ZMK is becoming the wireless default for premium prebuilts
A five-board flagship lineup running ZMK isn't a niche move; it's a market signal that an open-source firmware that lived in DIY-wireless territory for half a decade has crossed into mainstream-prebuilt anchor status. The vendor side has finally caught up with where the cohort already was.
For most of the last five years, the honest read on ZMK was that it was the firmware DIY-wireless people used and everyone else used QMK/VIA. The Glove80 shipped on it. The Voyager shipped on it. Corne wireless variants and the Sweep family ran it because someone with the patience to flash a devicetree file wanted Bluetooth and good battery life out of an open-source codebase, and ZMK was the codebase that had both. That read held cleanly through 2024. It started softening through 2025. At CES 2026, Keychron picked ZMK as the firmware platform for a five-board flagship line, and the read flipped. The firmware that served the DIY-wireless cohort for half a decade is now the firmware a Tier-1 vendor anchors a premium prebuilt lineup on. The line has been crossed.
The signal: a five-board flagship line, not a single SKU
A single Tier-1 wireless board on ZMK could be filed as a vendor hedging a bet — interesting, possibly a one-off, not yet a signal. Five is structural. The Q Ultra series Keychron announced at CES 2026 spans Q1 through Q6: 75%, 65%, TKL, 96%, and full-size, the same size ladder the wired Q line has been iterating on for three years. Every one of them runs ZMK rather than the QMK/VIA stack the existing Q line ships with. The headline number in the press release is the 660-hour battery claim and the 8K polling ceiling, but those numbers are downstream of the firmware decision. ZMK is what makes the power profile plausible at 8K polling on Bluetooth in the first place. The vendor didn't pick the firmware to fit the spec sheet; the firmware is what allowed the spec sheet to exist.
The other signal worth holding in mind is which SKU landed in reviewer hands first. The Q1 HE 8K Marble — the Hall-effect variant that anchors the polling story — reached reviewers in April, ahead of the mechanical variants of the same chassis. That ordering is not accidental, and it puts the Q Ultra launch at the exact intersection of two trends thock has been tracking in parallel. The Hall-effect-into-mid-premium piece this week called the mid-premium tier the slot HE has settled into; the Q1 HE 8K Marble is the SKU where the HE-mainstream trend and the ZMK-mainstream trend collide on a single board. Two trend lines, one product, one firmware platform underneath both. That's the shape of a category transition that's actually moving rather than a marketing one.
Where the demand floor came from
The vendor side moves last. The cohort that drove ZMK adoption did so years before Keychron noticed, and the split/ergo grown-up piece covers that history at length. ZSA's Voyager has shipped on ZMK since 2022, anchoring the low-profile split slot with battery-life numbers that QMK's wireless story couldn't match at the time. MoErgo's Glove80 is the column-staggered, sculpted-well counterpart and runs on the same firmware platform. Foostan's Corne, the open-hardware split that more readers have built than any other DIY board in the last five years, has had wireless variants on ZMK for long enough that the configuration is documented in beginner guides rather than expert forums. None of those products got to flagship-prebuilt volume, but together they built a demand floor — a cohort of typists who already had ZMK boards on their desks, already understood the keymap-as-code workflow, and already trusted the firmware's power profile.
That floor is what made the Q Ultra decision financeable. A vendor investing in ZMK firmware engineering in 2026 is investing into an ecosystem that has documentation, working hardware, a configurator effort in flight, and an installed base that will identify a regression and file it within hours. The same vendor making the same decision in 2022 would have been investing into a project that hadn't yet reached that maturity. The cohort did the work of getting the firmware to the line; the vendors are now crossing it.
The strengths the vendor side is buying
Two things ZMK does well are exactly the two things a wireless prebuilt sells on. The first is the power profile. ZMK was architected around low-power Bluetooth from the first commit, and the engineering decisions that follow from that — event-driven scanning, aggressive sleep states, a Zephyr-native BLE stack — compound into battery numbers that QMK's wireless implementations have never matched at parity. Keychron's 660-hour claim is the flag the lineup is planted on, and even if the reviewer-measured number lands at half that figure, half of 660 is still a multiple of what a QMK wireless board would deliver under the same usage. The firmware decision is what permits that headline.
The second is Bluetooth maturity. The wireless DIY cohort spent five years shaking the protocol-level bugs out of ZMK's BLE stack on real hardware — multi-device profiles that switch without fighting the host, sleep/wake transitions that don't lose the first keystroke, keepalive intervals that survive a Linux laptop's WiFi card aggressively reclaiming the 2.4GHz band. None of that work shows up on a spec sheet. All of it shows up in the reviews that determine whether a wireless prebuilt is a flagship or a return. A vendor shipping a wireless flagship in 2026 is not in a position to bet against five years of that field testing.
