Solid State Transformers Could Reshape AI Infrastructure

Solid State Transformers Could Reshape AI Infrastructure

VC Firms Pour $280 Million Into Startups Heron Power, DG Matrix and Amperesand

Three companies building next-generation power conversion technology have raised a combined $280 million in the past year, while another has been acquired by a major power vendor. These deals anticipate that the solid-state transformer (SST) will play a key role in the next phase of the AI buildout.

The transformer, a device whose core design dates to the 19th century, is giving way to software-defined power electronics that are purpose-built for the AI era.

The SST is emerging as a critical piece of the transition to 800 VDC power distribution, the architecture that NVIDIA and its partners are driving for next- generation AI infrastructure. SSTs also have the potential to help address “speed to power” for AI data center campuses, which face delays in the transformer supply chain.

The numbers tell part of the story:

Heron Power closed a $140 million Series B in February 2026, co-led by Andreessen Horowitzʼs American Dynamism Fund and Breakthrough Energy Ventures.

DG Matrix closed a $60 million Series A the same month, led by Engine Ventures with participation from Mitsubishi Heavy Industries, ABB, and others, bringing total capital raised to over $100 million.

Amperesand raised an $80 million Series A in November 2025, co-led by Walden Catalyst Ventures and Temasek, targeting 30 megawatts of commercial deployments in 2026.

Eaton completed its acquisition of Austin-based Resilient Power Systems in August 2025, bringing medium-voltage SST technology into one of the worldʼs largest power management companies.

Meanwhile, two other public companies, Enphase Energy and SolarEdge, have recently launched SST offerings specifically targeting AI data centers.

The common thread: the traditional transformer cannot keep pace with the speed, density, and power scale that AI infrastructure now demands.

“The transformer hasnʼt fundamentally changed in 100-plus years,” said Haroon Inam, Founder & CEO, DG Matrix. “Copper or aluminum windings around a steel core, likely oil-filled, built for a single purpose. That works when a server rack pulls 5 kilowatts and a workload is predictable.

“It doesnʼt work when a single NVIDIA MGX rack demands 600 kilowatts, swings from idle to full load in microseconds, and sits behind a grid interconnection queue thatʼs three to five years long.”

For the data center sector, SSTs are an emerging technology currently moving from the lab and pilot work into prototypes and first commercial offerings. Broad deployment will require operator comfort at data-center scale.

But the AI boom offers a compelling opportunity for SST specialists, and investors and equipment vendors are paying full attention. Hereʼs a deep dive into SSTs and the leading companies and supporting players in the STT ecosystem.

Why Solid-State Transformers Matter

A solid-state transformer (SST) replaces traditional windings and cores with power semiconductors, primarily silicon carbide (SiC) or gallium nitride (GaN) devices, and software-defined control logic.

The SST converts voltage like a conventional transformer, but also routes power from multiple sources simultaneously, responds to load changes in milliseconds, integrates battery backup without separate UPS hardware, and provides real-time grid stabilization.

The physical difference is significant. SST-based systems can reduce electrical equipment footprint by 70 to 80 percent, eliminating layers of switchgear, distribution transformers, and UPS systems that a conventional data center power chain requires.

By supporting multiple ports, they are also ideal for microgrid integrations that combine generation sources.

“Too much of todayʼs electrical infrastructure is passive, clunky equipment designed decades ago,” said Drew Baglino, Founder and CEO of Heron Power. “We need new, more capable solutions to keep pace with accelerating energy demand and the rapid growth of gigascale compute.”

Addressing Data Center Bottlenecks

The data center power problem has two distinct layers.

The first is speed to power. Utility power constraints are creating lengthy grid interconnection timelines, while lead times for medium- and high-voltage transformers can now run 18 to 36 months. SSTs can address this directly by collapsing procurement and installation timelines, and in some architectures can connect directly to medium-voltage distribution grids without a traditional transformer layer.

The second is density: getting power to the rack efficiently once it arrives. AI workloads have pushed rack power requirements past 100 kilowatts (kWs), with NVIDIAʼs Rubin Ultra platform targeting up to 600 kilowatts per rack. The conventional 54-volt DC in-rack distribution architecture hits hard physical limits at those densities.

AI training loads add a third wrinkle: a GPU cluster can swing from idle to full power in seconds. SSTs with integrated energy storage absorb those load spikes at the rack level, protecting both the facilityʼs internal distribution and the wider grid.The data center power problem has two distinct layers.

The first is speed to power. Utility power constraints are creating lengthy grid interconnection timelines, while lead times for medium- and high-voltage transformers can now run 18 to 36 months. SSTs can address this directly by collapsing procurement and installation timelines, and in some architectures can connect directly to medium-voltage distribution grids without a traditional transformer layer.

