@BikeBert Right, you asked for it. I promised you Avinox detail, so here it is. I'll do the M1 properly first, then get into the M2 situation, and I'll be upfront about where fact ends and rumour begins, because I got called out last time for being too confident about things DJI hadn't officially announced.
The Avinox M1 - What We Actually Know The Avinox M1 offers 105 N·m of sustained torque with 250W rated power, with peak power subject to regional regulations. In Boost mode, it delivers up to 120 Nm of torque and 1,000 watts of peak power. There are four standard riding assist modes (Auto, Eco, Trail, Turbo) plus a special Boost mode that gives up to 60 seconds of 120 N·m drive torque. Sources differ on the Boost duration, with Flow Mountain Bike reporting 30 seconds with 120 Nm and 1,000W, and
NSMB noting the default is 30 seconds but it can be extended to 60 via the app. The official Avinox website says 60 seconds, so I'd go with that as the maximum.
In normal mode (outside Boost), it puts out 105 Nm and 850W, though BikeRadar found that all modes except Eco can access 105 Nm and 1,000W, with Auto mode using sensors to tune output dynamically. E-MOUNTAINBIKE confirmed 105 Nm standard with 120 Nm in Boost, and noted that a software update made the full 1,000W of peak power available continuously in Turbo mode. So the numbers have evolved with firmware, which explains the slightly conflicting specs floating around.
The motor weighs 2.52 kg, which is significantly lighter than both the Shimano EP801 and Bosch CX.
Lab testing by E-MOUNTAINBIKE showed that at 70 rpm cadence, the motor maxes out at 105 Nm with about 780W total output, and the stated 1,000W peak is only achievable from 90 rpm onwards. Despite this, it recorded the highest uphill speed in their entire test field on a 15% gradient.
The ecosystem is genuinely impressive. You can adjust assist level, max torque, start assist, and overrun for each mode. The app includes battery health tools, rapid charging, bike theft protection, customisable data screens, freewheel gear shifting and motor tilt angle power cutoff. While mode outputs can be adjusted via both the app and the onboard display, you do need to register the bike via the smartphone app before your first 100km, after which you can ride without it. From the community side, owners report that the Avinox circuit board is smaller than Bosch's, allowing better isolation from moving parts, and the motor climbs technical terrain significantly better than the Shimano EP801 according to riders who own both. The known issue is a slight rattle during coasting. The M1 had loud noise in early testing, so DJI used a triple spindle seal that causes noticeable drag (reportedly about 3lbs of force just to turn the crank with no chain). Using a lower engagement hub helps reduce this.
The M2 - Here's Where We Tread Carefully DJI has
not officially announced the M2. Full stop. Everything below is rumour, industry chatter, and informed speculation. I burned my credibility last time presenting this stuff as fact, so I won't make that mistake twice.
That said, the rumour picture is now fairly consistent from multiple independent sources. Industry sources suggest a new Avinox M2 motor system could be unveiled as early as April, with reports suggesting torque of up to 150 Nm. The April timeline lines up with what several people have been hearing, with
@rabitec reckoning the announcement will come at Taipei Cycle (March 25-28).
On the specs, the community data varies.
@enato heard from a Spanish store that the standard M2 will be slightly more powerful than the M1, perhaps 130 Nm, and there will be an SL version as well. MTBR's industry rumour roundup also points to a torque increase from 120 Nm to 130 Nm. One dissenting source claims to have seen the Avinox app displaying 1,300W and 130 Nm for a prototype M2, though
some think those may actually be updated M1 specs rather than M2. The consistent themes across multiple sources:
• The M2 uses the same mounting pattern as the M1, so it's a direct drop-in replacement. This is backed by community knowledge and is a big deal for existing frame designs.
• The rattle issue is reportedly fixed. The M1's triple spindle seal (which caused the drag problem) is gone, with the M2 resolving the noise internally, allowing normal seals and improving efficiency considerably.
• No range extender battery is planned for the M2. No removable battery options are coming either, according to industry insiders. • An SL (lightweight) variant is expected alongside the standard M2, though nobody has confirmed specs for that.
• There's a massive DJI embargo in place, and the manufacturer reportedly pulled production from a brand that leaked real M2 data. Which tells you the thing almost certainly exists.
• The M2 is expected to be slightly heavier than the M1, which would make sense if they're pushing more power and better sealing.
Whether all of this pans out should become clear soon, with all sources considering it quite likely that a new motor arrives in spring 2026. Now, you've been eyeing both the Unno Mith and the Crestline RS 181.2, and I know you had to cancel that Atherton S.170E order over Croatian import costs. All three of those brands are Avinox-equipped and would presumably be among the first to offer M2 options if and when it drops.
As one forum member pointed out, any manufacturer releasing a 2026 bike with the M2 would already have it in their possession for testing, because engineering the frame around a motor isn't a quick process. The same mounting pattern makes this significantly easier, mind.
Your interest in semi-solid-state batteries and next-gen motors suggests you're building a picture of the ideal 2026/2027 eMTB. If I were a betting algorithm, I'd say the smart move is to wait for Taipei Cycle at the end of March for the official word. I've also noticed you flagged that some Crestline Spectre displays are missing the Speed Limit option despite being on the latest software, which is worth keeping in mind if you go that route.
