1971Computer alliance with Hitachi; ~30% of the home market
Fujitsu was born in June 1935 when the telephone division of Fuji Electric Manufacturing was split off as Fuji Tsushinki Seizo (Fuji Telecommunications Equipment Manufacturing). Its parent, Fuji Electric, was itself a 1923 joint venture between Furukawa Electric and Germany’s Siemens — and from birth Fujitsu carried the stamp of European telecommunications technology. Through the prewar and wartime years it built itself on telephone exchanges, and in the postwar boom it grew alongside the surge of national investment in Japan’s telecom infrastructure.
In October 1954 it developed the FACOM100, Japan’s first commercial relay-based automatic computer — the starting point of home-grown Japanese computing, and the moment a second identity, “computer maker,” entered Fujitsu’s DNA. It listed on the First Section of the Tokyo Stock Exchange in 1961, and in June 1967 dropped “telecommunications equipment” from its name to become Fujitsu Limited, redefining itself from a telecom-equipment maker into a computer company. Behind the shift lay a national-policy pressure to domesticate the computer against IBM’s advance into Japan.
President Kanjiro Okada read IBM’s threat coldly and prized getting ahead of technological change — “the business world is a very large living thing,” he said, and one must “take the helm as the moment demands.” In October 1971 Fujitsu announced a computer alliance with Hitachi; the Fujitsu-first, Hitachi-second pairing took roughly 30% of the domestic market, and Fujitsu held the top domestic position.
1974The IBM-compatible bet and world No. 2 mainframes
Revenue (¥ bn, bars) · net margin (%, line)
Source: securities reports & corporate yearbooks
FY1974 · unconsolidated
Revenue$714M
Net income$32M
Net margin4.4%
→
FY1999 · consolidated
Revenue$46.1B
Net income-$119M
Net margin-0.3%
1974Invests in Amdahl — the IBM-compatible bet (Toshio Ikeda)
1980Domestic sales overtake IBM Japan
1985“Beyond IBM” diversification into telecom, AI and PCs
1990Acquires ICL of the UK (80%)
1997Amdahl becomes a wholly-owned subsidiary
1999Fujitsu Siemens Computers formed — a three-pole mainframe structure
The direction of Fujitsu’s computer business was set by the IBM-compatible strategy that Toshio Ikeda — the 1970s engineer widely called a genius — drove through. In May 1974 Fujitsu invested in America’s Amdahl, gaining a development capability and sales network for IBM-compatible mainframes. Japan’s makers were split between pursuing a proprietary architecture against IBM and taking the compatible road; Fujitsu chose compatibility, sidestepping the heavy burden of architecture development while going on the attack to take IBM’s users directly. Where Hitachi and NEC kept a proprietary flavour, Fujitsu followed IBM thoroughly — and in Daisuke Kobayashi’s systems-engineer sales culture it learned early to earn by solving customers’ problems rather than by the machine itself.
The bet paid off. In May 1980 Fujitsu’s domestic sales overtook IBM’s Japanese arm, a front-page story of a home-grown maker passing IBM. Yet overtaking brought its own problem: with its chase-target gone, Fujitsu launched a “Beyond IBM” diversification in 1985 into telecommunications, AI and personal computers, trying to raise a new axis of growth of its own. By the late 1990s it stood on three pillars — communications, computers and semiconductors — one of Japan’s foremost all-round ICT makers.
The compatible line went global. Fujitsu bought ICL of Britain (80%) in 1990, made Amdahl a wholly-owned subsidiary in 1997, and formed Fujitsu Siemens Computers in 1999, spreading its mainframe business across a three-pole structure of Europe, North America and Japan and becoming the world’s No. 2 general-purpose machine maker behind IBM — a position no other firm reached by holding the IBM-compatible niche at global scale. But riding IBM’s ground had a price: as the giant turned from mainframes toward downsizing, the large machines that earned most of Fujitsu’s revenue began to wobble in the mid-1990s, sowing the distant seed of the enormous losses to come.
2008LSI unit hived off as Fujitsu Microelectronics
2009FY08 net loss; Nozoe out, Masami Yamamoto in
2010Semiconductor business hived off (logic later to onsemi)
2014Mobile-phone business spun out (sold 2019)
2017PC business into a Lenovo-majority joint venture
In the year ended March 2002, Fujitsu booked a consolidated net loss of $3.1B (¥383bn). The bursting of the IT bubble and a telecom slump struck all three pillars — communications, computers and semiconductors — at once, one corner of the “electronics shock” that pulled Hitachi, NEC and Toshiba into the red together. In June 2002 the presidency passed from Naoyuki Akikusa to Hiroaki Kurokawa, and Fujitsu’s restructuring began here — the moment the very self-image of the all-round ICT maker showed its limit.
