stock," "automation leader," and "cyclical cash cow." Now at least some capital is trading it as a "Physical AI industrial gateway." This upward shift in the valuation center has a factual basis, but it has not yet been fully validated by the income statement, so the stock's sensitivity to news and orders will be higher than before.

Business Model and Moat

Revenue Structure

First, correct a metric that is easy to confuse in the market. Under the company's FY2026 annual-report statutory presentation, FY2025 full-year revenue by the four major segments was: FA JPY 208.5 billion, or 24.3%; Robot JPY 378.6 billion, or 44.1%; Robomachine JPY 129.6 billion, or 15.1%; and Service JPY 141.1 billion, or 16.5%. The Robot 46.6%, FA 23.4%, Service 15.9%, and Robomachine 14.1% often seen in secondary sources actually correspond to the Q4 FY2025 single-quarter sales mix, not the full-year mix. This distinction matters for judging the business model because a single quarter can magnify the hottest current segment, while the full year is closer to the company's real revenue skeleton.

Source: company FY2026 annual results announcement.

If the question is simply "how does FANUC make money," the answer is not "robots." More precisely, it makes money from three layers. The first layer is the base control layer, such as CNC and servos. They determine how machine tools move, how accurately they move, and how stable they are. Once customers choose the system, introduce it, train around it, and build maintenance around it, switching costs are extremely high. The second layer packages these base capabilities into robots and machine-tool products. The third layer is service. The company clearly emphasizes that as long as customers continue using FANUC products, it will keep providing maintenance support. This means FANUC's profit comes from service, parts, maintenance, and repeat-introduction opportunities that occur throughout the equipment life cycle, rather than ending after a one-time equipment sale.

This also explains why Service appears to account for only 16.5% but is much more important than the surface revenue share suggests. Service is not a side branch. It is the adhesive of the whole business model. The company has more than 280 service locations in more than 100 countries and deploys more than 2,300 service and support personnel globally. For industries such as autos, electronics, medical devices, and precision machining, downtime losses far exceed one-time purchase-price differences, so "can it be repaired immediately," "are parts in stock," and "is there a local engineer" often matter more than the quoted price in explaining why customers choose FANUC.

Cost Structure and Operating Leverage

FANUC's fixed costs are not low. R&D, factories, reliability testing, service networks, parts systems, engineers, and global locations mean it is not a light-asset company that can quickly "slim down" when demand falls. Precisely because of this, operating leverage is obvious when revenue rises, and margins hurt when revenue falls. During the FY2024 robot inventory correction, operating margin fell to 17.8%. With the FY2026 revenue recovery, operating margin returned to 21.4%, and the company's FY2027 guidance even points to 23.3%. This is fixed-cost infrastructure being diluted again during a demand recovery, not a sudden improvement in the profit structure.

On the other hand, FANUC's capex is not a model that must expand at high intensity every year. The company has indeed invested heavily in capacity, BCP, and service facilities in recent years, but it did not keep pushing capex higher after FY2025. It is more like a capital-goods leader with "intermittent heavy investment + long harvesting periods." This structure allows it to benefit more from scale than pure systems integrators and to endure cycles better than new hardware companies that keep burning cash to expand.

Moat

FANUC's first true moat is the control layer and installed base. The company itself does not provide a freely verifiable precise global CNC share in public materials, but it clearly lists CNC as a business with "top global share," and public materials consistently emphasize CNC as the company's foundational technology. More importantly, China's high-end machine-tool control field still depends heavily on foreign suppliers. Public reporting citing analyst data says foreign suppliers still account for about two-thirds of China's machine-tool market, with FANUC, Mitsubishi, and Siemens at about 33%, 20%, and 16%, respectively. This shows that in the most critical, most precise, and least risk-tolerant system-switching scenarios, FANUC still stands in the high-end layer.

The second moat is the integration of "control + servo + robots + machine tools + service." Many companies can sell robots, many can make controllers, and many can do machine-tool software, but few have spent decades connecting all four through the same engineering, reliability logic, and global after-sales system the way FANUC has. Customers continue choosing FANUC in certain applications not only because of single-product performance parameters, but because it can make the entire automation chain more predictable. The company's slogan "Reliable, Predictable, Easy to Repair" is a plain-language expression of this moat.

The third moat is the service network. In industrial automation, service is a base value that determines whether customers dare to place orders, not an auxiliary value. The latest official disclosure shows more than 280 global service locations covering more than 100 countries. A new India service building completed in 2026 and a USD 90 million investment in additional U.S. factories and logistics centers point to the same thing: when the cycle improves, what the company is most willing to add is "being closer to customers," not "concepts." Such investment is not glamorous, but it makes it hard for latecomers to fully penetrate its customer relationships with low prices.