The second is density: getting power to the rack efficiently once it arrives. AI workloads have pushed rack power requirements past 100 kilowatts (kWs), with NVIDIAʼs Rubin Ultra platform targeting up to 600 kilowatts per rack. The conventional 54-volt DC in-rack distribution architecture hits hard physical limits at those densities.

AI training loads add a third wrinkle: a GPU cluster can swing from idle to full power in seconds. SSTs with integrated energy storage absorb those load spikes at the rack level, protecting both the facilityʼs internal distribution and the wider grid.

The 800 VDC Transition

NVIDIA is working with more than 20 AI infrastructure providers, including CoreWeave, Lambda, Nebius, and Oracle Cloud Infrastructure, to design data centers built around 800-volt direct current power distribution, or 800 VDC. The target platform is NVIDIAʼs Rubin Ultra GPU generation, expected in 2027, which will support rack densities up to 600 kilowatts. Todayʼs AI clusters run at roughly 120 kilowatts per rack.

Moving to 800 VDC eliminates most of the intermediate conversion steps in the power chain, with NVIDIA projecting up to a 5 percent improvement in end-to-end power efficiency, a 70 percent reduction in maintenance costs from fewer power supply unit failures, and a meaningful reduction in cooling load by eliminating AC/DC conversion hardware inside the racks.

“800 VDC fixes the architecture, not just the voltage,” said Inam of DG Matrix. “By taking the data center to DC at the row, or ideally at the building, operators eliminate the rectifier-inverter conversion step entirely. NVIDIAʼs MGX reference design recognizes this, which is why they are standardizing 800 VDC across the ecosystem.

The end goal is achieving 34.5kV AC to 800V DC power conversion, creating a simplified pathway from grid to rack, with native per-rack 800V DC outputs aligned to industry reference designs. That single-step conversion from medium-voltage grid power to rack-ready 800 VDC is the architectural move outlined in NVIDIAʼs blueprint.

“As data centers approach gigawatt-scale campuses built around megawatt AI compute racks, they are rightfully taking inspiration from mature gigascale DC technologies like energy storage and solar,” said Baglino of Heron Power. “Direct 34.5kV AC to 800V DC conversion is the efficient, streamlined, scalable power architecture that global leaders in AI need to achieve their goals.”

That said, NVIDIA has noted that deploying 800 VDC at the facility level introduces new challenges in safety, standards, and workforce training, and that the transition will occur in phases.

The first phase of the 800 VDC transition is expected to feature “sidecar” racks installed alongside compute racks, enabling denser GPU and CPU racks while separating out power components for 800 VDC conversion.

Full-scale production is tied to the 2027 Kyber rack rollout, and the neoclouds optimizing for NVIDIAʼs AI factory designs are likely to move first.

The SST Data Center Ecosystem

Hereʼs a look at the leading companies developing SSTs and the supporting ecosystem.

NOTE: We will discuss some public companies in this section. The information provided here is for informational and educational purposes only, and does not constitute investment advice or any other professional advice. Please consult with a professional for specific advice tailored to your situation.

DG Matrix

DG Matrix is the pacesetter in productizing solid-state transformers for data center use. It has developed the Interport SST platform, which routes power from multiple sources simultaneously to multiple loads at differing voltages.

The company says the Interport platform is “purpose-built” to integrate with the NVIDIA MGX modular rack architecture to deliver 800 VDC to GPU racks. DG Matrix has also conducted a yearlong program of testing energy storage solutions for dynamic GPU load response in collaboration with NVIDIA.

“This is a pivotal moment for the industry,” said Inam. “Our multi-port capability enables integrated behind-the-meter power aggregation from multiple AC and DC sources along with simultaneous AC and DC outputs, offering a future-proofed platform today.”

Three partnerships reinforce the companyʼs commercial momentum.

DG Matrix just announced a strategic partnership with modular developer InfraPartners, which will pair its prefabricated data centers with the DG Matrix Interport 360 to create a “grid-to-rack” power management solution.

Exowatt, the renewable energy startup backed by a16z and Sam Altman, selected Interport as its preferred power conversion platform for gigawatt-scale solar data center builds. See our podcast episode with Exowatt for more on its technology.

DG Matrix PowerSecure has formed a strategic collaboration with DG Matrix for AI data center and electrification deployments.

DG Matrix has installed a working pilot at Exowatt, and says it is shipping production units for other customers for multiple use cases.

“We are working with hyperscalers, energy companies, and industrial customers across North America and globally, with multiple gigawatt-class data centers in the pipeline,” said Inam.

The company recently got a vote of confidence from data center pioneer Christian Belady, who joined as an advisor.

“AI has rewritten the power requirements of a data center,” said Belady. “DG Matrix has built the first commercially available multi-port solid-state transformer by collapsing the data center electrical system of discrete devices into a single device … That is the kind of architectural shift the industry needs.”

