Renesas Electronics - Company History

A semiconductor business cut away from NEC

In November 2002 NEC used a corporate split to carve out its semiconductor business, excluding commodity DRAM, and established NEC Electronics in Kawasaki as a wholly owned subsidiary[1]. In July the following year the new company listed on the First Section of the Tokyo Stock Exchange[2], lowering the parent's stake and gaining capital independence in the same movement. This coincided with a period in which the worldwide slump in the semiconductor market of the early 2000s left the electrical makers unable to keep their chip divisions in-house, and restructurings that put the unit out to be valued independently by the capital markets were proceeding in several countries. Rather than a national policy of protecting semiconductors, it had the character of each major electrical maker cutting weight from its own balance sheet, and alongside the launch of Elpida Memory the same year it became a symbol of the Japanese electronics industry's semiconductor tidy-up.

NEC Electronics moved quickly to reorganise its domestic manufacturing subsidiaries and outsource back-end work, transferring the back-end operations of Yamagata NEC to ASE of Taiwan in 2004[3] and closing an Irish back-end line in 2006[4], compressing fixed costs again and again. Even so the net loss came to $890.6M (¥98bn) in FY05, $356.9M (¥42bn) in FY06 and $135M (¥16bn) in FY07: for the first several years after listing it was in the red almost every period, and doubt hung over whether it could sustain itself alone. It was dragging along, unchanged, the structural problem carried by every chip company owned by an electrical major — high fixed costs and a scattered set of sites that could not be broken down in a short time, in design and in manufacturing alike. Set against the overseas specialists, which had moved to a fab-lite model separating design from manufacturing and were competing unencumbered, the delay in structural change was plain.

Alongside the financial rebuild, NEC Electronics also pursued an attacking strategy that aimed at world share in system LSI. Tosaka Kaoru (戸坂馨), who completed graduate school at the University of Tokyo in 1966, joined NEC and served as senior executive vice-president and president of the NEC Electron Devices Company before becoming president in November 2002, set as his axis of differentiation the idea of a vertically integrated company holding everything from production technology to support in-house, offering each customer the semiconductor solution best suited to it. At the core was the μPD61171 system LSI for DVD recorders. In a system-LSI business where development costs run to the order of $8.6M (¥1bn) even when they are cheap, the structure was such that anything short of first place in the world did not pay, and the foresight of starting early, in 1997, on a data-compression encoder that was then seen as something that could never spread into the home was what led to world share in later years. Some $544.7M (¥60bn) was put into 300mm production equipment. Tosaka was followed by Nakajima Toshio (中島俊雄) in November 2005 and Yamaguchi Junshi (山口純史) in June 2009, who inherited the task of escaping the loss-making constitution that had persisted for years after listing.

The bet on merging the Hitachi, Mitsubishi and NEC chip units

In April 2010 NEC Electronics merged with Renesas Technology, whose origins lay in the semiconductor divisions of Hitachi and Mitsubishi Electric, and Renesas Electronics was born[5]. In the management line-up Akao Yasushi (赤尾泰), president of the former Renesas Technology, took the presidency of the new company, while Yamaguchi Junshi, president of the former NEC Electronics, stepped back to become chairman with representative authority[6]. It was an integration without precedent, bringing the semiconductor divisions of three of Japan's largest electrical makers into one, and it handed the company the position of world number one in automotive microcontrollers. At the same time it took on more than twenty manufacturing subsidiaries and development sites in Japan, and set off without being able to decide to close the overlapping ones in short order, producing an organisation in which the scale merits of integration and the weight of fixed costs lived side by side. The work of unifying the sales lines and design sites of three camps that shared automotive customers consumed, on its own, much of the management bandwidth of the early integration years.

In March 2011, immediately after the merger, the Great East Japan Earthquake struck, and the Naka plant in Ibaraki — the main site for automotive microcontrollers — was damaged and halted for several months[7]. The net loss was $1.3B (¥115bn) in FY10, $784.7M (¥63bn) in FY11 and $2.1B (¥168bn) in FY12, three consecutive years in the red, and the cumulative net loss passed $4.3B (¥340bn). The wireless modem business taken over from Nokia also underperformed[8], and the company fell into a state where, far from integration synergies, it could not even see how the crisis would be contained; carmakers moved as one to look for alternative suppliers of automotive microcontrollers. Unease spread through the whole Japanese semiconductor supply chain, and Toyota accelerated moves to qualify equivalent parts from other vendors through its Tier 1 suppliers around the world. The risk of depending on a single source for automotive semiconductors became common knowledge across the industry.

