Fujitsu — Company History

Splitting from Fuji Electric: a telecommunications maker sets out

Fujitsu's business begins at the point where Fuji Electric started assembling and repairing telephones in 1930. The parent, Fuji Denki Seizo (富士電機製造), had been established in 1923 on the basis of a technical tie-up between Furukawa Electric and Siemens of Germany, and imported, manufactured and sold electrical machinery and telecommunications equipment of every kind. As the spread of telephone networks lifted demand for communications equipment, in June 1935 Fuji Electric separated the division and founded Fuji Tsushinki Seizo (富士通信機製造). Three product lines came across — telephone switching equipment, telephones and loading coils — and in 1937 the manufacture of carrier transmission equipment was added. In November 1938 the head office moved to Kamiodanaka in Kawasaki, Kanagawa, and in May 1949 the company listed on the Tokyo Stock Exchange on the day trading there reopened.

The shape of a full-line communications and electronics manufacturer settled after the war, and its first step was the move into electronic computers in May 1951. The technical tie-up with Siemens was revived in April 1952, radio communications equipment followed in August 1953 and electronic components in April 1954. The company moved early into applications as well, completing Japan's first automatic control unit for machine tools in November 1956. In 1964 it exhibited the FACOM 231 — the only computer in Japan built purely on domestic technology developed in-house — at the New York World's Fair, where it won high marks from manufacturers and users at home and abroad. Shinko Electric Industries became a group company in June 1957; the company listed on the Osaka Stock Exchange in December 1960 and on the Nagoya Stock Exchange in October 1961; and in May 1962 it established Fujitsu Laboratories, holding the work of development inside the firm.

The threat of IBM in Japan, and the pressure to build at home

By October 1959 the application for IBM's plan to enter Japan was already a matter of fact, and building a co-operative structure that would let the country move quickly to domestic production had become urgent. In June 1960 the outline of the trade and foreign-exchange liberalisation plan was settled, and Japan, as a member of the international community, was required to adapt actively to a liberalised regime. The gap in scale was absolute: as of 1962 IBM employed 75,000 people, had booked revenue of $1,694.96m the previous year — 22nd largest in the United States — and held branches in 93 countries through wholly owned subsidiaries; no Japanese company, it was said, bore comparison. Against that, the computer divisions of Japan's seven makers came to fewer than 7,000 people even fully mobilised, and with liberalisation approaching the question was where on earth are we to look for a way out?

President Okada Kanjiro (岡田完二郎) placed the axis of management on the capacity to respond to change. The economy is a very large living creature. An aggressive policy is fine, but unless you act while the wave is with you it will fail. It is a matter of steering as the moment requires, he said (Keizai Jidai, May 1963), and went on: Look at Fortune magazine ... and more than half the companies that were inside the top hundred thirty years ago have dropped out. That is because they could not keep up with the change (same). He was a manager of the type who did not follow the technical detail himself, but reached his final judgement from the loudness of the reporting engineer's voice and the light in his eyes. In June 1967 the company changed its name from Fuji Tsushinki to Fujitsu, putting on its signboard the shift from a pure telecommunications maker to an integrated communications and electronics manufacturer. In the year to March 1968 sales divided evenly: switching equipment and telephones at 27% and radio and transmission equipment at 23% made communications 50%, while electronic equipment at 27% and components at 15% made electronics 42%.

Japan's domestic computer makers held that the technology itself had been developed, yet felt the threat of the adoption of the IBM System/370 series, regarded at the time as the most advanced machine there was. The domestic line-up was rearranged under that pressure. In October 1971 a tie-up in the computer business between Fujitsu and Hitachi was made public. Fujitsu stood first in sales and Hitachi second, and this alliance of the top two was reported as taking 30% of the domestic market and forming a body able to stand against IBM at home. With capital liberalisation three years away, momentum for reorganisation was rising rapidly across the computer industry. Within the state-led framework for nurturing a domestic computer industry, Fujitsu secured the leading position at home.

The Amdahl investment and selling through systems engineers

At the fork between developing a proprietary architecture in-house and taking IBM's customers with compatible machines, Fujitsu chose the latter. In May 1974, led by Ikeda Toshio (池田敏雄), the engineer known as a genius, it invested in Amdahl of the United States and gained both a development capability and a sales network for IBM-compatible mainframes. The investment itself dated back to 1972, and the trading relationship — component supply and OEM production of some models — was already close. The posture was to take IBM's users directly while avoiding the heavy burden of architecture development, and here Fujitsu parted ways with Hitachi and NEC, which kept designs of their own. In an age when the state was nurturing domestic computers, it was a late-comer's wager on riding the ecosystem of the world's largest maker.