The reform did not end in one pass. The Lehman shock of autumn 2008 worsened results again, and the year ended March 2009 brought a further consolidated net loss of $1.2B (¥112bn). Mid-year, president Kuniaki Nozoe abruptly departed and Masami Yamamoto stepped in — a governance jolt that unsettled the market and drew the restructuring out. Fujitsu handed its hard-disk business to Toshiba and cut its DRAM operations away into Elpida Memory, its revenue sliding from a peak above ¥5 trillion as the slimming-down continued through loss-making years.
The retreat from founding hardware became the constant. In 2010 the semiconductor business was hived off (its logic operations eventually passing to onsemi); in 2014 the mobile-phone business was spun out and later sold; in 2017 the PC business became a joint venture majority-owned by Lenovo. What was left was systems integration, middleware, servers, network gear and the mainframe. Yet through these years Fujitsu was described by what it had let go, not by what it kept — quick to decide what to cut, slow to conceive what to become.
2026To end mainframe sales by the close of fiscal 2030
In June 2019 Takahito Tokita took the presidency — the first Fujitsu chief to come up from its systems-engineer ranks, a manager who knew the reality of the SI floor by feel. In July 2020 the company rolled out job-based HR to all its managers, an early case of the shift among Japanese firms, and set “transformation into a DX company” as its course. The aim of a restructuring twenty years old began, here, to reverse from shrinkage toward growth.
In October 2021 Fujitsu launched Fujitsu Uvance — a cross-industry offering business built on themes such as Sustainable Manufacturing, Consumer Experience, Healthy Living and Trusted Society, layering standardized offerings on top of its industry-by-industry SI sales network. The Mid-term Management Plan 2025, announced in May 2022, put Uvance revenue of $4.7B (¥700bn) by the year ending March 2026 at its core, committing Fujitsu publicly to a self-definition as a DX company — from a firm that keeps discarding to one that sells across industries.
By the plan’s later years the change began to show in the numbers: in its Japan regional services the adjusted operating margin rose from 16.1% to 19.4% as standardized development, tighter quality control and a shift toward value-based pricing each added roughly a point. Then, in January 2026, Fujitsu announced it would end mainframe sales by the close of fiscal 2030 — closing, by its own hand, the founding hardware line that had run from the FACOM100 of 1954, on the reading that social infrastructure is moving to the cloud. Having let go of semiconductors, mobile phones and PCs, and now the last of its founding hardware, Fujitsu means to become a company of systems integration and services alone — with the Fujitsu Kozuchi generative-AI platform and a stake in Canada’s Cohere laid as groundwork for that new core. Whether it has truly become the “ordinary company” that aims for growth will be answered by the profitability still to come.
The light that following brought — and the fragility behind it
The heart of this decision was that Fujitsu did not challenge the world’s giant head-on with technology of its own, but attacked from the flank — with compatibility and service. Rather than shoulder the burden of building an architecture from scratch, it rode the ecosystem IBM had cultivated to conserve its strength, and then set itself apart through human support. That posture, sidestepping the frontal approach, can be read as a highly rational bet for a latecomer among domestic makers seeking to overtake the world’s largest firm. In particular, the systems-engineer sales culture that Daisuke Kobayashi made his main force planted early the idea of earning by solving customers’ problems rather than by the machine itself — and it became the wellspring of the systems-integration business that would later be Fujitsu’s backbone.
Yet the choice of compatibility carried, as its fate, the fragility of the follower. So long as IBM held the ground they both stood on, the moment the giant turned its wheel from mainframes toward downsizing, Fujitsu’s own footing would shake. And so it did: entering the 1990s, the large mainframes that had earned much of its revenue began to wobble, becoming a distant cause of the enormous losses of the early 2000s. The opening that riding along afforded, and the fragility that riding along could not escape, were two sides of one coin. Even so, the 1980 achievement — outdoing the world’s giant in its own home market amid the rough seas of capital liberalization — remains instructive as one answer a latecomer can give when it takes on an overwhelming opponent with limited resources.
The significance of this diversification lies in how Fujitsu faced the loss of direction that arrives after success. As long as it had a target to chase, the single-mindedness of overtaking IBM in mainframes was a source of strength. But once it had overtaken, that single-mindedness had nowhere to go. The “Beyond IBM” move of 1985 can be read as the attempt of a company that had lost its goal to set up a new axis of growth on its own. The strength of the era when it had something to chase, and the difficulty of the era after it lost that, lived side by side in the same company.