Effective nationalisation by the INCJ, and a shrinking equilibrium

In September 2013 a third-party share allotment was carried out with the Innovation Network Corporation of Japan (INCJ) at its centre and Toyota, Nissan, Keihin, Denso, Canon, Nikon, Panasonic and Yaskawa Electric as allottees[9]. The INCJ became the largest shareholder, with the major carmakers and electrical makers that wanted to avoid any uncertainty in the supply of automotive semiconductors riding along, and the effect was a rescue with public money[10]. From then on the phrase nationalised Renesas took hold in the media, and a distinctly Japanese form of restructuring — government and an automotive alliance defending the supply network together — was cut into the semiconductor industry. The management team entrusted with the rebuild came with a restoration plan of reorganising sites and concentrating on automotive microcontrollers, and secured from the customer carmakers an agreement that stability of supply took priority over short-term improvement in profit.

The first move in the rebuild was restructuring and the disposal of businesses. In 2013 the three back-end plants in northern Japan, Kansai and Kyushu were transferred to J-Devices[11] and the mobile subsidiaries in Europe and India were sold to Broadcom[12]; the following year the front-end and back-end manufacturing subsidiaries were consolidated into two companies[13]. Operating profit for FY13 came to $692.6M (¥68bn), the first substantial profit since the merger, and the shrinking-equilibrium model of concentrating management resources on automotive microcontrollers showed results for the time being. The management line-up moved with the rebuild as well: Tsurumaru Tetsuya (鶴丸哲哉) took over from Akao as president in February 2013, and in June of the same year Sakuta Hisao (作田久男), who had come up at Omron, was brought in from outside as representative director, chairman and chief executive. Revenue kept shrinking, and the shape of earning fully in the areas that paid at the cost of letting go of growth opportunities was left as the problem for the next regime. The risk of being outrun by Western rivals in the higher-margin adjacent fields of analogue and mixed-signal semiconductors was already, at this point, a subject of debate inside the board.

A first attempt to get off one automotive leg

In June 2015 Endo Takao (遠藤隆雄), who had come from Oracle Japan, took over from Sakuta as representative director, chairman and CEO, and the two-headed structure continued alongside Tsurumaru, president since February 2013 and president and COO from June of that year[14]. Endo, however, stepped down after six months; Tsurumaru took the CEO title as well from December 2015, and Kure Bunsei (呉文精) succeeded him as president and CEO in June 2016[15]. Renesas in this period, standing as world number one in automotive microcontrollers, posted operating profit of $986.4M (¥104bn) in FY14 and $857.7M (¥104bn) in FY15 and recovered a profitable constitution. Yet FY16 revenue shrank to $4.3B (¥471bn), and the company remained caught in a structure of squeezing profit out through restructuring — of never stopping shrinking. With a single portfolio dependent on the automotive sector, it could not take in the fruits of growth areas such as smartphones, data centres and the internet of things, and it went on sacrificing its future in exchange for stability. Even within the automotive field, against NVIDIA and Mobileye rising with SoCs for advanced driver assistance, Renesas's microcontroller-centred position was on the defensive.

The turning point was the acquisition of Intersil of the United States in February 2017[16]. It took in a mid-sized player in analogue and power semiconductors, and was the first deal aimed at moving from one automotive-microcontroller leg into the analogue field. FY17 revenue recovered to $7.0B (¥780bn) and the operating margin improved to 10.1 per cent, but the market's attention settled on whether the post-merger integration would genuinely work. There had been few examples of a chip company owned by an electrical major fully integrating an independent analogue maker from abroad, and the whole industry watched to see whether Japanese-style management could hold together a Silicon Valley development culture. The decision to leave most of the design leadership with the former Intersil side was an answer of sorts to that concern.

