Japan's power semiconductor industry has moved dramatically in just three weeks. Starting with the March 27 three-way merger announcement between Rohm, Toshiba, and Mitsubishi Electric, we've seen Denso declare itself a "semiconductor maker," Rohm's CEO insist on leadership, Toshiba eye a comeback IPO, and Mitsubishi Electric quietly open a major new Fukuoka plant. Four companies, four agendas, and a summer deadline looming. Here's what actually happened, in order.

Related Articles This is a follow-up. For background, see our earlier coverage:

The Story So Far: The Earthquake of March 6–27

On March 6, news broke that Denso, Toyota Group's auto parts giant, had proposed acquiring all of Rohm's shares via tender offer, at roughly ¥1.3 trillion ($8.2 billion). Rohm's stock hit the daily limit, jumping 18%. The market was stunned.

This external pressure accelerated quiet talks already underway between Rohm and Toshiba. On March 27, Mitsubishi Electric joined the conversation, and the three companies officially announced merger talks for their power semiconductor businesses. If realized, the combined entity would hold roughly 11% global market share, surpassing U.S. rival onsemi to become the world's No. 2 player, behind only Germany's Infineon.

That covered the seismic 21 days of March 6–27. But the story didn't stop there.

Denso: March 31, a "Semiconductor Maker" Declaration

Just four days after the three-way merger announcement, Denso held its investor day "DENSO DIALOG DAY 2026" in Tokyo on March 31, unveiling its new mid-term plan "CORE 2030."

The scale is staggering. Targets for fiscal 2030: sales exceeding ¥8 trillion ($50 billion), operating margin above 10%, and ROE above 11%. To get there, Denso will invest ¥3.7 trillion ($23 billion) in R&D, ¥2.2 trillion ($14 billion) in capital expenditure, and ¥700 billion ($4.4 billion) in IT, IP, and human resources over five years, a total of ¥6.6 trillion ($41 billion). This investment scale is unprecedented for a Japanese auto parts maker.

The most striking moment came when CEO Shinnosuke Hayashi declared, "Semiconductors drive vehicle value." Denso now plans to expand its semiconductor business across three domains (automotive, industrial equipment, and consumer electronics), effectively announcing its transformation from auto parts company into semiconductor maker.

On the Rohm bid, Hayashi reaffirmed that Denso is "not excluding possibilities broadly," signaling the proposal remains active. He also highlighted Denso's own 3D-structure SiC power semiconductors, which deliver 70% lower losses and 30% smaller size than conventional chips in new BEV inverters.

Notably, Hayashi made clear Denso has no plans to exit internal combustion engines. "The acceleration of electrification that people talked about 3–5 years ago has actually slowed, that's the correct view," he said, committing to a multi-pathway strategy covering ICE, HEV, PHEV, BEV, and FCEV. The message: be realistic about EV slowdown, but aggressive in semiconductors.

Rohm: "Too Many Cooks Spoil the Broth," CEO Defends Leadership

Three days after Denso's plan drop, on April 3, the Nikkei reported comments from Rohm CEO Katsumi Higashi. On the three-way alliance, he said, "We want to be at the center ourselves" and "Too many cooks spoil the broth."

This was a clear shift from the neutral stance Rohm maintained on March 27. While the formal position remains that the special committee is evaluating the Denso offer, management's lean toward the three-way alliance is now visible.

The context matters. Rohm posted a ¥50 billion ($310 million) net loss in fiscal 2025, its first loss in 12 years, driven by premature SiC capacity investments that ran into the EV slowdown. Joining Denso would refocus Rohm on automotive, but limit access to industrial and consumer markets where Rohm is strong. The three-way alliance preserves independence but adds coordination overhead.

Rohm also contributed ¥300 billion ($1.9 billion) to the JIP-led consortium that took Toshiba private in 2023, creating a capital linkage that makes three-way integration easier, and a Denso buyout more complicated. CEO Higashi's "leadership" stance reflects not just Kyoto pride, but cold calculation about capital structure.

Toshiba: "Half a Step Toward Re-Listing," An Unexpected Comeback Story

On April 9, the Nikkei analyzed the three-way merger through an unusual lens: Toshiba's path back to public markets.