The fourth moat is the balance sheet. Many investors see large cash balances only as "low efficiency," but in the capital-goods industry they also mean something else: when supply chains fluctuate, orders fluctuate, and customers delay investment, the company can still maintain R&D, inventory, spare parts, and global service. FANUC has proven through multiple cycles that it will not be forced to cut critical investment because of a cyclical trough. This is also why its industry position can persist across cycles rather than be displaced by more aggressive players in every downturn.

The pseudo-moat that should not be exaggerated is the recently hot humanoid robot narrative. The company's Physical AI positioning is real, but it has not publicly committed to complete humanoid robots as its main strategy. What can be confirmed for now is that project inquiries are increasing, some large projects are converting into orders, and Google/NVIDIA cooperation is progressing. What cannot be confirmed is whether these moves will become a sufficiently large new profit pool within 2 to 3 years. Treating the "option" as the "main engine" overestimates it; treating it entirely as noise underestimates the possibility of an upward shift in the valuation center.

Management and Governance

The company is currently led by President and CEO Kenji Yamaguchi. Public materials show dense investor-dialogue arrangements: in FY2024, the CEO, CFO, and shareholder-relations department held 323 dialogues, and the board discussed capital cost, ROE, cash levels, long-term strategy, and CEO succession. The company is structured as a company with an audit and supervisory committee, and public disclosure emphasizes that independent outside directors make up a majority of the board. Within Japanese manufacturing, this is already a relatively clear response to the TSE reform on "management conscious of capital cost and stock price."

Governance, however, is not strong enough to receive an unconditional full score. The 2024 European ROBOCUT EMC compliance testing issue already showed that even FANUC's mature quality and compliance systems can have problems. The second point of debate is capital allocation. The company states its shareholder return policy clearly: a 60% payout ratio, flexible repurchases when the stock price is appropriate, and, in general, annual cancellation of treasury stock exceeding 5%. But actual execution is not always as generous as it appears: in April 2026, the board again authorized repurchases of up to JPY 5.0 billion and up to 10 million shares, but in May it bought back no shares. Management explained in the Q&A that a key reason the previous plan was not fully used was that the stock price remained strong, and that it would consider the stock-price level and institutional investor feedback. For long-term shareholders, this means FANUC's capital returns are disciplined, but it will not aggressively repurchase at any price.

Industry and Horizontal Competitors

Industry Structure and Cycle

FANUC is in the "factory automation base," not a single industry. It simultaneously sits in several profit pools: CNC, industrial robots, dedicated intelligent machine tools, and industrial service. Long-term demand comes from three directions: manufacturing labor shortages, rising quality-consistency requirements, and higher automation density in China, North America, and Europe. IFR's World Robotics 2025 data show that global new industrial robot installations reached 542,000 units in 2024, twice the level of ten years earlier; China installed 295,000 units in one year, accounting for 54% of the global total; and IFR expects global installations to rise further to 575,000 units in 2025. In robot density, Western Europe had 267 units per 10,000 manufacturing employees in 2024, North America had 204, Asia had 131, and South Korea reached 1,220. The long-term direction is clear: automation is a structural manufacturing upgrade, not a one-off theme.

This does not prevent the industry from being highly cyclical. For FANUC, the most relevant cycle has always been the superposition of the capex cycle, inventory cycle, and technology iteration cycle, not the consumer cycle. In an upcycle, the main beneficiaries are customer line expansion, new machine introductions, whole-factory automation, machine-tool investment, and robot replacement. In a downcycle, the most vulnerable areas are project delays, channel inventory, slower auto investment, and equipment destocking in electronics. The company itself states directly in its annual report that its products are capital goods and demand is significantly affected by economic changes, so it must operate with a long-term perspective rather than be pulled around by short-term events.

Policy and geopolitics are also becoming increasingly important. The first is the TSE capital-cost reform. It does not directly change orders, but it raises the market's requirements for the efficiency of the company's cash use. The second is tariffs and export controls. In its FY2026 guidance, the company lists U.S. tariffs and their effects on the global economy as uncertainties, and the Q&A also mentions Middle East conditions, oil-based materials, and procurement risks. Looking deeper, industrial automation is also constrained by geopolitical reindustrialization. For FANUC this is a double-edged sword: U.S. manufacturing reshoring may increase demand, while localized production, supply-chain security, and technology compliance can raise investment requirements.

Competitive Landscape

The companies that truly compete head-on with FANUC are a group of players with completely different ecological positions, not a single company.