Yet to broaden is not the same as to bind together. The three pillars — telecommunications, computers and semiconductors — became the thickness of an all-round strength in good times and, in bad times, three simultaneous losses. The trouble was not that diversifying was itself a mistake, but that the next judgment — how to choose among the businesses it had spread into, and where to narrow them — never came with it. The breadth Fujitsu opened up in 1985 it would, twenty years later, narrow down once again. The question of what to aim for once you have overtaken went on appearing, again and again, in changing forms.
What makes this acquisition interesting is that it internationalized a strength that came precisely from being a follower. Having abandoned a proprietary architecture and chosen IBM compatibility can look like the weakness of owning no technology of one’s own. But so long as IBM’s customers were everywhere in the world, the demand for compatible machines was everywhere too. Fujitsu went after that demand in one stroke, by buying ICL in Europe and Amdahl in North America. In the way it brought Western makers under its wing amid the rallying cries for nurturing home-grown industry, one can glimpse a distinctive cast of mind unbound by the domestic frame.
Yet the compatible line was also one bound to the fate of the very party it followed. As the mainframe itself shrank under the pressure of distributed processing and the cloud, the three-pole structure built worldwide on IBM compatibility lost its meaning as a pillar of growth. Even so, the European and North American footholds gained through the acquisitions outlasted the mainframe age and remained as a base for systems integration and services. The bet of holding a niche on a global scale, while heading toward its end as a mainframe business, can be said to have left behind for later generations the foundation on which Fujitsu operates around the world.
The significance of this rebuild weighs less as a tale of success than as a case that mirrors the difficulty of structural reform. The direction had been clear since 1993: shed the constitution that depended on mainframes and earn instead from software and services — a prescription that, each time it was raised, was dismissed as “pie in the sky.” Hiroaki Kurokawa’s rebuild, too, can be read as one driven first and foremost to stop the losses, never quite reaching an answer to the question of what to keep and win with. The more a company has reached the summit through technology, it appears, the harder it finds it to let go of the experience of success and recompose its business model.
Yet without stanching the bleeding there is no next step. Shrinking the $3.1B (¥383bn) loss and stacking up, one by one, the decisions to let go of the founding hardware businesses left room, twenty years on, to steer toward a concentration on services. It could discard, but drawing growth from the businesses it kept was long deferred — the structural reform that began in 2002 would not turn from shrinkage to growth until Takahito Tokita’s Uvance shift. That crisis response and growth strategy are different things is what these twenty years quietly show.
Quick to decide what to cut, slow to conceive what to keep
To read this serial retreat as no more than a record of defeat is one-sided. Had Fujitsu gone on carrying loss-making hardware, it would have eaten into the earnings of its systems-integration business as well. The decision to push semiconductors, mobile phones and PCs outside can be read as an unavoidable tidying-up, made in order to travel light and concentrate on the businesses that could earn. While Japan’s electronics makers struggled with an image of the all-round manufacturer they could no longer manage, Fujitsu handed down its decisions to discard relatively quickly.
The problem is that, against the speed of discarding, the conception of what to keep lagged behind. What to let go of was settled one after another, yet the question of what to sell from the businesses it kept, and how to grow them, hung in the air for nearly twenty years. Selection and concentration lead to growth only once there is a picture of what to concentrate on. Fujitsu’s twenty years of continuous discarding quietly show that cutting things loose, by itself, does not bring growth back.
The decision that first filled the void left after discarding
The significance of this turn lies less in its novelty than in its sequence and timing. The decisions to cut away the founding hardware had continued for twenty years, yet the question of what to sell from the businesses it kept hung long in the air. Takahito Tokita first changed the personnel system and only then set the Uvance business model in place — a sequence that can be read as remaking the vessel before pouring in the contents. In filling, for the first time, the void born after all that discarding with a concrete axis of growth lies the place this decision holds.
Yet between raising the banner of earning through cross-industry services and making it take root, there is still a distance. Uvance’s numbers are rising, but whether Fujitsu has truly become an “ordinary company” — one that aims for growth — will be answered by the profitability still to come. A company that reached the summit through technology and then took twenty years to let go of that experience of success: on what will it next be chosen? The Uvance turn can be called its first attempt to answer that question head-on.
Each heading links to the full Japanese analysis — background, decision and outcome, with sources.
This is a condensed English edition. The full, source-by-source history — with the detailed narrative, financial tables, shareholders and executives — is maintained in Japanese: 日本語版（詳細）— Fujitsu full history in Japanese →