Company history

Renesas begins not with a founder but with a subtraction. In November 2002 NEC split off its semiconductor business — everything except commodity DRAM — into a wholly owned company in Kawasaki, NEC Electronics, and listed it on the Tokyo Stock Exchange in July 2003. The timing was industry-wide: with the semiconductor market slumping in the early 2000s, Japan’s electronics conglomerates could no longer carry chip divisions on their own balance sheets, and the carve-outs that followed — Elpida Memory the same year — were less an act of industrial policy than an act of housekeeping. The new company was pushed out to be judged by the capital markets without ever having been designed as a standalone business.

What it inherited was fixed cost. NEC Electronics spent its first years cutting: the back-end operation of Yamagata NEC went to Taiwan’s ASE in 2004, an Irish assembly line closed in 2006, domestic manufacturing subsidiaries were reshuffled again and again. It was not enough. Net losses ran ¥98.1 billion in FY05, ¥41.5 billion in FY06 and ¥15.9 billion in FY07 — near-continuous red ink from the moment of listing. While foreign specialists were moving fab-lite, separating design from manufacture to compete unencumbered, Renesas’s predecessor carried its own plants and a scattered network of sites it could not close quickly. The structural lag it is still remembered for was built in at birth.

The company was not only retreating. Tosaka Kaoru, who became president in November 2002, staked the firm on vertical integration — owning everything from process technology to customer support — and on system LSI, where anything short of the world’s top share failed to earn back development costs that started at ¥1 billion per design. The showpiece was the μPD61171 for DVD recorders, which paid off a bet placed back in 1997, when compression encoders for the home were widely dismissed as a market that would never exist. Renesas also put ¥60 billion into 300mm capacity — the same instinct to build in-house that would later become the thing it had to unlearn.

In April 2010 NEC Electronics merged with Renesas Technology — itself the chip arms of Hitachi and Mitsubishi Electric — to form Renesas Electronics. Nothing like it had been attempted in Japan: three of the country’s largest electronics groups folding their semiconductor divisions into one company. It bought the world’s leading share in automotive microcontrollers. It also bought more than twenty domestic manufacturing subsidiaries and design sites, overlapping sales organisations serving the same carmakers, and no ability to decide quickly which of them to close. Scale and fixed cost arrived in the same box, and merely unifying three sets of engineering and sales lines consumed the management bandwidth of the first years.

Eleven months later the Great East Japan Earthquake stopped the Naka plant in Ibaraki — the main source of automotive microcontrollers — for months. Net losses ran ¥115.0 billion in FY10, ¥62.6 billion in FY11 and ¥167.6 billion in FY12: more than ¥340 billion in three years, with the wireless modem business acquired from Nokia adding to the drag. Carmakers did not wait. Toyota and others began qualifying alternative suppliers through their tier-one vendors worldwide, and single-source dependence on automotive silicon became a permanent industry concern rather than a Renesas problem.

The rescue, in September 2013, was neither purely public nor purely private. A buyout approach from KKR alarmed Toyota and Nissan, who lobbied the Ministry of Economy, Trade and Industry; what emerged instead was a $1.5B (¥150bn) third-party share issue led by the state-backed Innovation Network Corporation of Japan, with eight automotive and electronics customers — Toyota, Nissan, Keihin, Denso, Canon, Nikon, Panasonic and Yaskawa — riding alongside. Because the investors were customers who could not afford to lose supply, stability outranked short-term profit. Three back-end plants went to J-Devices, the European and Indian mobile subsidiaries to Broadcom, and by FY13 Renesas posted ¥67.6 billion of operating profit — its first real profit since the merger, earned by shrinking to the automotive core.

Under Endo Takao (2014) and then Kure Bunsei (2016), the rescue plan worked and stopped working at the same time. Operating profit held above ¥100 billion in FY14 and FY15, but revenue kept contracting — ¥471.0 billion in FY16 — because the profit came from cuts rather than growth. A portfolio anchored to automotive microcontrollers could not reach smartphones, data centres or IoT, and even within cars, the rise of NVIDIA and Mobileye in ADAS processors put a microcontroller company on the defensive. Stability had been bought with the future.