Yaskawa Electric is the most direct comparison. In official terms, Yaskawa makes AC servos, inverters, CNC controllers, industrial robots, and systems integration. Its business structure is more of an integrated mechatronics route centered on "motion control + robots" than FANUC's. Its strengths are deep accumulation in motors, drives, and motion control, deep roots in welding and some application scenarios, and, as of May 2026, the launch of the new long-term plan Vision 2035 and medium-term plan Dash 35, showing that it is also actively moving from a traditional automation company toward higher-level mechatronics solutions. Why do customers buy Yaskawa? Usually because it has strengths in motion control, drives, and certain application processes, and because it is more flexible in price and solution combinations, not because it looks more like standard infrastructure than FANUC. For FANUC, Yaskawa's threat has always been incremental erosion at the intersection of robots and motion control, not "destroying the whole moat."

ABB is another type of competitor. In 2025 disclosures, ABB described Robotics as the world's second-largest robotics business, with 2024 revenue of USD 2.3 billion. In October 2025, ABB decided to sell its Robotics business to SoftBank at an enterprise value of USD 5.375 billion, abandoning its original spin-off listing plan. By February 2026, ABB's annual report had disclosed Robotics as discontinued operations. Why do customers buy ABB? More because it has deep system capabilities in large manufacturing customers in Europe and the United States, software, simulation, systems integration, and cross-factory automation. The difference between FANUC and ABB is that FANUC is more like a control base and high-reliability mass-production platform, while ABB is more like a robot asset inside a global automation group with software and solution capabilities. ABB's asset disposal also shows that the sector is being repriced internally: robots are no longer just "a department inside a large group," but need to be evaluated independently.

Keyence is not a direct competitor, but it is the capital market's most common reference for "how high a multiple should high-quality Japanese automation assets receive." KEYENCE officially emphasizes USD 7.0 billion in sales, average annual growth above 10% over 25 years, 350,000 customers served, and 250 offices in 46 countries. Why do customers buy Keyence? They buy sensors, identification, measurement, machine vision, and an extremely strong direct-sales engineering system, not large capital equipment. It differs greatly from FANUC: Keyence is lighter, faster, higher-turnover, and more like a "premium tool merchant" in automation; FANUC is heavier, more dependent on installed base, and more deeply rooted in the capital-equipment layer. Therefore the valuation gap between the two companies is essentially the difference between "light-asset high turnover" and "heavy-asset high-moat cyclical leader."

Mitsubishi Electric represents another type of broad FA major. It is a broader industrial electrical and FA system player, not a pure FANUC comparable. Why would customers choose Mitsubishi rather than FANUC? The answer is usually that the whole factory or the whole FA stack is already in the Mitsubishi system, including PLCs, drives, controls, and energy management, rather than because "one product is the best." For FANUC, Mitsubishi is more like an ecosystem competitor than a point-product competitor.

Looking further upstream, precision transmission and reducer companies such as Harmonic Drive Systems and Nabtesco correspond to the components profit pool. Their impact on FANUC is through robot body cost, performance, and supply-chain security, not customer capture. The profit pools FANUC truly competes for directly remain the high-end control and robot platform layers.

Ecological Positioning

FANUC's most accurate industry position is a base-layer automation platform leader, not a "robot company." The gap it fills is this: when customers need a globally replicable, globally maintainable, cross-process extensible automation platform with reliable control layers, FANUC can provide CNC, robots, dedicated machines, and maintenance systems. The profit pools it directly competes for are the control and robot layers held by automation giants such as Siemens, Mitsubishi, Yaskawa, and ABB. The players most likely to compete for its profit pools are two types: low-cost Chinese CNC/robot manufacturers moving upward from the bottom through price, and upper-layer AI/software/simulation platforms trying to "abstract" the control layer. In the short term, the former is more concrete; in the medium to long term, the latter deserves more attention.

If the industry sees technological substitution, price war, or demand decline, FANUC's position usually gets hit first and then repairs. It gets hit first because it sells capital goods and customers can most easily postpone procurement. It repairs later because low-price alternatives find it hard to quickly replace its installed base, service system, and control ecosystem. That is why FANUC's stock often moves more violently than the business, while the business is often more stable than the market fears.

Current Fundamentals and Valuation Analysis

What Happened in the Last Four Quarters and What Is Happening Now

The latest full fiscal-year FY2025 numbers are very clear: revenue was JPY 857.8 billion, up 7.6% year over year; operating profit was JPY 183.8 billion, up 15.7%; net profit attributable to owners of parent was JPY 166.5 billion, up 12.9%; and operating margin rose from 19.9% to 21.4%. The core sales recovery drivers were China FA and robots in China and the Americas. The company's FY2026 guidance continues to lift this recovery line: revenue of JPY 909.6 billion, operating profit of JPY 212.2 billion, and net profit of JPY 184.9 billion, meaning management does not view FY2025 as a one-off rebound year.