The company is a privately-held, VC-backed startup. Investors in DG Matrix include ABB, which is also a legacy transformer incumbent, another signal that established power players are engaging with SST innovation.

Heron Power

Heron Power was founded by Baglino, former SVP of Powertrain and Energy at Tesla, where he led development of the Powerwall and Megapack and the buildout of a 50 GWh battery factory. That manufacturing scale-up experience is directly relevant to what Heron is attempting: not just a new product, but a new product category at industrial volume.

Heron Link converts medium-voltage AC, including 34.5 kV distribution, directly to

800 VDC in a single conversion step, natively compatible with NVIDIA's 800 VDC rack architecture. Each unit handles up to 4.2 megawatts and includes integrated battery storage that can eliminate the need for separate UPS systems.

NVIDIA has recognized Heron as a data center power system provider within its Open Innovation Ecosystem. Heron recently shared its blueprint for 800 VDC data centers outlining its strategy for supporting Kyber racks.

“At Heron we are manifesting an alternative future, where modern power electronics enable projects to come online faster, the grid to operate more reliably, and scale affordably,” said Baglino.

The companyʼs $140 million Series B funding provides resources for building a 40 GW per year U.S. manufacturing facility. Internal field demonstrations are expected mid- 2026, with partner installations in early 2027 and manufacturing ramp through 2028.

Amperesand

Amperesand was spun out of Singaporeʼs Nanyang Technological University and Temasekʼs Xora Innovation following years of SiC power electronics research, and has built a leadership team with deep experience in large-scale power systems.

The company initially focused on decarbonization for ports and industrial customers, but says it also is targeting the huge market for data center power infrastructure. Its medium voltage SST platform consolidates medium-voltage transformers, switchgear, and UPS systems into a single factory-built product.

The companyʼs oversubscribed $80 million Series A closed with new investors including Industry Ventures, Acclimate Ventures, and SG Growth Capital alongside continuing backers. First commercial units are being delivered this year to the Port of Singapore in collaboration with PSA International, and the company says it has multiple pilots with undisclosed hyperscale AI data center operators.

Eaton / Resilient Power Systems

Eatonʼs acquisition of Resilient Power Systems, completed in August 2025, brings institutional manufacturing scale and a global field service organization to medium- voltage SST technology. Resilientʼs initial product connected EV charging depots directly to distribution grids. Data centers are the next application.

“Resilientʼs innovative technology, offering high density electrical power in a smaller footprint than comparable solutions,” said Heath Monesmith, President and COO of Eatonʼs Electrical Sector. “We are excited for the opportunity to scale this next- generation medium voltage solid-state transformer technology for growing global markets like data center and energy storage.”

Eatonʼs Q2 2025 investor materials described the Resilient acquisition as strengthening its power distribution offering for data centers and called SST technology “a critical building block in future high-power AI data center designs.”

What Eaton (NYSE:ETN) brings is what the startups lack: customer relationships with utilities and large enterprises, and manufacturing scale.

Enphase Energy

The most recent major entrant is Enphase Energy. On April 28, the residential solar microinverter leader announced development of the IQ Solid-State Transformer, saying it was purpose-built for AI data centers.

The IQ SST deploys a “supercluster” of 342 intelligent power modules, which use GaN switching rather than SiC. The system converts medium-voltage AC directly to 800 VDC or ±400 VDC in a single stage, with 98.5 percent targeted efficiency and 99.999 percent availability through distributed redundancy.

“AI is changing how power must be delivered to compute infrastructure,” said Badri Kothandaraman, President and CEO of Enphase (Nasdaq: ENPH). “For two decades, Enphase has built distributed, semiconductor- and software-defined power conversion systems at scale. As AI racks move toward 800 VDC architectures and megawatt-scale densities, we see a large opportunity to apply this expertise to data center power infrastructure.”

The companyʼs timeline calls for full system demonstrations in late 2026, customer pilots in 2027, and volume shipments in 2028.

SolarEdge Technologies

In November 2025 SolarEdge announced a collaboration with Infineon Technologies to develop a modular 2 to 5 MW SST for AI and hyperscale data centers. The joint development targets greater than 99 percent conversion efficiency and direct medium-voltage to 800 to 1500 VDC conversion, combining SolarEdgeʼs DC architecture expertise with Infineonʼs SiC switching technology.

“The AI revolution is redefining power infrastructure,” said Shuki Nir, CEO of SolarEdge (Nasdaq:SEDG). “Collaborating with Infineon brings world-class semiconductor innovation to our efforts to build smarter, more efficient energy systems for the AI era.”

SolarEdge followed in March 2026 with a published roadmap report arguing that legacy AC-based data center power systems may soon be unable to support accelerating compute density. The company has begun engaging with potential customers but has not specified commercial timelines.

