Infineon and Eaton Push Silicon Carbide Into 800 VDC AI Power

Illustration of an Eaton medium-voltage solid-state transformer using Infineon silicon carbide power devices for an 800 VDC AI data center

Infineon and Eaton are pushing silicon-carbide power electronics deeper into the AI data center, pairing Infineon SiC devices with Eaton’s MVSST 2.0 medium-voltage solid-state transformer for emerging 800 VDC power architectures.

In the companies’ September 29 announcement, Infineon said it will supply SiC power devices for Eaton’s MVSST 2.0 platform targeting deployments in Asia-Pacific and Europe, the Middle East and Africa. The design is intended to connect medium-voltage grid power more directly to the DC electrical architecture feeding high-density AI computing.

Solid-state transformers remove conversion steps

Conventional data-center power paths can pass electricity through several stages before it reaches server racks. Eaton’s solid-state approach uses high-frequency power electronics instead of relying only on traditional low-frequency magnetic conversion, allowing the platform to reduce conversion stages while controlling voltage electronically.

Infineon’s role is at the semiconductor layer. Silicon-carbide switches can operate at higher voltages and switching frequencies than conventional silicon devices, helping power-conversion systems increase power density while reducing losses and the size of some passive components.

Eaton is moving MVSST hardware toward live data-center trials

The partnership builds on hardware Eaton is already taking toward field validation. In a separate engineering program, Eaton says a 2.5 MW medium-voltage solid-state-transformer container will be installed at Singapore’s Jurong Island for live trials through the Sustainable Tropical Data Centre Testbed.

That scale matters because 800 VDC is moving from a component-level idea toward complete grid-to-rack architectures. BitcoinVersus.Tech recently covered SolarEdge and Infineon’s microsecond protection for 800 VDC AI power, ABB’s 800 VDC data-center distribution system, and Delta’s 800 VDC power and megawatt-scale cooling architecture. Eaton’s transformer addresses a different part of the chain: conversion from medium-voltage infrastructure toward the DC bus.

The next SiC step goes beyond today’s voltage class

Infineon and Eaton say they are also exploring future solid-state-transformer platforms using 2.3 kV and 3.3 kV SiC power modules. Higher-voltage devices could let designers handle more of the conversion task with fewer series-connected components while targeting greater efficiency, density and reliability.

The immediate MVSST 2.0 platform is aimed at high-density computing and emerging 800 VDC distribution. Infineon says the system is among the early medium-voltage solid-state transformer platforms to achieve IEC certification, an important step for equipment intended to sit between utility-scale electrical infrastructure and expensive AI compute.

For AI data centers, the semiconductor race increasingly extends outside the GPU. Power MOSFETs, circuit breakers, transformers, battery systems and cooling hardware all determine how many accelerators a site can reliably energize. The Infineon-Eaton collaboration puts silicon carbide directly into that grid-to-rack competition.


BitcoinVersus.Tech

Advertisement: Follow BitcoinVersus.Tech for independent coverage of semiconductors, data centers, artificial intelligence, Bitcoin mining, power infrastructure and energy.

Follow BitcoinVersus.Tech for original semiconductor, data-center power and infrastructure reporting.

Disclaimer: BitcoinVersus.Tech provides news, analysis and educational information for general informational purposes. Nothing in this article is financial, investment, legal or professional advice. Readers should independently verify information and make their own decisions.

10 responses to “Infineon and Eaton Push Silicon Carbide Into 800 VDC AI Power”

  1. […] following the physical side of that rack transition through 3.5 MW liquid-cooling distribution and 800 VDC AI power architectures. Lattice is focusing on the control layer that has to make increasingly dense infrastructure […]

    Like

  2. […] has been following, including xAI’s large Megapack battery installation behind Colossus 2, Infineon and Eaton’s silicon-carbide push into 800 VDC AI power, SolarEdge and Infineon’s solid-state protection for 800 VDC systems, and Delta’s 800 […]

    Like

  3. […] is not equivalent to usable compute capacity. Our coverage of Bitdeer’s AI-cloud expansion and silicon-carbide power systems for 800 VDC AI infrastructure shows how sites increasingly compete on the complete electrical and thermal stack rather than on […]

    Like

  4. […] materials story complements our coverage of silicon carbide in AI power distribution, GaN and SiC power development, and higher-voltage semiconductor research. Those are separate […]

    Like

  5. […] makes this agreement distinct from Infineon and Eaton’s transformer-platform collaboration. Here, local storage and peak shaving are part of the development scope alongside […]

    Like

  6. […] previously covered Infineon and Eaton’s silicon-carbide work for 800 VDC infrastructure. The Skeleton agreement adds a different emphasis: pairing conversion hardware with fast energy […]

    Like

  7. […] infrastructure. BitcoinVersus recently covered ABB’s 800 VDC AI data-center power system, while Infineon and Eaton are pushing silicon carbide into the same 800 VDC transition. The common goal is to move more power through fewer conversion stages without allowing switching […]

    Like

  8. […] of a long-term industrial-policy bet spanning power electronics, communications, defense systems, AI data-center power, electric vehicles, and grid […]

    Like

  9. […] This is the same broader semiconductor transition BitcoinVersus has been following in 800 VDC silicon-carbide infrastructure and in other SiC and GaN AI power […]

    Like

  10. […] move toward 800 VDC AI power is partly about current. For the same amount of power, raising voltage reduces current. Lower […]

    Like

Leave a Reply