Order improvement matters more than revenue. In Q4 FY2025, consolidated orders were JPY 252.0 billion, up 19.2% year over year and 14.5% quarter over quarter. FA orders rose about 40% year over year, and China-region orders rose about 55%. At the same time, in Robot single-quarter sales, China rose 34.6% year over year, the Americas rose 34.8%, and Europe rose 45.1%, while Asia excluding China still fell 49.3%. These numbers show that the current recovery is led by China and the Americas, followed by Europe, while the rest of Asia remains divergent, rather than every corner of the world rising in sync.

The company's Q&A further gave the market a subtle signal. Management acknowledged that part of Q4 orders related to delayed-delivery projects, so the FY2026 forecast is not simply Q4 multiplied by four. At the same time, it said the overall demand environment for FA, Robot, and Robomachine will remain "solid" for a period, and that Physical AI-related inquiries and projects have clearly increased both domestically and overseas, with several-thousand-unit projects gradually becoming orders. This means current fundamentals have a real recovery, but the recovery also contains early orders and project-timing noise. The hardest market judgment lies exactly here.

What the Market Is Trading Now

Today's FANUC is being traded on three things, not one. The first is cyclical recovery. At the hardest point of the 2024 inventory adjustment, investors worried robot orders would fall into long-tail destocking. By spring 2026, the market had begun asking again whether this was the beginning of a new upcycle. The second is manufacturing reinvestment in China and the United States. The company specifically mentioned that China AI semiconductors, data centers, humanoid-robot parts processing, medical devices, and new-energy vehicles are driving machine-tool demand and early CNC orders. In the United States, investment is shifting from pure EVs toward a coexistence of hybrids and ICE, while general industrial demand remains stable and reshoring continues. The third is the Physical AI narrative.

The boundary between real fundamentals and market narrative is whether there is already an independently measurable new profit pool. So far, the answer is no. The Google and NVIDIA partnerships are real, and Physical AI project conversions are real, but the company clearly says it does not separately track the number of units purely attributable to Physical AI. The reason the stock reached a record in May 2026 is that capital markets were willing to assign option value first, not because the income statement suddenly gained a new business. This option is reasonable, but it is also the easiest part to overheat.

Sell-side expectations also show wide disagreement. Google Finance aggregated 12-month target prices roughly between JPY 5,600 and JPY 8,630, with an average of about JPY 7,603. When the current intraday price was around JPY 7,273, the market was no longer treating it as an "obviously undervalued" stock, but as one that needs later order delivery to prove it can move higher.

Bull-Bear Debate

The strongest bullish evidence is that volume, pricing, and profit have not diverged in this recovery. Revenue is up, margins are up, cash flow remains strong, and orders continue rising. China's high-end CNC and robot demand is coming not only from new-energy vehicles, but also from AI semiconductors, data-center parts, medical devices, and general industry. Add the company's thick cash and decent shareholder-return framework, and the market is naturally willing to value it above ordinary manufacturers.

The strongest bearish evidence is that the recovery already contains many components that are easy to overestimate. Q4 orders already included early ordering; the company itself clearly says they cannot simply be multiplied by four; China orders contain AI/humanoid/data-center concepts, but the company also admits information on the low-priced domestic CNC market is incomplete and that it does not cover the most extreme low-price layer; U.S. tariffs and Middle East conditions are still included in annual risks. Most importantly, 35 to 39 times P/E is no longer a "buy first and ask later" position. If the recovery pace is slower than expected, the first pullback will show up in valuation, not earnings.

Historical Valuation and Peer Valuation

FANUC's current valuation is clearly not in a "deep undervaluation zone." Based on the 2026-06-12 closing price of JPY 6,950, it trades at about 38.9 times FY2025 EPS of JPY 178.47, about 35.1 times FY2026 guidance EPS of JPY 198.14, about 3.5 times book value per share of JPY 1,998.45, and a dividend yield of about 1.5%. This set of numbers is not cheap for a capital-goods company, showing that the market has already put "high quality + recovery + AI option" into the price.