For more on SolarEdge, hereʼs our Data Center Richness podcast episode with Dafna Granot, the Senior Manager for Strategy & Innovation at SolarEdge.

Delta Electronics

Delta Electronics has the most tangible proof point in the field: a production SST deployment in an operating AI data center. In February 2026, Deltaʼs SST system went live at a Chindata Group hyperscale campus in North China, installed for internet platform company Meituan.

Delta says its SST system employs a modular design, with a single power cabinet delivering up to 1 MW while occupying just one square meter, offering a 50% space savings compared with conventional solutions.

Deltaʼs system uses proprietary solid insulation technology and SiC power conversion to deliver multi-level DC outputs of 240V, 400V, and 800V. Meituan, Chindata, and Delta jointly conducted technical research and development and ultimately decided to deploy Deltaʼs indoor-type SST system.

“During the process of implementing innovation, we faced multiple challenges,” said Xin-Ping Ye, Product Marketing Director of the ICT Infrastructure Department, Delta GreenTech China, who said the teams “custom-engineered an intelligent power-delivery architecture featuring high power density and multi-voltage outputs.”

At NVIDIA GTC in March 2026, Delta also showcased 800 VDC in-row power racks developed in collaboration with NVIDIA.

Hitachi, Siemens, Schneider Electric,Vertiv

Hitachi Energy, Siemens, Schneider Electric and Vertiv all maintain active SST research programs targeting data centers, though none has announced a commercially available product.

All four companies are partners in NVIDIAʼs Omniverse DSX initiative and are expected to be players in the 800 VDC transition. It remains to be seen whether these companies will offer their own solution, partner, or follow Eatonʼs strategy and enter the SST market through acquisition. Vertiv and Schneider each have a long history of acquiring promising providers in data center infrastructure.

The Semiconductor Layer

Conventional power electronics run on silicon. Solid-state transformers depend instead on wide-bandgap (WBG) materials, principally silicon carbide (SiC), with gallium nitride (GaN) in a growing supporting role. “Bandgap” refers to the energy required to push electrons into conduction. As a wide-bandgap material, SiC handles higher voltages, higher temperatures, and faster switching than traditional silicon.

That matters because faster switching at higher frequencies lets SST designs shrink their magnetic components and shed heat, with lower switching and conduction losses. In practice, SiC tends to carry the medium-voltage front end where utility power first enters, while GaN is strongest in the lower-voltage, high-frequency conversion stages closer to the load.

For power-semiconductor makers, the business opportunity is supplying the medium- and high-voltage SiC (and GaN) that make SSTs and the wider HVDC chain viable.
Their approaches diverge: some sell devices into third-party SST integrators, others push a flagship high-voltage part as the enabling technology, and all anchor their positioning to NVIDIA's roadmap.

Semiconductor Providers for SSTs

Infineon Technologies is a global leader in power semiconductors. Its CoolSiC line anchors a portfolio that spans silicon, silicon carbide, and gallium nitride. Infineonʼs strategy is to be the default component and subsystem supplier sitting underneath multiple competing SST designs. In March 2026 it partnered with DG Matrix, which will source the latest SiC for its Interport platform. It is also working with SolarEdge on a modular 2-to-5 MW SST building block, targeting greater than 99% efficiency for 800 VDC data centers. In September 2025 it signed an agreement with ROHM to collaborate on SiC products used in AI data centers.

Navitas Semiconductor (Nasdaq:NVTS) is a wide-bandgap power specialist built on GaNFast gallium nitride and GeneSiC silicon carbide. Navitas holds the most visible customer endorsement in the SST value chain, having been named a collaborator on NVIDIA's 800V HVDC architecture for the Kyber racks powering Rubin Ultra GPUs.

Navitas supplies high-voltage SiC for the grid-to-800V conversion stage, including SSTs, and GaN for downstream rack delivery, and positions itself as an enabler embedded in NVIDIA's roadmap rather than an SST integrator in its own right.

Wolfspeed - This companyʼs role in the SST value chain rests on its high-voltage silicon carbide, headlined by the industry's first commercially available 10 kV SiC MOSFET aimed squarely at the medium-voltage grid interface that NVIDIA's 800 VDC vision depends on, now backed by new 3.3 kV modules. Having emerged from a 2025 Chapter 11 restructuring that cut debt by about 70%, it sells device leadership while rebuilding financial credibility with customers. This month Wolfspeed (NYSE:WOLF) launched a dedicated data center solutions team and regional office in the Bay Area.

onsemi (ON Semiconductor, Nasdaq:ON) frames its role across the entire SST and HVDC chain, from substation conversion down to the processor, listing solid-state transformers explicitly among its targets and collaborating with NVIDIA on the 800 VDC transition on the strength of decades of silicon and EliteSiC. It aims to be a broad systems supplier across the chain rather than a single-stage SST specialist.