⚡ Just three days after Denso's withdrawal. On April 28, Mitsubishi Electric CEO Kei Urutsuma dropped a bombshell at the company's earnings briefing. "I want to spin off the power semiconductor businesses of all three companies and form a joint venture." That's a different framework than the "full semiconductor integration" Rohm and Toshiba had outlined. A new wrinkle has been added to Japan's vision of a global No.2 power-semi alliance. Wolfspeed, Infineon, STMicro — global rivals aren't waiting. Can Japan's "national JV" actually reach agreement by summer?

📌 Related Articles This is the latest chapter in a four-part series. For background, see:

What Urutsuma's "Spinoff JV" Statement Actually Means

On April 28, 2026, at Mitsubishi Electric's fiscal year 2026 earnings briefing in Tokyo, CEO Kei Urutsuma responded to a question about the three-way integration framework with this:

"I want to spin off the power semiconductor businesses of all three companies and establish a joint venture."

He continued: "I think the integration of power semiconductors and the integration of other businesses must inevitably proceed separately. We want to handle this in parallel and reach agreement quickly."

This isn't just executive enthusiasm — it's a substantive proposal about the architecture of the integration. Here's why it matters. The basic agreement Rohm and Toshiba announced on March 27 explicitly mentioned "integration of both companies' semiconductor businesses including not only power semiconductors but also analog, logic, and optical devices." In other words, Rohm and Toshiba were planning a "full semiconductor business integration."

Now Urutsuma is saying: "Let's just spin off the power semi parts and create a JV." That's a different structure entirely.

Why Mitsubishi Electric Insists on "Power Semi Only"

The reason is straightforward. Mitsubishi Electric isn't a "full-line semiconductor maker" like Rohm or Toshiba.

Urutsuma himself put it bluntly: "From our standpoint, there's no meaning in us joining the integration of non-power-semiconductor businesses. The integration must inevitably be centered on power semiconductors."

Mitsubishi Electric's semiconductor business is built around IGBT (insulated-gate bipolar transistor) modules and high-voltage power devices. The company holds top global share in segments like HVDC (high-voltage direct-current transmission) modules and certain industrial-equipment categories. But it doesn't compete in analog ICs, logic chips, or optical devices — areas where Rohm and Toshiba are strong.

Conversely, Rohm and Toshiba have broad semiconductor portfolios beyond power semis. From the start, both companies' executives have emphasized synergies like "combining analog and power semis enables integrated solutions for automotive and industrial customers." Imagine providing "power management ICs + power semis + sensors" as a single offering for an EV powertrain.

Urutsuma's proposal doesn't reject this synergy outright. His position is that "Rohm + Toshiba's full integration" and "the three-way power-semi JV" should proceed in parallel. But operationally, that creates significant complexity.

On Denso's Withdrawal

The same day, April 28, Denso officially announced the withdrawal of its acquisition proposal for Rohm. The 50-day "Black Ships" saga had ended.

On this, Urutsuma was deliberately measured: "It's another company's decision and not for me to comment on. Either way, we agreed to discuss three-way integration, so I think it's important that we steadily move that forward."

But behind the surface neutrality, Denso's withdrawal is a major tailwind for the three-way talks. Denso's bid had functioned as the "external pressure" that pulled Rohm back into the three-way framework. With that pressure gone, Rohm, Toshiba, and Mitsubishi Electric can focus on internal coordination.

And the first issue in that internal coordination is precisely the structural debate Urutsuma surfaced: "full semiconductor integration vs. power-semi-only JV."

Mitsubishi's Cards: New Fukuoka Plant and Kumamoto SiC Fab

Behind Mitsubishi Electric's confident posture is impeccable timing on its own capital investments.

On April 15, the company held the completion ceremony for its new "Power Device A Building (PA Building)" at the Fukuoka district of its Power Device Works in Fukuoka City. Costing about ¥10 billion ($63 million), the five-story 25,000-square-meter back-end (assembly and inspection) facility will improve production efficiency by 40% versus existing operations. Full operations begin in October.

In November 2025, Mitsubishi Electric also began operating a new 8-inch wafer SiC (silicon carbide) power semiconductor fab in Kikuchi, Kumamoto Prefecture (Shisui district). SiC production capacity is set to expand approximately five-fold in fiscal 2026, supported by a partnership with U.S.-based Coherent for 8-inch substrate supply.

Masayoshi Takemi, head of Mitsubishi Electric's Semiconductor & Device Group, summarized the company's market view at the Fukuoka ceremony: "Around 2023–2024 we expected the EV space to grow. But various Chinese SiC makers have emerged, leveraging subsidies to manufacture at low cost, and we now don't expect EV-related business to deliver the profits originally projected. Mitsubishi Electric's mission is to focus on infrastructure-grade products where our high-voltage and high-reliability strengths are essential — products that lifelines depend on. We also have high expectations for the new 800V power architecture rapidly growing in data centers."