Among peers, Yaskawa is also seeing valuation lifted by Physical AI and robot-cycle recovery. The Google Finance page on 2026-06-15 showed its market cap, P/E, and EPS had all clearly risen. Keyence has long enjoyed a higher premium because it is lighter and higher-turnover. ABB's robotics business, by contrast, was carved out of the group, showing that global capital markets are also shifting their pricing of robot assets from "subsidiary business inside a large group" toward "automation asset that can tell an independent story." FANUC's valuation is therefore not isolated, but it is certainly not a bargain being "dragged down" by peers. The cheap companies in today's industry often lack its kind of moat; those with this kind of moat are generally not cheap.

Cash-Flow Look-Through and Absolute Valuation

Start with cash flow. In directly verifiable primary data, FY2024 operating cash flow / net profit attributable to owners of parent was about 1.73 times, and FY2025 was about 1.51 times. The direction is very clear: FANUC's profit turns into cash. The company did not provide a directly copyable five-year OCF table in this text extraction, so I will not write an unverified five-year average, but at least in the past two years it was not a company using accounting profit to dress up the balance sheet.

The company does not publicly split maintenance capex and expansion capex. In the valuation, I take a conservative approach: treat all FY2025 capex of JPY 40.1 billion as maintenance. Under this calculation, owner earnings are about JPY 210.8 billion, corresponding to an owner-earnings yield of about 3.1% on the current market cap and an owner-earnings P/E of about 32.4 times. This is indeed slightly lower than the headline P/E of 39 times, but far from "cheap once viewed differently." Whether the company is expensive depends on the market's already high expectations for the coming years' recovery, AI optionality, and capital efficiency, not on accounting quality.

This is a valuation scenario analysis under the research framework, not investment advice. The range endpoints are derived from the company's FY2026 guidance, cash-flow delivery over the past two years, balance-sheet safety cushion, and the quality premium currently granted by the market.

Expectation Gap and Margin-of-Safety Review

The market's current implied expectation is "the worst is over, the recovery will continue, and Physical AI can raise the valuation center," not merely "the worst is over." If any of these is weaker than expected, the stock can pull back before any profit blow-up occurs. The next item that can truly rewrite the bull-bear judgment is still orders, not press conferences. If China FA and Robot orders maintain double-digit growth for two consecutive quarters, and U.S. general industry plus data-center-related demand can take over, the market will be more willing to view FY2026 guidance as conservative. Conversely, if management again emphasizes "pulled-forward orders" and "hard to predict," accompanied by margin misses, valuation will contract quickly.

Under the conservative scenario above, the current price is not at a discount to conservative intrinsic value. Instead, it carries roughly a 5% to 11% premium, so the margin of safety is zero. The most fragile assumption across the three scenarios is also the valuation multiple, not revenue. If the market stops assigning an "industrial AI option" and returns to a "high-quality cyclical stock" perspective, there can be meaningful de-rating even if earnings hold up. If the valuation multiple in the base scenario is cut by 30%, fair value would compress from JPY 6,800 to JPY 7,400 to roughly JPY 4,800 to JPY 5,200. That is why I think it now looks more like a stock that can be held but should not be chased. Based on these factors, my conclusion on FANUC's current price margin of safety is: none.

Risk Catalysts and Cross-Sectional/Longitudinal Summary

Risk Analysis

The risk that most needs to be taken seriously is discontinuous recovery in Chinese and global manufacturing capex. I assign medium probability and high impact. The transmission path is direct: FA and Robot orders slow first, revenue recognition is delayed, the fixed-cost platform pushes operating margin below 20%, and the market then compresses the valuation from "high-quality recovery stock" back to "ordinary cyclical stock." What needs monitoring is quarterly orders, especially year-over-year growth in China FA and China/Americas Robot, not news headlines. Once double-digit growth turns into two consecutive quarters of zero or negative growth, the story changes.

The second risk is that the market prices the Physical AI option first but profit arrives later or not at all. Probability is medium to high, and impact is medium to high. The company's cooperation with Google and NVIDIA is real, and the company has clearly said several-thousand-unit projects are converting into orders. But it has also clearly said it cannot split out orders purely brought by Physical AI. In other words, the narrative is already ahead of discloseable revenue. If over the next 12 to 24 months this remains only demos without independently perceptible incremental revenue, the stock decline may come not from an earnings blow-up, but from premium evaporation.

The third risk is that low-priced local Chinese CNC and robot vendors move upward from the bottom. Probability is medium and impact is medium to high. The company itself acknowledged in the Q&A that China has many manufacturers and very low-cost CNC, and that FANUC does not cover that layer. Financial Times-cited analyst data also show that China's high-end machine-tool control is still dominated by foreign suppliers, which itself means domestic substitution pressure will persist. In the short term, it erodes the low-end market and price anchor first. In the medium to long term, it may penetrate upward into the high end. FANUC's response is to maintain high reliability, service, and high-end share, not price war. But as long as China's market price system moves lower, its China mix and margins will eventually be affected.