The weaknesses they're inheriting
The honest case against ZMK as a mainstream prebuilt platform is the same case it's had for years, and the pattern named in the vendor-first customs piece applies here too — a vendor that takes a position on firmware inherits the firmware's reputation, and ZMK's reputation in two specific areas is still under construction. The first is the configurator. QMK has VIA. VIA is the reason a reader who has never touched a keymap-as-code workflow can remap caps lock to control on their second day with a Q3. ZMK's answer is ZMK Studio, which is in active development and showing real progress, but is not yet the drop-in equivalent VIA has been for QMK-stack boards. A Q Ultra owner who wants to swap a key in the first week needs a workflow that hides the devicetree file the historical ZMK setup exposes. If Keychron's fork doesn't ship that workflow at launch, the line will lose readers who came in on the wired Q line's VIA-flattened customization curve.
The second is feature parity on advanced macros and combos. QMK's combo and macro infrastructure has accumulated a decade of community-driven features — Tap Dance, Layer Lock, Mod-Tap nuance the elder statesmen of the bench have opinions about. ZMK has implementations of most of those concepts, but not always at the same maturity, and the documentation gap between "the feature exists" and "the feature is the obvious right call" is the gap the configurator question really lives inside. A reader migrating from a Q1 with a complex QMK keymap to a Q1 Ultra on ZMK should not be the one carrying the cost of that gap. The Q Ultra's reception in the first month of reviews will be downstream of how that migration feels.
What we're watching
Three threads. The first is the configurator question. ZMK Studio needs to ship a workflow that a wired-Q owner recognizes as a sibling of VIA, not a flat replacement that demands re-learning. The Q Ultra is the largest single product release ZMK Studio will have to support; whether it ships ready or lands a quarter behind the hardware is the most consequential firmware deliverable of the year. If it ships ready, the configurator-gravity case against ZMK softens substantially. If it doesn't, the gap becomes the review story rather than the launch story.
The second is which Tier-1 vendor follows Keychron. The Cherry MX2A retrospective traced how a mainstream-performance tier moves when the incumbent stops defending it; firmware platforms are subject to the same dynamic. QMK's position as the prebuilt default is structural — every existing Tier-1 wireless board in 2026 still ships on it — but structural defaults are stable until they're not. The question is whether the next flagship wireless announcement from a major prebuilt vendor lists QMK as the firmware platform or quietly adopts ZMK without making it the headline. The second pattern is the one to watch for. A firmware migration that stops being a story is the firmware migration that has actually won.
The third is feature parity on advanced macros. If the boutique custom side — the readers who run six-layer keymaps with combos and home-row mods — files real friction with ZMK at flagship hardware, the trend tightens around mid-premium prebuilts and softens at the enthusiast tier. If they don't, the firmware ceiling that limited ZMK adoption is essentially gone. The live Trends Tracker has the ZMK row sloping up and the wireless-firmware sub-thread sloping up alongside it; neither line has spiked, both are climbing on the cadence of a category settling into a default slot. That's the shape this piece is calling. The Q Ultra launch is the test case. The next twelve months tell us whether QMK keeps the wired prebuilt tier and cedes the wireless one, or whether the firmware question stops being a binary altogether.
Keep reading
- News
Keychron's Q Ultra ZMK series brings 660h battery to the prebuilt-premium tier
Keychron used CES 2026 to announce the Q Ultra line — Q1 through Q6 prebuilts running ZMK rather than QMK/VIA. The 660h battery claim is the headline; the firmware swap is the actual story.
thock4 min read - Trends
How Hall-effect quietly became the new mid-premium default
The polling-rate hype is loud and easy to wave off — 8K, 32K, numbers chasing each other up the spec sheet. The quieter story is that the magnetic-switch tier has stopped being a Wooting-shaped niche and has settled into the $100–$230 prebuilt slot the scene actually buys from. The volume has moved, not just the marketing.
thock7 min read - Trends
When customs became vendor-first
From 2018 through 2022 the typical custom shipped through one host vendor and a fan of regional proxies; by 2026 the baseline custom is a configurator on a vendor's own storefront. Boards moved. Keycap group buys mostly didn't.
thock5 min read - Trends
The split/ergo cohort grew up
Split and ergonomic boards spent a decade as a hobbyist sub-cohort orbiting a stagger-majority. By 2026 the Voyager line, the Glove80, and a steady trickle of r/mk transition posts have pushed split/ergo across a visibility threshold — still a minority, but now a reference category instead of a footnote.
thock5 min read