Morale on the ground ran high, and president Kobayashi Taiyu (小林大祐) called his people a crazy crowd. A group that cut its honeymoons short to get back to work was said to be holding its ground without a step backwards against IBM, the mammoth of the world (Nikkei Business, 30 August 1976). Kobayashi put the shape of selling on support rather than price: what mattered was the after-sales service that let customers get results from the machines, and for that the company had to keep a corps of systems engineers ready to be supplied the moment a customer asked. In May 1980 it was reported that Fujitsu's domestic sales had passed those of IBM's Japanese subsidiary, and the news drew attention as the final stage of a pillar of Japanese industrial policy — building a computer industry with international competitiveness.

The emptiness after overtaking, and diversification beyond IBM

Even in a single national market, passing IBM's sales was a considerable event, and Fujitsu drew particular notice from abroad. Yet an age had already arrived in which being the top computer maker did not, on its own, lift a company's image. The collapse of the IBM myth posed a more fundamental problem for strategy. The established system — a large computer installed in a computer room, processing information centrally — began to shake as distributed processing on personal computers and the like spread, and cracks appeared in the general-purpose-computer-centred order that the IBM machine symbolised. With the growth of data communications, self-contained systems grew more fluid still, and an era arrived in which computing and communications together aimed at higher-order information processing. All of it forced a correction of the excessive tilt towards general-purpose computers that the years of chasing IBM had produced.

Once the target is a composite enterprise such as NEC, with its C&C banner, merely chasing furiously gives no measure of the range, and a strategy becomes necessary. In November 1985, under president Yamamoto Takuma (山本卓眞), Fujitsu raised the slogan beyond IBM and drove a product offensive into new markets outside the mainframe — communications equipment, AI and personal computers. In June of the same year its personal computer strategy was reported as a new-product offensive that astonished the distribution market. It had depth in people too: among the 1984 graduates of the University of Tokyo's departments of electrical and electronic engineering, NEC took eight, Toshiba and Mitsubishi Electric seven each, and Hitachi and Fujitsu six each. In February 1986 it set up NIF as a joint venture with Nissho Iwai, planting the seed of the business that would become Nifty. Breaking out of a one-legged reliance on mainframes gave the depth of a three-pillar structure and, at the same time, created a structure that carried communications, computers and semiconductors together.

ICL and Amdahl: buying the world's second-largest SE force

In October 1990 Fujitsu acquired 80% of ICL of Britain, making it a group company the following month. ICL had begun as a British state-sponsored company and, unlike Fujitsu with its IBM-compatible line, built mainframes to its own specification. After the acquisition it withdrew from mainframe manufacturing in stages, and in January 1997 sold D2D, the manufacturing arm that had handled OEM production of personal computers and similar products. ICL's real business became system services for computers supplied on OEM terms by Fujitsu, and 10,000 of its 20,000 employees were systems engineers. The pattern of letting manufacturing go and devoting itself to sales and service was tested here first.

By 20 September 1997 Fujitsu had acquired, through a tender offer costing $913.9M (¥111bn), all the shares of Amdahl of the United States, in which it held 42%, making it a wholly owned subsidiary in October. The roughly $413.2M (¥50bn) of goodwill arising from the purchase was to be amortised at $41.3M (¥5bn) a year over ten years, and interest-bearing debt rose past $14.9B (¥1.8tn) with the money raised for Amdahl. A record profit — which would have been the first since the $373.1M (¥89bn) of the year to March 1985, thirteen years earlier — was given up as a result. The aim, even so, was clear: Amdahl had bought DMR of Canada in 1995 and Trecom of the United States in 1996, with the result that 6,000 of its 10,000 employees were systems engineers and system services accounted for 67% of the previous year's sales. The purchase left Fujitsu with 20,000 systems engineers in Japan, 10,000 in Europe and 6,000 in North America — 36,000 in all, second in scale only to IBM's roughly 60,000 worldwide.