The fourth risk is long-term capital-allocation efficiency improving less than expected. Probability is medium and impact is medium. FANUC's strong cash balance sheet provides a safety cushion, but it also raises the market's requirements for capital efficiency. The company has already written "identifying capital cost and improving ROE" into its governance language, and the TSE continues to push reform around capital cost and stock-price-conscious management. Yet after the latest repurchase authorization was announced, not one share was bought in May 2026. If ROE stays at 8% to 9% for the next few years, cash continues accumulating, and growth investment does not clearly rewrite the revenue curve, the valuation will increasingly look like a discount case of "unfinished governance improvement."

The fifth risk is that quality and compliance issues damage brand trust. Probability is low to medium, and impact is medium to high. The 2024 ROBOCUT EMC issue is worth remembering because FANUC's brand is built on reliability, transparency, and lifetime maintenance, not because it already created huge compensation claims. Industrial automation customers have low tolerance for "possible line stoppage" and even lower tolerance for "possible non-compliance." One incident may not be fatal, but if repeated, valuation will be damaged before profit.

Catalysts and Tracking Indicators

The most important positive catalysts fall into three categories. The first is order delivery continuing to beat expectations, especially China FA, U.S. general industrial robots, data-center-related automation, and Robomachine demand for AI semiconductor parts processing. The second is Physical AI moving from "many inquiries and many partnerships" to "visible revenue, separable products, and repeatable customer cases." The third is substantive capital-allocation improvement, such as actual repurchase execution, progress on treasury-share cancellation, and clearer ROE and cash-use targets.

Negative catalysts can almost all be seen first in quarterly reports: Q4 order heat clearly cooling, management again emphasizing pulled-forward orders and uncertainty, weakening China general industrial demand, discontinuous Robot recovery in the Americas and Europe, gross margin and operating margin failing to return to FY2026 guidance, or Physical AI discussion heat fading without orders taking over. Another external catalyst that cannot be ignored is rates and style. Current valuation is not low. Once global rates rise and the growth style retreats, assets like FANUC, with decent quality but low cash-flow yield, will see multiple compression first.

These indicators matter because they correspond to three fundamental questions: is the cycle continuing to recover; is the AI option becoming monetizable incremental demand; and will capital markets continue granting a high-quality premium? Most can be tracked through company quarterly reports, Q&A, industry data, and management's public statements.

Cross-Sectional/Longitudinal Summary

From the longitudinal view, what FANUC has truly proven is its ability to turn a difficult, boring, but extremely critical industrial base-layer problem into a global standard, not its ability to chase hot themes. After being spun out of Fujitsu's NC team, it spent decades refining just one thing: industrial automation. It first made control systems and servos so deeply embedded that customers did not want to leave, then migrated capabilities into robots and dedicated machine tools, and finally used a global service network to turn installed base into a relationship network that can renew, repurchase, and expand. Its past success certainly benefited from historical tailwinds, including the rise of Chinese manufacturing, electronics manufacturing expansion, EV investment, and rising automation penetration. But without a control-layer moat, quality culture, and service system, it could not have repeatedly converted those tailwinds into market position.

From the horizontal view, FANUC's most real advantage versus peers is that control, execution, machine-tool compatibility, reliability, and global maintenance all hold together, not that every local technical indicator is the best. Yaskawa is more like a strong motion-control and application-process competitor. ABB is more like the robot platform of an international automation group with stronger software and system capabilities. Keyence is a completely different, lighter, faster, higher-turnover automation tool merchant. Why do customers choose FANUC over the long term? Because in factories, stable output matters far more than stage effects. The practical questions of "can it be repaired immediately if it breaks," "will engineers know how to use it when lines change," and "can multiple factories worldwide use the same standard" are its core competitiveness.

These success factors remain today, but the stock price has already prepaid a sizable share of that success. The current valuation discounts two layers of future delivery in advance, rather than rewarding past history: first, order recovery can smoothly convert into revenue and margins; second, Physical AI can at least convince the market that it has a second curve that will gradually grow. The problem is that the company itself says the first cannot simply be assessed by multiplying Q4 by four, and it also says the second cannot be separately tracked as orders purely brought by Physical AI. That is the hardest part of FANUC as a stock today: the company has no major problem, but the price has moved ahead.