The turn came in February 2017 with the acquisition of Intersil of the United States, a mid-sized analogue and power player — the first move off the automotive monoculture. Revenue recovered to ¥780.2 billion in FY17 at a 10.1% operating margin, but the open question was integration: few Japanese electronics-descended chipmakers had ever absorbed an independent American analogue house, and leaving most design leadership with the former Intersil team was itself an answer to that doubt. Kure followed in September 2018 with IDT for more than ¥700 billion, adding timing devices for data centres plus communications, sensing and memory interfaces. He argued that autonomous driving raised the value of microcontrollers and SoCs, claimed a two-to-three-year process lead over NXP and Infineon, and deliberately declined to fight Intel and NVIDIA on deep learning — Renesas would keep the safety-critical, volume-production automotive ground instead. The deal closed in March 2019.

That July, Shibata Hidetoshi — an INCJ alumnus who had joined as CFO in 2013 — became CEO and named the problem plainly: the company had to stop shrinking. Growth by acquisition, not profit by restructuring, would be the line. His timing looked terrible. The market turned, FY19 operating profit fell to ¥6.2 billion against a ¥6.3 billion net loss, and a balance sheet carrying two large acquisitions met a downturn; rating outlooks went negative and credit spreads widened. Shibata refused to slow investment and spent the year persuading shareholders that the route had not changed.

The third deal closed in August 2021: Dialog Semiconductor of the UK, a low-power mixed-signal specialist known for supplying power-management ICs to Apple, for about ¥600 billion. It added consumer and wearable power efficiency to a portfolio weighted toward cars and industry — and, as a by-product, spread a customer base that had been concentrated on a few automakers, while introducing a new concentration risk in Apple. Then the pandemic chip shortage arrived. Revenue reached ¥993.9 billion with ¥183.6 billion of operating profit in FY21 and ¥1,500.8 billion with ¥424.2 billion in FY22 — the best results since the merger, and effectively a different company from the one the state had rescued. Shibata later treated the boom warily, citing excess channel inventory around 2020 as the mistake to design against.

Renesas reorganised into two pillars, automotive and industrial/infrastructure/IoT, and by FY22 the second was the larger: ¥845.8 billion of revenue against ¥645.0 billion. Smaller purchases filled in the gaps — Celeno (Wi-Fi, 2021), Reality Analytics and Steradian (radar, 2022), Panthronics (NFC, 2023), Transphorm (GaN power, 2024) — with a single aim: to own enough IP to sell a complete subsystem, from sensor through power to connectivity, rather than a chip. It was a shift in how the company sells, following the solution model European and American rivals had already adopted, several years late.

The most unusual move came in August 2024, when Renesas paid roughly $5.9B (¥900bn) for Altium, the Australian-listed PCB design software maker — a hardware company buying its way upstream into electronic design automation. Shibata called the combined offering “Renesas 365” and described a company becoming a platform business, embedding design tools and silicon together in the customer’s workflow and charging by subscription rather than by shipment, with a target of a sixfold market capitalisation by 2030. A year in, Altium’s annual recurring revenue was growing about 15%. Then 2025 delivered the counter-lesson: Wolfspeed, Renesas’s contracted external source of silicon carbide power devices, filed for Chapter 11. Renesas converted its deposit into convertible notes and equity, booked ¥237.6 billion of impairment and valuation losses, and reported a ¥51.8 billion GAAP net loss even as non-GAAP operating profit hit a record ¥386.9 billion. It suspended its own SiC development for good — Shibata stated the company does not intend to build the technology itself — and settled for sourcing a strategic product from outside, an admission that would have been unthinkable in the in-house cultures of Hitachi, Mitsubishi and NEC it came from.