This is a deliberate strategic positioning: Mitsubishi Electric is not engaging in head-on price competition with Chinese makers in EV automotive segments. Instead, it's playing in "high-voltage, high-reliability" niches like data center 800V power architectures and HVDC grid infrastructure.

If three-way integration succeeds, Mitsubishi's "infrastructure and industrial" strength complements Rohm's "automotive, consumer, and industrial" portfolio and Toshiba's "automotive, consumer, and industrial silicon" lineup. But whether that complementarity is housed in an independent JV or absorbed into a Rohm-Toshiba full-company merger fundamentally changes operational autonomy.

Where Global Rivals Stand — The Time Race

While the three companies aim for "agreement by summer" and "full integration in 2-3 years," global rivals aren't standing still.

Wolfspeed (US): The world's largest SiC wafer supplier with 33.7% share. Filed for Chapter 11 in June 2025 and emerged from bankruptcy on September 29 of the same year, having reduced total debt by about 70% and annual cash interest expense by approximately 60%. Re-launched as a "vertically integrated U.S.-made silicon carbide supplier" centered on its 200mm Mohawk Valley fab in New York. Struggling with EV demand slowdown and Chinese pricing pressure, but riding tailwinds from the Trump administration's anti-China semiconductor policies and the broader "U.S. reshoring" trend.

Infineon (Germany): Dominant global leader in power semiconductors with 22–25% share. Bringing its 200mm SiC megafab in Kulim, Malaysia, online and pushing scale and cost leadership. The biggest competitor for the three-way alliance.

STMicroelectronics (France-Italy): According to Trendforce, ST held 32.6% of SiC power devices globally in 2023, the top position. Building a new 200mm SiC fab in Catania, Italy, plus a Chinese JV with Sanan Optoelectronics for SiC production — a two-front strategy. Has a long-term Tesla supply contract, making it strong in automotive SiC.

onsemi (US): Started 8-inch SiC volume production in Roznov, Czech Republic, pushing into European markets. Affected by EV slowdown but pivoting to data center and AI server demand.

The three-way alliance's combined share is around 11%, second to Infineon globally and roughly tied with onsemi for No. 2. But in pure SiC alone, the alliance trails STMicroelectronics significantly. Industry analysts have noted that the "world No. 2" framing is somewhat optimistic when read against a simple sum of shares.

Chinese Pursuit Forces Speed

What makes time pressure most palpable is the rise of Chinese makers.

In SiC wafers, Wolfspeed (33.7%) is followed by China's Tankebule (17.3%) and SICC (17.1%) in second and third place. Combined, the two Chinese makers approach Wolfspeed's standalone share.

In power semiconductors themselves, BYD Semiconductor, CRRC Times Electric, Silan Microelectronics, Yangjie Technology, and CRRC Times Semiconductor are pursuing aggressive low-price strategies backed by government subsidies. Japan's contract foundry JS Foundry lost the price war and went bankrupt in July 2025.

Against this backdrop, the three-way alliance's "agreement by summer + 2-3 years for full integration" timeline can look leisurely. Urutsuma's emphasis on "early agreement" reflects awareness of Chinese speed.

Why the JV Format Matters

A joint venture (JV) structure carries clear benefits for all three companies.

First, each company keeps its independence. Rohm preserves its identity as an independent Kyoto-based maker while contributing only the power semiconductor business to the JV. Toshiba Device & Storage and Mitsubishi Electric do the same. Less corporate culture and HR system clash than full mergers.

Second, the JV creates an exit ramp through eventual IPO. For Toshiba, currently under JIP (Japan Industrial Partners) ownership, an independent JV listing would be a stepping stone toward effective re-listing of the parent's chip business. All stakeholders benefit.

Third, economic security alignment. Japan's Ministry of Economy, Trade and Industry (METI) recognized Rohm and Toshiba's joint power semiconductor manufacturing plan as a "Supply Assurance Plan" in December 2023, providing combined subsidies of ¥129.4 billion (~$820 million). Mitsubishi Electric's Kumamoto 8-inch fab also received ¥22 billion (~$140 million). Bundling publicly funded businesses under a "national JV" aligns well with policy intent.

But JVs come with challenges. Governance (ownership ratios and decision rights), HR system unification, brand strategy, headquarters location, CEO selection — all must be navigated. Past examples like Japan Display Inc. (JDI) and Elpida Memory show how excessive insistence on "three equal partners" can paralyze decision-making. Rohm CEO Higashi has repeatedly said "too many cooks spoil the broth," but living that principle in real negotiations is what counts now.

Implications for the EV Market

If three-way integration succeeds, what does it mean for EV competition?

