Why Single-Stage AC-AC Solid-State Transformers Are the Holy Grail —

A solid-state transformer (SST) can be built in two fundamentally different topologies: single-stage AC-to-AC conversion, or two-stage AC-to-DC-to-AC conversion. The single-stage architecture is the more elegant engineering solution. Every commercial SST shipped or in active development in 2026 uses the two-stage architecture instead. The reason is mostly economic, and the implications shape the entire commercial SST landscape.

A single-stage SST converts utility-frequency AC input directly to load-frequency AC output through a single power-electronics conversion. The medium-frequency transformer (MFT) at the heart of the converter provides galvanic isolation and step-down. The North Carolina State University FREEDM Systems Center targeted a single-stage 480-volt-input, 480-volt-output, 20-kilowatt unit at 99 percent efficiency, but produced only laboratory hardware.

A two-stage SST converts AC input to a DC intermediate bus through a first converter, then synthesizes the AC output from the DC bus through a second converter. The two-stage topology adds losses (each converter stage costs 1 to 2 percent in efficiency, so two-stage architectures land at 96 to 98 percent rather than the single-stage’s 99 percent target). It adds parts count. And it adds DC-link capacitors, which are the single largest reliability liability in modern power electronics.

The two-stage topology won commercially for three reasons. First, modularity: the AC-DC stage and the DC-AC stage can be designed and tested independently, and the design can be reused across products (the same AC-DC front end can feed AI data center 800 VDC distribution, an EV charger DC bus, or a battery storage system). Second, controllability: the DC intermediate bus is the natural injection point for storage, for power-quality compensation, and for multi-port architectures like DG Matrix’s multi-port SST. Third, semiconductor maturity: two-stage modular multilevel topologies were the first to mature with the 1.7-kilovolt and 3.3-kilovolt silicon carbide MOSFET generations from Wolfspeed and Infineon.

The two dominant commercial two-stage families are dual-active-bridge (DAB) and modular multilevel converter (MMC). DAB topologies dominate the megawatt-charger and battery-storage segments because they tolerate wide voltage variations on both sides. MMC topologies dominate the medium-voltage data-center interconnect segment because they scale to higher voltage at the cost of more devices. ABB, Hitachi Energy, Eaton, and Amperesand each ship MMC-based products. Heron Power and DG Matrix favor hybrid DAB-MMC architectures. The single-stage architecture remains a research goal — not a commercial product.

Why It Matters

Buyers comparing SST products should weight the architectural choice by their actual use case, not by efficiency point estimates. AI data center and battery-coupled applications benefit substantially from the two-stage DC bus; railway and motor-drive applications, where the load is predominantly AC-AC, are the rare cases where the single-stage academic argument might actually translate to commercial advantage. Vendors publishing single-stage roadmaps without a clear use case identified are signaling research direction, not product strategy.

Critical Perspective

The single-stage vs two-stage SST debate is presented in vendor literature as primarily an efficiency story, but the more honest framing is that single-stage AC-AC is the academically elegant solution that nobody can produce at scale with current semiconductor pricing. The 2 percent efficiency advantage cited for single-stage at laboratory load is real, but the absolute number (99 percent vs 96-98 percent) understates how small the operational difference is when both architectures are operating below their rated capacity, which they do most of the time. The two-stage architecture’s DC bus enables three commercially valuable features that the single-stage cannot offer: native DC-coupled storage, multiple downstream DC voltages from a single conversion stage, and fault-current limiting via DC isolation. Commercial buyers consistently choose the two-stage product set because those three features matter more to their use cases than the marginal efficiency improvement. The single-stage narrative persists in academic papers because it makes for better thesis material, not because commercial customers are missing the opportunity.

Related Coverage

• FREEDM Systems Center at NCSU: The 2008 NSF Center That Turned Solid-State Transformers Into Real Devices: FREEDM’s early research enabled solid-state transformer development.
• DG Matrix Ships First Commercially Available Multiport Solid-State Transformer Aligned With NVIDIA 800VDC AI Factory Spec: Commercial multiport SSTs align with AI factory 800VDC specs.
• ABB Takes Minority Stake in DG Matrix as Part of $20M Seed Round: The Largest Western Transformer Maker Bets on a Startup: ABB invests in DG Matrix, a startup in transformer tech.
• The Federal R&D Investment Ecosystem Under the US Solid-State Transformer Industry: ARPA-E, DOE OE, NSF FREEDM, and Why It Matters: Federal R&D funding supports the US solid-state transformer industry.