Why GaN specialists like Orchard Audio still lead the technology's edge

The small brands are still winning the GaN race

In February 2026, Orchard Audio quietly released the Starkrimson 25 Mono Premium, a pair of gallium nitride monoblock amplifiers priced at $2,500 USD per pair. The announcement didn't arrive with a splashy press campaign or a coordinated global launch event. There was no full-page spread in a glossy magazine. Instead, it landed the way most genuinely interesting amplifier developments do these days — through a small, focused company talking directly to an audience that already cares deeply about what's happening at the bleeding edge of amplification technology. And that, more than any single specification, is the real story here.

Because if you've been paying attention to where gallium nitride audio technology is actually advancing, the pattern has been consistent for years: it's not the Sonys, the Denons, or even the boutique European giants who are pushing GaN FET amplification into genuinely new territory. It's tight, direct-to-consumer operations like Orchard Audio, run by engineers who are obsessively focused on a single technology and are willing to iterate quickly, licence their work and explain their decisions in real terms to real buyers. That's worth unpacking in some depth, because for Australian enthusiasts thinking seriously about their next amplifier purchase, the implications are significant.

What the Starkrimson 25 Mono actually represents

The Starkrimson 25 Mono Premium is described by Orchard Audio as being extremely close to a Class A/B amplifier in warmth and soundstage — which, if you understand the historical baggage around Class-D and its various derivatives, is a genuinely meaningful claim. For decades, the knock on switching amplifiers was that they traded the organic character of linear designs for efficiency and thermal manageability. Early Class-D implementations often had a sterile, slightly mechanical quality that made them fine for subwoofers and PA applications but difficult to love in a critical two-channel context.

GaN — gallium nitride — changes the switching equation materially. The core advantage is speed. Where a conventional silicon-based Class-D amplifier might switch at somewhere in the range of 300 to 400 kHz, Orchard Audio's GaN designs operate at around 800 kHz — roughly two to three times faster. This matters because the faster the switching frequency, the easier it becomes to filter out the high-frequency artefacts that switching amplifiers inherently produce, and the wider the usable audio bandwidth becomes before that filtering starts to interact with the audible range. The practical consequence is an amplifier that can approach the transient precision and tonal density of a good Class A/B design while retaining the efficiency and relatively compact thermal profile that makes GaN architectures so attractive in the first place.

The Starkrimson 25 Mono sits below the flagship Starkrimson Ultra line, which delivers a formidable 500 watts per channel and represents the top of Orchard's current range. The 25 Mono is clearly aimed at a different listener — one who doesn't necessarily need cinema-scale dynamics but wants the tonal qualities that GaN switching at these frequencies makes possible, in a monoblock form factor that allows complete channel separation and flexible placement around a listening room. At $2,500 USD per pair, it's positioned as an accessible entry point into what Orchard considers the genuine performance tier of their lineup.

Florida 2026 and the broader Orchard ecosystem

A few weeks after the Starkrimson 25 launch, Orchard Audio appeared at the Florida 2026 audio show, which Sound Technology covered in early March. The showing gave the wider audiophile community a clearer picture of where the company is heading as a full ecosystem rather than just an amplifier manufacturer.

Front and centre was the Starkrimson Stereo Ultra DMC 2.5, priced at $4,250 USD. This is the stereo flagship — a single-chassis design that inherits the core GaN switching architecture and represents Orchard's thinking about how to deliver that 800 kHz switching performance in a configuration that suits listeners who prefer a single box over a monoblock pair. The DMC designation refers to the integrated digital motor control architecture that Orchard has developed, which is a fairly significant engineering commitment — it's not simply a matter of dropping GaN FETs into an existing circuit topology.

Alongside the amplifiers, Orchard showed the PecanPi, their streamer, DAC, and preamplifier combination, priced at $1,500 USD. This is where my particular interest sits. The PecanPi represents Orchard's answer to a question that a lot of serious listeners are asking: if you're going to build a system around a specialist GaN amplifier, why not have the source and preamplifier stages designed by the same team, to the same philosophy? It's a compelling argument. The interaction between a digital-to-analogue converter, a preamplifier stage, and the input sensitivity of a downstream power amplifier — what we'd call the system's overall gain structure — has an enormous influence on the final sound. Having all of that designed by one engineer or team, with matching impedances and consistent voicing goals, removes a layer of uncertainty that plagues mix-and-match system building.

For Australian buyers, the PecanPi as an all-in-one source unit is particularly interesting because it places Orchard in direct competition with products like the Cambridge Audio CXN100 (check price) and other networked source components, while offering the advantage of being designed expressly to work in an Orchard-centric system. Whether it can match those established products on raw streaming feature sets and app experience is a question worth asking, but the integration argument is real.

Leo Ayzenshtat's influence beyond Orchard's own products

One of the more under-discussed aspects of Orchard Audio's position in the GaN amplification landscape is the fact that founder Leo Ayzenshtat has licensed his GaN FET technology to other manufacturers. This is significant. It means that Orchard's engineering work isn't just influencing the products that bear the Starkrimson name — it's quietly shaping the broader field. When a small specialist operation's intellectual property ends up powering products from other brands, it tells you something important about the quality and novelty of what that operation has actually developed.

