Funding & VC

SolidUltra’s €3 Million Plan for 800-Volt AI Data Centers

SolidUltra secured €3 million to develop modular 800VDC transformers for AI data centers, targeting a 2028 launch, 10,000 systems and about 1 gigawatt by 2032.

By Amara Osei

4 min read

SolidUltra Secures €3 Million for Modular Power Transformers Built for AI Data Centers
SolidUltra Secures €3 Million for Modular Power Transformers Built for AI Data CentersAkela999 / pixabay

By Amara Osei

SolidUltra said it secured €3 million within three months of its June 2026 incorporation to develop modular solid-state transformers for AI data centers, targeting commercial deployments in 2028.

The Massy, France-based company is developing a medium-voltage solid-state transformer, or SST, to convert grid electricity into the 800-volt direct current planned for future high-density computing facilities. SolidUltra said the financing will fund engineering, product development, industrialization and early customer engagement before its targeted 2028 deployments. French power-electronics group Watt & Well, a founding partner, backed the financing.

What is SolidUltra building?

Conventional data-center power systems commonly pair a low-frequency transformer with additional conversion equipment before power reaches computing hardware. SolidUltra said each stage carries its own electrical losses, cooling requirements and physical footprint, trade-offs that grow harder to manage as operators plan megawatt-scale racks.

SolidUltra intends its SST to replace that arrangement with integrated grid-to-800VDC conversion. The equipment supplies AI infrastructure; it does not perform AI inference.

According to SolidUltra’s website, the modular design uses 100-kilowatt cells connected in series on the medium-voltage input and in parallel on the DC output. The cells use automotive-grade silicon-carbide components and provide redundant, field-replaceable modules, allowing operators to scale with repeated power blocks rather than redesign equipment for each capacity increase.

SolidUltra’s funding announcement projects a range from hundreds of kilowatts to 10 megawatts and beyond. The company says operators can expand conversion capacity alongside computing demand by assembling repeated power blocks.

Why does 800VDC matter to AI facilities?

SolidUltra says higher distribution voltage requires less current at a given power level, which can reduce conductor demands and resistive losses. NVIDIA’s 800VDC roadmap covers both existing AC-powered facilities and new DC-native AI factories, with work through the Open Compute Project to align equipment requirements across the ecosystem.

SolidUltra argues that rising compute density has made electrical design a growing constraint on AI expansion. The company says adding accelerators requires reliable power without excessive space or energy loss, not merely access to chips.

SolidUltra notes that an 800VDC supply remains an intermediate distribution voltage; chips require further conversion before electricity reaches them. The company is pursuing a new power chain built around direct medium-voltage conversion, rather than local conversion equipment added inside an existing facility.

The startup said its alignment with the 800VDC transition does not, by itself, establish a commercial partnership with NVIDIA.

What must SolidUltra prove before deployment?

SolidUltra set five development targets for the planned product:

  • Conversion efficiency: 98.5%.
  • Power losses: Three times lower than conventional architectures.
  • Equipment footprint: Reduced by a factor of four.
  • Total cost of ownership: Approximately 30% lower.
  • Power-conversion carbon footprint: Up to a tenfold reduction.

All five remain development targets rather than independently demonstrated operating results from commercial installations, SolidUltra said. The carbon target applies to conversion infrastructure, while cost and loss comparisons depend on the conventional baseline. Conversion efficiency is also distinct from whole-data-center efficiency.

SolidUltra says its field-replaceable power modules are designed to be swapped in under 15 minutes. The company targets 99.999% availability, with granular replacement intended to limit the capacity affected when one module needs service.

The company also aims to shorten equipment deployment from years to weeks through a factory-built design. SolidUltra says faster installation could bring prepared sites online sooner, but it does not eliminate utility-capacity, grid-connection, permitting or other facility-electrical requirements.

Who is backing SolidUltra?

According to SolidUltra, CEO Raul Iglesias, Chief Business Officer Romain Beaume, Alain Rouault and Watt & Well co-founded the company in June 2026. The model pairs a company focused on AI infrastructure with Watt & Well’s established power-electronics engineering and manufacturing capabilities.

SolidUltra said Watt & Well develops modular power equipment for electric-vehicle charging, including AC/DC and DC/DC applications. The partner also provides bidirectional conversion products for battery energy storage and serves energy and aerospace applications.

That engineering and manufacturing experience is relevant to producing power hardware. SolidUltra said its proposed data-center system still faces separate development and qualification work.

“AI is not only changing computing. It is fundamentally changing the electrical infrastructure required to power it,” Iglesias said in the funding announcement. He described the €3 million investment as a step toward industrial deployment with Watt & Well’s backing.

What does the 2028 launch require?

SolidUltra plans a 2028 commercial launch, then aims to scale production to 10,000 modular systems by 2032. The company said those systems would represent approximately 1 gigawatt of installed power.

SolidUltra said the 10,000-system figure is a production ambition, not a disclosed order backlog. The €3 million provides resources to advance engineering, industrialization and early customer engagement as the company moves from its proposed architecture toward deployable equipment.

SolidUltra identifies a timing mismatch: AI racks are becoming denser while supporting electrical infrastructure takes years and considerable space. The next test is commercial execution—qualifying the hardware, converting early customer engagement into deployments and reaching the 10,000-system production target at scale.

Source: unite.ai

Share this article:

More from Amara Osei

Amara Osei

Show full bio

Senior reporter covering companies, deals and strategy in entrepreneurship, startups and venture funding at Business Bearings.

1,718 articles

Related articles

« Previous articleNext article »