Mainframes falter, and the break with building hardware in-house

Fujitsu had grown together with the general-purpose computer, but that course reached a turning point in the early 1990s. The large mainframe at the heart of a computer division that had earned more than 70% of sales began to wobble, as the advance of downsizing — symbolised by IBM's fall into loss — shook the standing of the big general-purpose machine. Sekizawa Tadashi (関澤義), who became president in 1990, led 50,000 people under a banner of hands-on management and in 1993 declared a break with doing hardware in-house, alongside a policy of taking cost out from the design stage. In April 1991 the company began selling mobile phones, widening the base in communications. Whether the restructuring of this period succeeded turned on whether Fujitsu could move away from a hardware-centred culture quickly enough, while demand for mainframes still remained.

In April 1998 Sekizawa handed the presidency, as the form book had predicted, to Akikusa Naoyuki (秋草直之), who came from the software side. Improving earnings and lightening debt were set as the immediate tasks of the new regime, and in April 1999 the company spent $228.4M (¥26bn) to buy Nifty. In October of the same year it established Fujitsu Siemens Computers as a joint venture with Siemens of Germany, broadening its European base. Amid talk of the end of the era of five integrated electrical manufacturers, in August 1999 Fujitsu and Hitachi were named among the winners. Even so, the Fujitsu of this period still carried a financial structure with more interest-bearing debt than IBM's.

¥300bn of restructuring and a ¥382.5bn loss

At the press conference on the structural reform plan on 20 August 2001, Akikusa said: On the premise of zero growth, we are liquidating various past businesses and plants. In the fashionable phrase, it is a reform with a solid backbone. A net loss of $455.9M (¥55bn) had been posted in the first quarter, the full-year net loss was expected to reach a record $1.8B (¥220bn), and $2.5B (¥300bn) of restructuring charges were the main cause. The plan put $1.2B (¥145bn) — about half the special charges — into semiconductors, the source of the deterioration, cutting three of the eleven domestic front-end lines and, in back-end work, disposing of two of the seven companies in Asia. Headcount was to fall by 16,400 — 11,400 overseas and 5,000 in Japan — and 4,700 people were redeployed, 1,400 of them moved into sales and systems engineering. The goal was a V-shaped recovery to $3.5B (¥400bn) of operating profit and a return on equity above 10% in fiscal 2003.

Doubts remained even about a restructuring plan that looked bold, because the withdrawals went no further than desktop hard disc drives: the American plant was not sold outright, the domestic plant only just brought on stream was kept, and the reluctance to give up the flash memory business stood exposed. Information equipment plants were consolidated but neither closed nor sold, and most of the job cuts fell overseas. What this distress laid bare was that the structural reform proclaimed for a decade — escaping a hardware-dependent constitution and making software and services the pillar of earnings — had been no more than a picture of a rice cake. In the year to March 2002 Fujitsu posted a consolidated net loss of $3.1B (¥383bn), a scale without precedent, and fell into straits in which talk ran for a time even of being sent to the Industrial Revitalization Corporation. In April 2002 it split off its server and storage-system businesses jointly with PFU to establish Fujitsu IT Products.

Killing loss-making projects and turning to a profitable structure

In June 2003 Kurokawa Hiroaki (黒川博昭), who had come up through systems engineering, took the baton from Akikusa. In the year to March 2004, with the restructuring of the hardware divisions in sight, loss-making system development projects broke out in the information services business that was supposed to be the main pillar, and $517.6M (¥60bn) of extraordinary losses was booked. It was the result of an industry fighting over a small pie without regard to margins at a time when customers' appetite for system investment had cooled; Fujitsu had even bid ¥1 for a local authority's system. In fiscal 2004 a large project involving more than a hundred systems engineers was expected to run $183.1M (¥20bn) into the red, and because people kept being thrown at it to see the development through, costs finally swelled fourfold. The deeper the loss, the more hands were needed; the more hands were put in, the more the loss grew.

The remedy was the SI Assurance Division, a special unit reporting directly to Kurokawa, in which some fifty of the company's ablest systems engineers examined the progress of large projects worth billions of yen, step by step. It was given the authority to propose declining the contract to the customer where no improvement could be expected, and a rule was applied under which business with a cost ratio above 90% would not, as a principle, be taken. The value of loss-making projects in fiscal 2005 fell to $103.5M (¥11bn), less than a third of the previous year. The tidying of businesses went on: in March 2005 the plasma display module business was transferred to Hitachi, and the following April a contract was signed to transfer the liquid crystal device business to Sharp. Plasma prices had halved from May 2004, and the joint venture with Hitachi was expected to post an operating loss of close to $92.5M (¥10bn) for fiscal 2004. By fiscal 2006 the company had come back within reach of a record net profit of $765.4M (¥89bn).