Over the next 1 year, the key variables are whether China FA and Robot orders can continue, and whether new demand outside autos from U.S. general industry, data-center automation, and related areas can take over. Over the next 3 years, the key variables are ROE and capital allocation: can the company, while maintaining a high safety cushion, convince the market that cash is a reserve for future growth and returns rather than trapped assets? Over the next 5 years, the key variables are two seemingly unrelated but highly connected questions: first, will the industrial world more deeply accept open Physical AI toolchains; second, can FANUC still defend its high-end control-layer position without being squeezed simultaneously by upper-layer platforms and lower-layer low-price alternatives?

What would make it a better investment? The answer is simple: either the price provides a thicker margin of safety, or the next two quarters prove today's high multiple with orders and margins, rather than the business telling a bigger story. Conversely, under what conditions should the original view be overturned? If China's recovery is clearly just pulled-forward orders, Physical AI stays in press releases for a long time, capital-efficiency improvement remains slow, or quality governance shows new flaws, then investors should admit they bought "a good company priced too early at a high level," not "a high-quality cyclical leader at a reasonable price."

Bull and Bear Reasons

Bull reasons:

• The CNC control layer and global installed base create the deepest switching costs, and public materials still treat CNC as a core global top-share business.

• Q4 FY2025 orders rebounded strongly, and China FA plus China/Americas Robot have moved out of the inventory-correction stage.

• Cash and marketable securities are abundant, the liability structure is clean, and cycle-through ability is significantly stronger than most capital-goods peers.

• Physical AI-related cooperation and project conversions have produced real business contact points, not pure concepts.

• The service network covers more than 100 countries and more than 280 locations. The higher customers' downtime cost, the more valuable this network becomes.

Bear reasons:

• Current trailing P/E is about 38.9 times and forward P/E is about 35.1 times, leaving a very thin margin of safety.

• The company itself warns that Q4 orders cannot simply be extrapolated by four, and the recovery pace contains pulled-forward order elements.

• Physical AI orders are improving, but they still cannot be separately broken out as measurable revenue, so narrative is ahead of delivery.

• Low-priced Chinese CNC and domestic-substitution pressure will first erode pricing systems and mix, rather than causing share loss overnight.

• Capital-allocation efficiency remains a long-term controversy, and the first month after the 2026 new repurchase authorization saw no execution.

Pre-mortem

The first scenario most likely to make this investment lose 50% three years later is "recovery disproven + de-rating." More specifically: in 2027, China's machine-tool and auto-related investment cools, the early orders from spring 2026 do not turn into sustained shipments, and FA and Robot orders turn negative again; U.S. general industrial demand then fails to fill the gap. Operating margin could fall from the company's FY2026 guidance of 23.3% back to 18% to 19%, and the market could value it at 22 to 25 times as a "high-quality cyclical stock," no longer willing to pay more than 35 times forward P/E. If EPS also falls back to around JPY 170 to JPY 180, a stock price of JPY 3,800 to JPY 4,500 would not be extreme. In this scenario, the decline is composed of the double hit of earnings downgrade plus valuation compression.

The second scenario is "AI narrative fails to monetize + China's pricing system moves down." More specifically: by 2028, the Google/NVIDIA cooperation still mainly appears in demos, software connections, and a small number of pilots, without forming large enough separately measurable incremental revenue. At the same time, Chinese vendors continue moving upward from low-priced CNC and robots. FANUC may not immediately lose share in China's high-end market, but pricing and service-bundling ability begin to come under pressure, and net margin cannot return to recovery-period expectations. The market then re-prices it from an "industrial AI gateway" to a "mature automation company with a lot of cash," compressing P/E from 35 times to 20 to 24 times. Even if revenue does not collapse, the stock price could halve.

Final Research Conclusion

If the company and the stock are separated, FANUC is the kind of core Japanese industrial asset I am willing to track for a long time and also willing to hold in part over the long term. Its control layer, service network, balance sheet, and quality culture are all hard to replicate in global automation. The only issue is that today you are buying "a good company with clear quality, clear recovery, and a price that is clearly not cheap," not "a good company that has been wrongly sold off." Investment outcomes can differ greatly between the two.

My main concern is that the pace of growth delivery cannot keep up with the price the market has already assigned, not that it has no growth ahead. Around the current price, investors have already bought China order recovery, U.S. demand resilience, Physical AI cooperation, and a high-quality balance sheet together. To earn attractive returns from here, the company needs to prove these expectations one by one, not merely maintain the fact that "the company is good." For balanced investors, those odds are no longer attractive.

I would change my view in two situations. The first is a price decline to a level with a clear margin of safety. The second is that, without a price decline, the company uses two consecutive quarters of orders and margins to turn today's somewhat expensive valuation into "reasonable in hindsight." Until one of these happens, FANUC is better defined as "hold and wait" rather than "move aggressively."