Toshiba was taken private in late 2023 by a JIP-led consortium in a roughly ¥2 trillion deal. Since delisting, the company has been restructuring, with its semiconductor business (Toshiba Device & Storage) positioned as a growth engine for an eventual re-IPO.

According to sources, Toshiba internally concluded early on that "a three-way alliance is the optimal solution against Infineon," and had been quietly negotiating with Rohm since 2024 and Mitsubishi Electric since 2025. Denso's Rohm bid provided the external shock that forced these talks into the open.

For Toshiba, the three-way merger becomes a vehicle to spin off its semiconductor business and re-list, an exit strategy for JIP as well. Unlike past failures like Japan Display Inc. (JDI), which combined weak divisions under government pressure, this deal carries the flavor of a deliberate, opportunistic re-listing strategy.

Mitsubishi Electric: April 15, Opening a New Fukuoka Plant Mid-Negotiation

On April 15, while merger talks continued, Mitsubishi Electric completed a major capital project in Fukuoka.

At its Power Device Works Fukuoka site (Nishi Ward, Fukuoka City), the company held the completion ceremony for a new ¥10 billion ($63 million) building called "Power Device A-tou (PA-tou)." The five-story, 25,000-square-meter facility will handle back-end processing (assembly and inspection) of power semiconductors. With automated material handling robots, productivity will improve 40% versus existing lines. Full operations begin October 2026.

This timing is not accidental. The plan was originally announced in March 2023, but completing it during active merger talks is strategically significant. Mitsubishi Electric spent years advocating for industry consolidation without being able to execute it concretely. Having slipped into the three-way framework thanks to the Denso shock, completing its own capex signals "we've already made our move", strengthening its leverage in negotiations over integration terms and role distribution.

The Summer Focus: Three Open Questions

What's become clear in three weeks is that all four companies are now taking defined strategic positions. On March 27, the story was "let's begin talks." Now, each player's interests and priorities are openly colliding.

The three companies aim to "establish a direction by this summer." Three key questions remain.

First, Rohm's special committee decision. Whether Denso's bid or the three-way alliance better serves shareholder value is the central fork in the road. The committee, composed of independent directors, is structured to decide independently of management, but CEO Higashi's comments suggest management is leaning three-way.

Second, ownership structure and integration form. Will it be a joint venture, or a Rohm-centered merger? The three companies have roughly similar revenues: Rohm (¥480 billion), Mitsubishi Electric's device business (¥290 billion), and Toshiba's semiconductor business (¥445 billion), making leadership a contested question.

Third, the race against global competition. Germany's Infineon is scaling its 200mm SiC megafab in Malaysia. Onsemi is expanding in Czechia and the U.S. STMicro is building in Italy and via a China JV. BYD and CRRC Times have started in-house SiC production. Every month Japan spends on integration talks is a month competitors move ahead.

Is the "Japan Power Semiconductor Revival" Real?

Past "Japan alliances" like JDI and Elpida Memory failed. Can this be different?

The optimistic case rests on genuine technical complementarity. Rohm's SiC, Toshiba's silicon and 300mm wafers, Mitsubishi Electric's module technology, each brings a distinct strength, not overlapping weakness.

The pessimistic case is timing. The more complex the integration, the longer decisions take, and the further behind global rivals Japan falls. And it's still unclear whether Rohm will truly reject Denso, or whether some compromise (e.g., partial Denso stake plus three-way alliance) will emerge.

The shift visible in these three weeks is that Japanese companies have moved from "wait and see" to "take action." But action isn't the goal. What design emerges by summer, and whether it can actually compete globally, is what ultimately matters.

Update: After the three-week window covered here, Denso withdrew its Rohm acquisition proposal in late April 2026. The power semiconductor reshuffle is now proceeding around the three-way integration of Rohm, Toshiba, and Mitsubishi Electric.

Are there examples in your country of auto parts makers trying to acquire semiconductor companies, or of industry-wide consolidation plays? How have "national champion" mergers worked out where you live? We'd love to hear your perspective.

References