Panasonic's September torque update and the battery passport deadline in February point to the same thing: the two most expensive parts of an e-bike have become software products. Yet I think most supplier negotiations are still mostly about hardware.

120 Nm from an update

In early September, Panasonic Industry announced that it is raising the maximum torque of its GXM mid-drive to 120 Nm across the portfolio with a software update, no hardware change. Same motor, same housing, different product.

To me this means one thing: peak torque is no longer a property of the hardware. It is a parameter the supplier can grant or hold back. So the key question for an OEM is not how many Nm the motor delivers today, but what it gets out of that motor over the model's full life cycle.

This is also why the role of connected e-bike apps seems to be growing, especially apps that can deliver over-the-air (OTA) updates. The app is becoming the channel through which a bike gains features and fixes after it has left the shop. Whoever controls that channel controls a growing share of the product's value.

Supplier talks, however, still revolve around unit price, torque and lead times. Update policy shows up in an annex, if at all. If I were choosing a drive partner today, these are the questions I would ask first:

  1. How many years of updates does a motor generation get, and is that in the contract?
  2. Who performs and pays for updates: the dealer, the distributor, or the rider through an app (OTA)?
  3. Who owns the app relationship with the rider: the bike brand or the drive supplier?
  4. Can the supplier unlock paid features directly, bypassing the brand?
  5. How is the system protected against tampering? The ECI's late-September manifesto already expects manufacturers to use cryptographically verified updates with anti-rollback protection.

Battery passport: the format is ready, the key is missing

Under Article 77 of the EU Batteries Regulation (EU) 2023/1542, batteries for light means of transport (LMT), e-bike batteries included, need a digital passport from 18 February 2027. That is roughly four and a half months away.

The formal side is moving: the implementing regulation on the passport registry has been in force since August, and the harmonised standards were cited in mid-July. But the implementing act under Article 77(9), which defines who may access the non-public data, missed its 18 August deadline and is now expected in Q4 2026. It will decide what a service workshop, a refurbisher and a recycler can see, and on what terms they may download or share that data.

My view: mid-sized LEV brands will get the data side of the passport ready, because plenty of software vendors sell it as a package. The access side they will not, even though it decides whether their service network, cell-level repair and second-life batteries actually work. Anyone who hard-codes access tiers into their system now will have to rebuild it once the rules appear. Anyone who designs them as configurable roles will only need to change settings.

The common thread

In both cases there is a key, the data itself, that decides what the product is worth in its second and third year: what the motor can do after an update, and who can work with a used battery. Today that key usually sits with the supplier, and contracts rarely say how much of it the brand, the distributor and the service partner get.

And let's not forget the end user, the rider. They need to know what parameters their bike has at every level, and be able to prepare for an upgrade. For most riders, that knowledge will arrive through the app, which is one more reason the app belongs in the contract, not just in the product roadmap.

Anyone who doesn't lock this in with their drive and battery partners now will be negotiating after the fact in 2027, from a weaker position. My suggestion is simple: in the next supplier round, put the update policy, OTA and app access, and data access on the first page of the term sheet, not in the annex.

How are you negotiating this today? I think few are, but tell me if I'm wrong.

Sources