【Company Profile Score】

• Fundamental quality: High

• Growth: Medium

• Moat: Strong

• Financial stability: Strong

• Management credibility: Medium

• Valuation attractiveness: Low

• Risk level: Medium

• Suitable investor type: Long-term growth / cyclical / value stock

【Investment Rating】

• Rating: Hold

• One-sentence investment thesis: The deepest-moat CNC leader has exited the correction, but the current price already includes the recovery and Physical AI option.

• Three-tier price signals: 【Ideal/Fair Buy Price】5000–5300 JPY Rationale: roughly equivalent to conservative intrinsic value of JPY 6200–6600 after applying about a 20% margin of safety.

• Holdable price: 6000–7500 JPY

• Clearly overvalued price: 9000–9500 JPY

• Current price classification: Holdable

• Is it worth waiting for a better price: Yes. The better buy-trigger range is around JPY 5,000–5,500, preferably with China FA orders not turning negative again. The opportunity cost of waiting is that if FY2026 to FY2027 truly becomes a sustained upcycle, the stock may not return to this range.

• Target holding period: 1–3 years

• Expected annualized return: conservative -2% to 1%; base 4% to 7%; optimistic 10% to 13%

• Maximum loss risk: under the scenario of "recovery disproven + valuation returning to ordinary cyclical stock," a 35% to 50% drawdown within 3 years is possible

• Signals triggering reassessment: Consolidated orders turn negative year over year for two consecutive quarters

• Operating margin stays below 20% for two consecutive quarters

• China FA or China Robot orders fall into clear decline again

• Physical AI remains only cooperation announcements for a long time, with no project conversion or customer cases

• Capital allocation remains conservative, with cash accumulating but no substantive improvement in repurchases or ROE

【Valuation Range】

• current: 6950 (as of the 2026-06-12 close)

• bear (conservative · ideal buy zone): [5000, 5300]

• base (reasonable · acceptable holding zone): [6000, 7500]

• bull (optimistic · above clearly overvalued line): [9000, 9500]

Additional Information

Key Data Tables

Source: FY2021-FY2025 from the company's Integrated Report 2025; FY2026 from the 2026-04-24 annual results announcement. FY2024 cash and cash equivalents are the prior year-end balance disclosed in the FY2025 annual report.

Source: company FY2026 annual results announcement.

Research Uncertainties

• I could not confirm the 1976 IPO issue price, fundraising size, and listing valuation in directly verifiable primary public materials this time, so I did not write them.

• For the common claim that "FANUC has 50% to 60% global CNC share," the primary public language I could verify this time only went as far as "top global share." More specific precise public figures lack sufficiently solid free primary support, so the body uses more conservative wording.

• Instant valuation metrics for Yaskawa, Keyence, and Mitsubishi Electric are not fully consistent across different public pages, so the horizontal section emphasizes business ecological position and capital-market qualitative comparison rather than forcing a multiple table that looks precise but has high noise.

• Physical AI order delivery still mainly comes from management descriptions and partnership announcements, and has not yet formed independently separable revenue disclosure, so the valuation can only treat it as an option, not a certain second curve.

References

• FANUC 2026-04-24 "Consolidated Annual Financial Results For the Year Ended March 31, 2026," supporting materials, and Q&A summary.

• FANUC "Integrated Report 2025," company history, governance, service network, and long-term financial highlights.

• FANUC 2026 Google / NVIDIA partnership press releases, India service building, and U.S. expansion announcements.

• IFR "World Robotics 2025" press release and executive summary.

• Tokyo Stock Exchange follow-up documents on "management conscious of capital cost and stock price."

• Financial Times reporting on China's high-end machine tools and CNC competitive landscape.

• ABB annual reports, announcements, and Reuters reporting on the robotics business separation/sale.

• Yaskawa official company homepage and 2026 long-term/medium-term plan information; Keyence official company overview.

Other Securities Mentioned in the Report

• 6506.TSE — Yaskawa Electric, the most direct Japanese robot and motion-control comparison

• 6861.TSE — Keyence, the key valuation reference for "high-quality automation assets" in capital markets

• 6503.TSE — Mitsubishi Electric, a broad FA ecosystem and control-system comparison

• ABB.US — ABB, a core global second-tier robot competitor whose robotics assets are being independently re-rated

• 9984.TSE — SoftBank Group, the proposed acquirer of ABB Robotics, reflecting capital restructuring of global robotics assets

• 6702.TSE — Fujitsu, FANUC's historical origin and the background to the 1972 spin-off

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.