On March 6, 2026, a Nikkei report shook the market. Denso, the core parts maker of the Toyota group, had proposed acquiring all shares of Kyoto-based chipmaker Rohm through a tender offer. The price was put at up to roughly 1.3 trillion yen (about $8.2 billion), which would have made it one of the largest consolidations Japan's power semiconductor sector had ever seen.

Rohm's stock hit its daily limit immediately, closing 18 percent higher at 3,243 yen, its sharpest rise in 26 years. Denso, the buyer, dropped more than 4 percent at one point as investors weighed the financing burden and integration risk.

Behind it all sits a sense of crisis about Japan's position in power semiconductors. These are the chips that convert and control electricity: driving EV motors, charging batteries, managing power in data centers, and generally sitting wherever energy needs to be used efficiently. Unlike the logic chips that serve as a smartphone's brain, they are unglamorous and absolutely essential.

From Partnership to Takeover Bid in 18 Months

The relationship escalated in stages. In September 2024 Denso announced a capital tie-up and took 0.3 percent of Rohm. In May 2025 the two published a basic agreement on a strategic partnership. That July, Denso spent about 38 billion yen to raise its stake to just under 5 percent. Then, in February 2026, came the tender offer for the whole company. Partnership, stake, bigger stake, bid: textbook, on paper.

Rohm had posted a net loss of 50 billion yen (about $320 million) for the year ended March 2025. Its early bet on silicon carbide (SiC) power semiconductors had backfired as EV market growth slowed and Chinese rivals attacked on price. Saddled with expensive capacity and weak earnings, Rohm must have looked like an opening.

Rohm's market capitalization sat around 1.1 trillion yen (about $7 billion) before the report, so 1.3 trillion implied a premium of roughly 18 percent. Against the 20 to 50 percent typical of tender offers, that was restrained.

The Stark Reality: Japan Falls Far Behind Europe

Why does consolidation keep coming up? The market share numbers answer it.

According to Omdia, Germany's Infineon Technologies leads power semiconductors with about 22.8 percent. America's onsemi follows at roughly 11.2 percent and Switzerland's STMicroelectronics at about 9.9 percent. Those three Western firms alone hold roughly half the market.

Japan's side of the ledger: Mitsubishi Electric, the domestic leader, has 5.5 percent. Fuji Electric has 4.9 percent. Toshiba and Rohm sit at 3.2 percent each. Add all four together and they still fall short of Infineon alone.

Chinese players are climbing too. Backed by generous state subsidies, BYD Semiconductor and CRRC Times Electric push on price and are gaining ground in industrial and low-to-mid voltage segments. JS Foundry, a Japanese attempt at contract manufacturing for power chips, lost that price war and went bankrupt in July 2025.

SiC: The Material Rewriting the EV Race

The material everyone watches now is silicon carbide. Compared with conventional silicon it loses far less power and stays stable at high temperatures. Put SiC in an EV inverter, the unit that controls the motor, and range reportedly improves by 5 to 10 percent, which is why Tesla, BYD, and others have adopted it.

The market is growing accordingly. SiC power semiconductors are projected at roughly $2.73 billion in 2025, reaching about $8.4 billion by 2030, an annual growth rate near 25 percent. The broader power semiconductor market should climb from about $55 billion in 2025 to somewhere between $67 billion and $80 billion by 2030.

But SiC is an oligopoly. The global top five, Infineon, STMicro, Wolfspeed of the US, onsemi, and Rohm, account for roughly 90 percent of world revenue. Rohm is the only Japanese company still in that league, which is precisely why Denso wanted it. Meanwhile Wolfspeed has fallen into financial distress, and Renesas Electronics abandoned its SiC mass production plan after its collaboration with Wolfspeed collapsed. The competitive landscape is in flux.

Wafer diameter is another axis. The shift from 150mm (6-inch) to 200mm (8-inch) is underway, and 8-inch yields roughly 2.2 times as many chips per wafer, cutting costs by as much as 40 percent. Infineon has brought a 200mm SiC megafab online in Malaysia and is pushing for cost leadership. Mitsubishi Electric started an 8-inch capable plant in Kumamoto in November 2025.

The Bid Was Withdrawn: Consolidation Shifts to Rohm, Toshiba, and Mitsubishi Electric

The short version: Denso's proposal did not land.

Rohm set up a special committee of outside directors to assess the impact on corporate value. On March 13 came reports that Rohm and Toshiba had entered talks on integrating their power semiconductor businesses; Rohm said it was in discussions on strengthening the business alliance. On March 24, Denso formally confirmed it had made a share acquisition proposal.

Then on April 28, 2026, Denso announced it was withdrawing the bid. It had not secured Rohm's agreement and concluded the deal would not enhance its own corporate value. Rohm's special committee said it had not reached a conclusion supporting the proposal. Rohm announced the same day that its review was closed.

The two did not part ways, though. On production and development centered on analog semiconductors, they agreed to advance their co-creation activities and keep talking, including personnel exchange. Denso still holds just under 5 percent of Rohm. On the same day, Denso announced a buyback tender for its own shares at 1,696 yen apiece, worth up to about 313.6 billion yen.

The center of gravity has moved. Nikkei reports that power semiconductor consolidation now looks likely to proceed around an alliance of Rohm, Toshiba, and Mitsubishi Electric. Mitsubishi Electric President Kei Uruma has argued for years that Japan cannot grow global share while domestic makers stay fragmented, and that the sooner consolidation happens the better. His company is strong in power semiconductor modules, which bundle multiple devices, and weak in single-function discretes; consolidation would fill the gap. Even Uruma conceded in November 2025 that the market is poor, that merging would leave excess capacity, and that any agreement requires both sides to swallow something.

The Ministry of Economy, Trade and Industry has been orchestrating too. Subsidies worth 129.4 billion yen underpinned a Rohm-Toshiba camp set against a Denso-Fuji Electric camp, but Rohm objected when Toshiba announced a tie-up with the Chinese SiC wafer giant SICC, and the camp cracked. Denso's proposal emerged from that confusion, and then retreated into it.

Not Rapidus: Japan's Other Semiconductor Strategy

Talk of Japan's semiconductor revival tends to fixate on Rapidus and its push for 2nm logic production in Chitose, Hokkaido. Power semiconductors are a different world entirely.

Rapidus is chasing the ultra-fine brain chips that go into smartphones and AI servers, where the game is shrinking circuit width to the limit. Power semiconductors care less about miniaturization and more about controlling large currents efficiently, and the materials (SiC, GaN) and processes are fundamentally different.

Japan is fighting on two fronts. Rapidus is the attempt to claw back ground in advanced logic; in power semiconductors, the play is consolidation among existing makers. Neither alone restores a semiconductor power.

Denso's proposal misfired, but the industry's direction has not changed. Japanese power semiconductor makers cannot go on competing globally one company at a time, and consolidation of some form still looks inevitable. Only the cast at center stage has changed.

Is power semiconductor consolidation underway in your country as EVs spread? Are automakers there buying chip companies? We'd love to hear about it.

References