Solid-State Transformers Are Tackling the Grid's Most Wasteful Problem

Electricity disappears somewhere between the power plant and your data center rack. Not dramatically — no explosions, no blackouts — just a slow, invisible bleed. Roughly six to eight percent of all electricity generated globally is lost to grid inefficiencies, most of it in the transformation process that every electron must survive before it powers anything useful. That's not a rounding error. At a global scale, it represents hundreds of millions of tonnes of CO2 annually and billions of dollars in wasted generation capacity.

Solid-state transformers won't fix the entire grid, but they might address the part that's about to buckle under the weight of the AI era.

What Makes a Solid-State Transformer Different

A conventional transformer is, at its core, a 19th-century technology wrapped in increasingly expensive copper and silicon steel. It works by electromagnetic induction — two coils wound around an iron core, stepping voltage up or down as needed. The physics are elegant, but the hardware is massive, heavy, and slow to manufacture.

Solid-state transformers replace the iron core with high-frequency power electronics, performing the same voltage conversion digitally at far higher frequencies. The result is a device that can be a fraction of the size and weight of its conventional counterpart, with greater controllability, bidirectional power flow capability, and — critically — significantly lower losses.

A recent study from the Energy Research Institute at Nanyang Technological University in Singapore quantified the efficiency opportunity: widespread SST deployment could reduce grid losses by up to 25 percent. That's not a theoretical ceiling — it's a credible estimate of what happens when you swap out aging electromagnetic infrastructure for precision power electronics at scale.

The tradeoff, historically, has been cost and complexity. High-frequency semiconductors capable of handling grid-level voltages are expensive, and the control systems required to manage them are sophisticated. That's why SSTs have spent years as a promising research concept rather than a commercial product.

That's changing.

Why Data Centers Are Driving the Urgency

The data center industry has developed a transformer problem that nobody planned for. As AI workloads intensify and GPU power density climbs, operators are deploying server racks that draw 30, 50, even 100+ kilowatts each — numbers that would have seemed absurd five years ago. The electrical infrastructure supporting those racks — including the transformers stepping down grid voltage for facility use — was not designed for this load profile.

The supply chain hasn't kept up. Lead times for conventional large power transformers now stretch 18 to 24 months in many markets. For a hyperscale operator trying to bring a new campus online in 12 months, that gap is existential. You can design the building, pour the concrete, and rack the servers faster than you can get the transformer to power them.

This is the bottleneck that SST startups are positioning themselves to solve — not just more efficient electricity conversion, but a fundamentally different supply chain dynamic. SSTs use fewer raw materials, require less specialized manufacturing, and — if production scales — could reduce both lead times and cost per megawatt of capacity.

Hyperscale Power's co-founder Daniel Rothmund framed it directly: "The power demand of new GPU models is increasing exponentially, creating a mismatch with traditional power systems. Server racks are becoming significantly denser, and our technology provides the ideal solution."

The company just closed a €5 million (~$5.7M) seed round led by World Fund and Vsquared Ventures — investors focused on climate tech and deep tech, respectively — with the funding earmarked for building Hyperscale Power's first physical prototype. Founded in 2025, the startup is targeting the same three-sector opportunity that defines SST demand right now: data centers, large-scale renewable energy projects, and EV charging infrastructure.

The Renewable Energy Connection

SSTs aren't just a data center story. Renewable energy integration is arguably where the technology matters even more.

Solar and wind generation is inherently distributed and variable. The grid infrastructure connecting those generation sources to load centers was designed around centralized, dispatchable power plants. Voltage fluctuations, reactive power management, and bidirectional flow — things that happen constantly in a renewable-heavy grid — stress conventional transformers in ways they weren't built to handle.

Solid-state transformers handle all of those conditions natively. Their power electronics can respond in microseconds to voltage changes, manage reactive power actively, and route electricity in either direction without mechanical modification. As grids absorb more solar and wind capacity, that flexibility becomes less of a nice-to-have and more of a prerequisite.

The efficiency gains compound here too. World Fund's managing partner Daria Saharova noted that a two percent efficiency gain across data center power consumption at a global scale "translates into millions of tonnes of CO2 avoided." Apply that logic across the entire grid — where six to eight percent of generation is currently lost — and the climate math becomes substantial.

A Crowded but Early Market

Hyperscale Power isn't alone. The SST startup space includes Heron Power, Ampersand, and DG Matrix, among others. DG Matrix is currently the best-funded of the cohort, having closed approximately $60 million in Series A financing last month — a round backed by Chevron Technology Ventures, Helios Climate Ventures, and ABB, the Swiss-Swedish industrial giant that brings serious grid hardware credibility to the cap table.

The presence of ABB as an investor in DG Matrix is worth paying attention to. ABB has been building conventional transformers for over a century. When a company like that bets on an SST startup, it signals one of two things: either they believe the technology is genuinely disruptive and they want optionality, or they're hedging against their own core business being disrupted. Probably both.

The €5M seed that Hyperscale Power raised is modest by comparison — but seed rounds in deep hardware are different animals than software rounds. The money isn't going to sales and marketing. It's going to build a prototype, validate the physics at scale, and generate the performance data needed to raise a Series A. For a company founded in 2025, that's exactly the right use of capital.

The insider perspective here: hardware deep tech lives and dies by the prototype. A credible demonstration unit that performs at spec opens doors to strategic partnerships, pilot deployments, and the kind of customer letters of intent that de-risk the next funding round. Hyperscale Power's first milestone isn't a product — it's proof.

What Happens Next

SST adoption won't be a sudden flip. The installed base of conventional transformers is enormous, utilities are conservative by nature, and grid interconnection standards weren't written with solid-state devices in mind. Regulatory pathways in most markets are still being established.

But the data center sector operates on faster cycles than utilities, and the combination of supply chain pressure, rising power density, and sustainability commitments is creating genuine pull for alternatives. Several hyperscale operators have already begun exploring SSTs for specific applications — medium-voltage distribution within campuses, for example — where the size and controllability advantages are most immediately valuable.

The next 24 months will likely determine which SST developers have the technical and commercial traction to survive the valley between seed funding and real revenue. Hyperscale Power's prototype will be a critical data point. So will DG Matrix's ability to translate its $60M war chest into deployed projects.

The iron-core transformer has had a 150-year run. It's not going away overnight — but the conditions that would accelerate its replacement are stacking up faster than most people in the industry expected. The companies that figure out how to manufacture SSTs reliably, at cost, and at volume will be selling into one of the most urgent infrastructure needs of the next decade. That's a strong place to be building toward.

[INTERNAL LINK: solid-state transformers]

[INTERNAL LINK: renewable energy integration]

[INTERNAL LINK: data center efficiency]

EDITOR NOTES:

• Consider cutting the paragraph about the historical tradeoff of cost and complexity if it feels redundant.
• Ensure to include a compelling CTA at the end that links to the InfraSale Marketplace.