The clearest winners are EV manufacturers. They're shifting away from over-reliance on a single dominant supplier (Infineon). A strong "Japan alliance" alternative gives them more leverage. Tesla, Ford, GM, Stellantis, Hyundai, BYD — all gain a "Plan B."

For Japanese automakers like Toyota, Honda, and Nissan, the disappearance of full Denso-Rohm vertical integration leaves an independent three-way JV as a "right-distance partner" they can engage with. That value is meaningful.

That said, EV market deceleration challenges the economic case for three-way integration. EV adoption outside China has plateaued since 2023, and SiC power semiconductor demand forecasts have been repeatedly cut. Mitsubishi Electric's pivot toward data center and HVDC anticipates exactly this trend.

Can "Rohm + Toshiba" and "Three-Way JV" Run in Parallel?

Realizing Urutsuma's vision requires a complex multi-layer structure.

Conceptually:

Layer 1 (JV): Spun-off power semiconductor businesses of all three companies. Covers SiC, GaN, and Si (IGBT/MOSFET). The flagship economic security entity.

Layer 2 (Rohm + Toshiba integration): Non-power-semi businesses — analog, logic, optical devices — integrated. Mitsubishi Electric does not participate.

Layer 3 (Each company's parent): Continues system businesses and consumer operations. Mitsubishi Electric retains heavy electric, factory automation, and home appliances. Rohm keeps sensor ICs and the like. Toshiba retains infrastructure and social-system businesses.

Running this three-layer structure smoothly requires precise design of inter-layer transaction terms, IP sharing rules, and personnel-mobility frameworks. For Rohm's Higashi, Mitsubishi's Urutsuma, and Toshiba's Shimada (with JIP) to deliver "agreement by summer," the structural challenges are non-trivial.

Reading the "Temperature Differences"

The nuances in the three CEOs' public statements are revealing.

Rohm CEO Katsumi Higashi (April 3, earnings briefing): "For sustainable growth, individual companies operating at small scale won't work. Companies in Japan with shared aspirations need to come together." "We have a desire to be at the center." "Too many cooks spoil the broth."

Mitsubishi Electric CEO Kei Urutsuma (April 28, earnings briefing): "I want to spin off the power semiconductor businesses of all three companies and form a JV." "There's no meaning for us to join the non-power-semi integration." "We want to reach agreement quickly."

Toshiba CEO Taro Shimada (limited public statements): At the March 27 basic-agreement announcement, commented that the move "will significantly contribute to the development of broad customer segments and wide industrial fields as Japan's semiconductor business."

Rohm emphasizes "we want to be at the center"; Mitsubishi insists on "power semi only"; Toshiba speaks in broad terms — each reflects the structural realities of their underlying businesses.

Three Focal Points Through Summer

Three issues will dominate negotiations.

① JV ownership ratios: Equal three-way (33.3% each) or weighted by business size? Mitsubishi Electric wants maximum influence in the "power semi JV" but Rohm wants to be the "center of the three-way alliance." Business size: Rohm ~¥480B ($3.0B), Mitsubishi Semiconductor & Device ~¥290B ($1.8B), Toshiba semiconductor ~¥445B ($2.8B) — Rohm and Toshiba are larger.

② Treatment of non-power-semi businesses: How will Mitsubishi engage with the Rohm-Toshiba integration of analog, optical, and other businesses? Urutsuma's stance is "non-participation," but IP licensing and sales cooperation may still apply.

③ Overseas operations: How will the three companies rationalize their China, Europe, and U.S. production and sales footprints? In markets where Chinese competition is fierce, sales structure rationalization is the most critical early-stage issue.

The Endgame Has Begun

What started as the Denso bid story on March 6 has evolved through Denso's April 25 withdrawal — eliminating the "external factor" — and arrived at April 28 with Mitsubishi Electric's framework proposal pushing the "internal design" phase into focus.

If we use a board-game metaphor, the player-selection phase ("who teams up with whom") is over. The strategy-implementation phase ("now that we've teamed up, how do we win?") is beginning.

Urutsuma's "spin off and form a JV" proposal is one of the most operationally feasible approaches to three-way integration. The open question is whether Rohm and Toshiba abandon their original "full semiconductor integration" synergy vision, or pursue it in parallel as a separate two-company track. This is the central debate through summer.

The global power semiconductor market doesn't pause. With Infineon, ST, Wolfspeed, onsemi, and Chinese makers all in fierce competition, Japan's alliance can't afford to dawdle. Early agreement vs. JDI-style stalemate vs. JV that becomes a hollow shell — the next several months will set Japan's power semiconductor industry on its long-term trajectory.

In your country, have you seen multiple companies of different sizes form a "joint venture" alliance? Does industry consolidation as national strategy actually work? How does Japan's "national JV" vision look from the outside? We'd love to hear your perspective.

References