As his second arrow, Kure announced in September 2018 that the company would buy IDT (Integrated Device Technology) of the United States for more than $6.3B (¥700bn). The aim was to take in the timing devices that set the priority of processing for data centres, and to add the growth areas of communications, sensors and memory interfaces to the analogue and mixed-signal business. Kure argued that microcontrollers and SoCs were becoming more important to realising autonomous driving, and said Renesas was two to three years ahead of NXP Semiconductors of the Netherlands and Infineon Technologies of Germany in miniaturisation, while taking a position of avoiding a head-on collision with Intel and NVIDIA of the United States, which were pressing into the automotive field. He made clear that he would not step into deep learning, the main battlefield of those two, and would not let go of the volume-car domain that bears directly on human safety. The IDT acquisition completed in March 2019 and was carried across the handover of the presidency to Shibata Hidetoshi (柴田英利) in July of the same year[17].

Shibata Hidetoshi takes over and declares an end to never-ending shrinkage

In July 2019 Shibata Hidetoshi, who had come from the INCJ and served as chief financial officer, became chief executive[18]. Shibata, who worked at JR Central before building hands-on M&A experience at an investment bank and at the INCJ and joined Renesas as CFO in 2013, argued repeatedly for the need to break out of a state of never-ending shrinkage, and announced a shift from securing profit through restructuring to a line of expansion through acquisitions and investment in growth. It was a change of direction under a chief executive who had come up through finance, and a declaration that management was changing gear from the stage of tidying up the negative legacy to the stage of going after the future. At the time of his appointment the integrations of Intersil and IDT were still being assessed, and Shibata took the decision to move on the next acquisition before seeing how they turned out, leaving objections alive inside the company. The board ultimately backed the acceleration, and the change of course was shared across the organisation.

A deterioration in the semiconductor market left FY19 operating profit at $56.9M (¥6bn) and a net loss of $57.8M (¥6bn), and Shibata was in difficulty almost as soon as he took office. Meeting a downturn in results while still carrying the debt from firing off the Intersil and IDT acquisitions back to back, the market saw for itself how heavily Shibata's line of continued attacking investment burdened financial strength in a recession. The rating agencies moved their outlook to negative more than once, and bond spreads widened against listed peers as the financial markets grew sterner. Shibata showed repeatedly at results briefings that he would not stop investing for growth, and worked patiently through dialogue with shareholders to persuade them to keep the course.

The Dialog acquisition, with the pandemic chip shortage as a tailwind

In August 2021 the company acquired Dialog Semiconductor of the United Kingdom[19], a specialist in low-power mixed-signal semiconductors known for supplying power management integrated circuits to Apple, in a deal of about $5.5B (¥600bn). It was the third acquisition after Intersil and IDT, and it thickened by another layer the analogue and power domain at Renesas. A new field — high-efficiency power for consumer devices and wearables — was added to a portfolio that had been weighted towards automotive and industrial uses, and the diversification of the customer portfolio came with it as a by-product. It marked the turn from a revenue structure concentrated on particular carmaker OEMs to a multi-layered customer base including smartphones and household appliances, and in theory raised the company's resistance to the economic cycle. Because it carried within it another concentration risk — dependence on Apple — managing the share of any single customer emerged as a governance question for the period after integration.

As the chip shortage grew severe worldwide under the pandemic, Renesas recorded revenue of $9.1B (¥994bn) and operating profit of $1.6B (¥174bn) in FY21, and revenue of $11.4B (¥1.5tn) and operating profit of $3.2B (¥424bn) in FY22, the best results since the merger. Strong demand from the automotive and industrial sectors coincided with a weak yen, and the company moved to an earnings model that looked like a different business from the one rebuilt under the INCJ. Shibata himself later remarked that about five years ago there was a stretch where things were not very good[20], citing the excess channel inventory of around 2020 as a point of reflection. He had been conscious from an early stage of the risk that inventory swells when a sharp rise in demand reverses, and the inventory policy of later years was rebuilt on the premise of that experience. As a result the inventory position of the consumer power management chips taken in through Dialog also became material that sharpened the market's interest in the company's forecasts.