The battery attracts much of the attention in an electric vehicle. Between that battery and the motor sits another engineering challenge: converting power in a package that is light enough, cool enough and reliable enough to earn its place in the machine.

A Nagoya University-led collaboration is addressing that challenge with the “Feather Inverter,” the subject of a special exhibit at NEPCON Japan Autumn at Makuhari Messe. Organizer RX Japan announced the September 9–11 event as a venue for showing the hardware and explaining the technologies behind it. The project is led by Professor Masayoshi Yamamoto of Nagoya University’s Institute of Materials and Systems for Sustainability. [1][3]

The name offers an inviting image. The engineering story is more substantial: weight depends on how semiconductor devices, cooling and supporting components work together. A lighter chip package does not automatically make a lighter, fully usable power-conversion system.

The boundary around the headline number

In the organizer’s published interview, Yamamoto reports a 70% weight reduction against the combined power-semiconductor module and cooler assembly of a Tesla Cybertruck inverter. [2] That is a development-side claim with a specific comparison boundary, not independent verification of a complete inverter’s weight.

It does not mean a vehicle becomes 70% lighter. Nor does it establish the same reduction for every component required to operate the inverter. Before comparing designs, the measurement must cover equivalent hardware doing an equivalent job.

Japan.co.jp’s assessment is that the next useful questions concern continuous output, cooling conditions and the components included on the scale. Peak output available briefly is a different specification from power maintained over time. A compact exhibit cannot answer those questions by appearance alone.

What an inverter actually does

Nagoya’s laboratory describes an inverter as a circuit that converts direct current into alternating current or high-frequency power. Its research spans SiC and GaN conversion, magnetic components and noise reduction. [4] Driving an AC motor from a DC battery is a familiar application of the underlying principle.

The inverter is the assembly; power semiconductors are components within it. Keeping those terms separate matters. A transistor’s capabilities must survive connection to other parts, electrical switching and the removal of heat before they become useful system performance.

ROHM’s application literature explains losses during conduction and switching, among other operating processes, and describes the role of silicon carbide, or SiC, in reducing conversion losses. [7] The material offers opportunities; the operating point still matters. Efficiency needs a load, temperature and measurement method attached to it.

Eight companies, several different jobs

The organizer identifies the following project groups. [2]

AreaParticipants
Power moduleAoi Electronics, Mitsubishi Electric, ROHM
CoolingTaiho Kogyo, Toray
Smoothing capacitorNichicon, Nagoya University
FilterPanasonic Industry, Kitagawa Industries

Its technical account describes an Aoi module integrating Mitsubishi silicon IGBT and ROHM SiC chips, and cooling that combines fine fins with a PPS-resin housing. [2]

In Japan.co.jp’s reading, the significance is the combination. This is not simply a story of replacing every established component with the newest material. It asks whether different strengths can be reconciled within one design.

The connections can determine the outcome. Optimizing a module in isolation may change what the cooler must do. Reducing one component’s dimensions may complicate its connection to the next. The practical challenge lies in retaining the required functions as the package loses mass.

The heat does not disappear

An illustrative calculation shows why cooling remains important even at high efficiency. At 100 kilowatts of input, a 1% loss amounts to 1 kilowatt of lost power. These are hypothetical values calculated by Japan.co.jp, not Feather Inverter specifications.

Reducing cooler mass must therefore be evaluated together with heat-removal performance. If a smaller assembly requires heavier supporting equipment elsewhere, the system-level benefit changes. The same applies to the energy consumed by cooling auxiliaries.

Toray describes heat resistance, dimensional stability and chemical resistance among the characteristics of its PPS resin family. [10] Such properties explain why material selection belongs in the design discussion. They do not substitute for testing the particular housing, joints and operating conditions of an actual converter.

Japan.co.jp’s analysis is that lightweight construction becomes persuasive when the whole thermal arrangement is accounted for. A promising material choice and a demonstrated service life are separate pieces of evidence.

The components that rarely make the headline

Capacitors help stabilize an inverter’s supply. Panasonic’s technical explanation describes film capacitors in voltage stabilization and smoothing, while Nichicon’s electrified-vehicle range includes smoothing, filter and snubber applications. [8][9]

These functions impose their own demands. Size cannot be assessed without considering capacitance, fluctuating current and heating. Filters likewise have a job that must remain satisfied when a design becomes lighter.

The broader lesson, in Japan.co.jp’s assessment, is to watch where a design moves its burdens. A reduction in one part can be valuable, or it can create extra requirements in another. Packaging, wiring and supporting components determine which outcome occurs.

That makes the collaboration interesting as an industrial project. It brings together suppliers whose work may be evaluated separately in a catalog but must operate together in the final machine. The demonstration offers a shared object around which those requirements can be discussed.

The long route from the Prius to SiC systems

The history behind the exhibit extends across several decades. In a 2014 announcement, Toyota said it had developed its own power semiconductors since the first Prius launched in 1997. It traced Toyota Central R&D Labs and Denso’s foundational SiC work to the 1980s, with Toyota joining the collaborative development in 2007. [5]

That announcement described a SiC power-control unit installed in a prototype hybrid and evaluated on a test course. [5] The important historical progression was from a material’s properties toward an assembled vehicle in operation.

Railways provided another route. NEDO’s account of SiC commercialization traces its supporting projects to 1998 and describes the technology’s development for railway inverters. [6] These programs were separate from the Feather Inverter, but they illustrate the long work of connecting devices, circuits and operating equipment.

Earlier automotive or rail results cannot be borrowed as performance figures for the current exhibit. Their value here is historical: Japan’s power-electronics capability developed through integration and testing as well as semiconductor research.

The new lightweight project adds another design objective to that history. Better electrical performance matters, but so do mass, space and the supporting hardware required to achieve it. Success has to be demonstrated at the level relevant to the application.

What follows a trade-show demonstration

RX Japan presents electric vehicles, electric aircraft and industrial equipment as potential applications. [1] Those are prospective destinations, not evidence of customer adoption or completed qualification. Each use will require its own operating assumptions and validation.

Japan.co.jp would look next for complete mass and dimensions, continuous power, efficiency across operating conditions, cooling requirements, durability and production cost. Serviceability also belongs in the discussion: reducing space is less useful if maintaining the equipment becomes impractical.

A comparison becomes especially informative when it identifies the reference hardware and all included parts. Independent replication or customer testing would then make it easier to judge how much of the claimed advantage carries into an application.

The exhibit’s most useful promise is a concrete one: material and component expertise can be assembled into a design that invites measurement. Whether that design becomes a product will depend on the evidence accumulated after the display case, under the conditions in which customers need it to work.

Exhibition details: The organizer lists September 9–11, 2026, 10 a.m.–5 p.m. daily, at Makuhari Messe. Admission is free with advance visitor registration. [1]