The licensing model also speaks to Ayzenshtat's approach to the technology. Rather than treating GaN amplification as a proprietary black box to be defended against all comers, he appears to view it as an area where wider adoption ultimately benefits the field — including, presumably, Orchard's own reputation as the originating source of the technology. It's an approach more commonly seen in the software world than in high-end audio hardware, and it's one reason why Orchard occupies a somewhat unusual position: simultaneously a boutique manufacturer of finished products and a technology developer whose work has commercial reach beyond their own catalogue.

This is worth bearing in mind when you encounter GaN amplification claims from other manufacturers. Some of that technology traces a direct line back to Orchard's work. The question worth asking is whether those derivative implementations operate at comparable switching frequencies and apply the same rigour to filtering and output stage design — or whether they're simply borrowing the GaN branding while implementing the underlying technology more conservatively.

Why the big brands haven't caught up

It's worth spending a moment on why the major Japanese, European and American consumer electronics companies haven't simply absorbed this technology and deployed it at scale. The honest answer is a combination of risk aversion, supply chain complexity, and organisational inertia.

GaN FETs — the actual semiconductor devices that make these switching speeds possible — are sourced from a relatively small number of suppliers, and designing with them reliably at audio frequencies requires specific expertise that isn't freely available off the shelf. Larger companies also tend to design amplifier stages that need to work across wide ranges of load impedance and sensitivity to satisfy broad product line requirements, which pushes them toward more conservative, well-characterised topologies. There's also the not-insignificant matter of certification and compliance testing — a process that's expensive and time-consuming, and that discourages the kind of iterative revision that Orchard can undertake as a direct-to-consumer company.

The result is a situation that's genuinely unusual in consumer electronics: the most technically advanced implementations of a core technology are coming from the smallest participants in the market. If you're shopping for a streaming amplifier and looking at the mainstream options, it's worth understanding this context. Products like the Marantz Model 40n (check price) represent excellent value and accomplished engineering in the conventional Class A/B idiom, but they're not where GaN switching amplification is being pushed forward.

The soundstage question and what it means for speaker matching

Orchard's claim that the Starkrimson 25 Mono delivers warmth and soundstage character close to Class A/B is the kind of assertion that audiophiles are rightly sceptical of until they've heard it in a controlled context. The historical reputation of switching amplifiers for a somewhat compressed, less dimensional soundstage presentation is not entirely mythological — it relates to real artefacts in the amplifier's frequency and phase response that earlier switching designs struggled to fully suppress.

The higher switching frequency of Orchard's GaN designs is theoretically relevant here. With the switching frequency pushed further above the audio band, the output filter's crossover region moves away from the frequencies where complex musical information lives, and the filter's group delay characteristics become less problematic. The result should be an amplifier that preserves more of the phase relationships in the source signal — and phase coherence is fundamental to convincing stereo imaging and depth.

For Australian listeners thinking about speaker pairing, this has practical implications. A GaN amplifier of this type should theoretically be a good match for speakers that reward precise transient delivery and accurate phase — compact two-way designs with well-engineered crossovers, for instance. If you're running something like the KEF LS50 Meta (check price) or similarly resolving standmounts, the argument for an amplifier that can faithfully deliver that level of source resolution becomes compelling. You can read more about that class of loudspeaker in our guide to the best standmount speakers for serious listening.

Practical considerations for Australian buyers

Orchard Audio sells direct, which is both an advantage and a complication for Australian customers. The advantage is that the pricing — $2,500 USD for the Starkrimson 25 Mono pair, $4,250 for the Stereo Ultra DMC 2.5, $1,500 for the PecanPi — doesn't go through an Australian distributor margin, which in high-end audio can add anywhere from 30 to 60 per cent to the landed cost. The complication is that warranty service, if something goes wrong, requires international communication and potentially international shipping. For a power amplifier, that's a non-trivial risk to weigh up.

Currency conversion also matters. At current rates, the Starkrimson 25 Mono pair lands somewhere in the $3,800 to $4,200 AUD range before GST and shipping, depending on how the US dollar moves. That puts it in genuine competition with local-distribution products from brands like Rotel, NAD's higher-end offerings and select European Class A/B designs. The value proposition is real, but it requires doing the currency arithmetic carefully and factoring in the direct-purchase risks honestly.

For those who want to understand where a GaN amplifier fits within a complete digital system — particularly if you're building around a streamer and DAC front end — our guide to the best DACs and network streamers provides a useful framework for thinking about source quality and how it interacts with downstream amplification.

The broader significance: specialists as proving grounds

The Starkrimson 25 Mono launch, taken alongside the Florida 2026 showing and Ayzenshtat's ongoing licensing activity, confirms something that's been true of GaN audio amplification for the better part of a decade now: the technology's genuine advances happen at the fringes of the industry, not at its centre. Small, direct-to-consumer operations with the freedom to iterate, the technical focus to push switching frequencies into genuinely new territory, and the motivation to explain their engineering decisions to buyers who are paying close attention — these are the companies moving the field.

That's not a criticism of the larger manufacturers. It's simply an accurate description of how technologically disruptive products tend to develop before mainstream adoption catches up. And for serious Australian audio enthusiasts who are willing to engage with the direct-purchase model and do their homework on system matching, it represents an opportunity to access amplification performance that the major brands simply aren't offering at comparable price points right now.

Watch this space carefully. The pace at which Orchard is expanding its lineup — from monoblocks to stereo flagships to integrated source components — suggests a company that is building deliberately toward a complete system proposition, not just iterating on a single product. Whether that trajectory continues, and whether the technology licensing expands further, will tell us a lot about the next chapter of GaN amplification in high-end audio.