A president's removal, and the governance vacuum it exposed

In June 2008 Nozoe Kuniaki (野副州旦) became president as Kurokawa's successor. On 25 September 2009 Nozoe resigned, citing illness, and thereafter kept his silence and stayed entirely out of sight. Five months after the resignation, however, on 26 February 2010, he sent the directors and corporate auditors a notice retracting it. His account was that when he met some of the officers before the regular board meeting of 25 September, he was pressed with the argument that a representative director associating with a fund said to be an anti-social force was an extremely high-risk and grave problem, and that his resignation was demanded so that Fujitsu would not be delisted — leaving him no choice but to accept. The immediate ground for the notice was that a hearing investigation by his lawyers had concluded there was no involvement whatsoever by the anti-social forces alleged.

Fujitsu convened an extraordinary board meeting on 6 March 2010 and resolved to remove Nozoe as senior adviser, citing the loss of the relationship of trust caused by the unilateral dispatch of the notice. The same day it issued a long disclosure correcting the stated reason for the resignation from illness. The Tokyo Stock Exchange judged the disclosure of 25 September 2009 to have been inappropriate and issued a verbal severe warning on 9 March 2010. Fujitsu at the time was one of Japan's largest corporate groups, with 183,508 employees and sales of $50.2B (¥4.69tn), and its governance had fallen into a serious state. On 1 April 2010 Yamamoto Masami (山本正已), a corporate senior vice-president, became president and took on the rebuilding.

Cutting the founding hardware loose, and Fujitsu Uvance

In semiconductors, the LSI business was split off in March 2008 to form Fujitsu Microelectronics, renamed Fujitsu Semiconductor in April 2010. In August 2012 part of Fujitsu Semiconductor was sold to a domestic foundry, and in March 2015 Socionext began trading as the merged system LSI business of Fujitsu and Panasonic. The mobile phone business was spun off in March 2014 as Fujitsu Connected Technologies and transferred to Polaris Capital in April 2019, ending the company's involvement. The personal computer business was turned into a joint venture with Lenovo in November 2017, with Lenovo taking the majority of Fujitsu Client Computing. Fujitsu and NEC — once called members of the Denden family for supplying telephone switching equipment to NTT — cut loose one business after another judged non-core over this decade, and turned themselves into vast IT vendors. With the structural reform complete, the semiconductor business was folded back into the parent in April 2023, and the hardware lineage running back to the founding reached its close here.

In June 2015 the presidency passed from Yamamoto Masami to Tanaka Tatsuya (田中達也), and in June 2019 Tokita Takahito (時田隆仁) became the first president to come from systems engineering. Tokita said that solving customers' problems with digital technology required the company itself to change, and set out an escape from contract development that would take in consulting upstream. In April 2020 it established a new company for DX consulting and brought in executives from SAP of Germany and Microsoft of the United States as officers of the parent. The wall in its way was Accenture, the American consulting firm, whose Japanese arm had more than tripled its headcount over the decade to 14,000 people. In July 2020 Fujitsu introduced job-based personnel management in earnest for managers, and renewed the Fujitsu Way, making transformation into a DX company the policy of the firm.

In October 2021 Fujitsu announced Fujitsu Uvance as the framework for a cross-industry offering business. The medium-term management plan of May 2022, running to 2025, set sales of $4.7B (¥700bn) from Fujitsu Uvance in fiscal 2025 as its central target. The reorganisation continued: Fujitsu Japan was created in 2020 to strengthen the domestic business, and in March 2022 more than 3,000 jobs were cut. The pace of change, however, cast a shadow over quality control. From March 2023 onwards, incidents in which certificates belonging to other people were issued by the My Number card certificate-issuing system supplied by Fujitsu Japan occurred one after another at local authorities across the country, and together with the leak of information from a corporate internet service that came to light in 2022, this led the Ministry of Internal Affairs and Communications to issue administrative guidance at the end of June 2023 to Fujitsu and to the subsidiary providing the cloud service. After twenty years of decisions to let things go, the work of answering what to sell with the businesses it kept is where Fujitsu now stands.