So you think its mostly a ONT side issue then.
The order books for the likes of BT, buying equipment at the scale they do would also likely negotiate a pretty hefty discount.
At the exchange side, if they already have XGS-PON/GPON combo OLTs, then there's nothing to be saved by getting rid of GPON.
Bingo

and if it is GPON only, then maintenance swapout likely isn't a huge undertaking. (I seem to remember Adtran/Nokia at least saying the upgrade process is "easy")
Also at the exchange side - specification for GPON being older doesn't mean the kit is old. If anything it means at the time of planning they would have likely ran into less issues, or at least less issues that the vendor didn't already know about.
An example - look at VDSL rollout, the Huawei modems that they had multiple revisions for - would have been so ridiculously cheap in comparison to alternatives, so even with multiple replacements of the BTOR Modem - it's likely cost wasn't an issue. (Especially as they're collecting the old dead modem, they may have additional deals with the vendor to replace dead kit - e.g. cradle to grave style hardware lifecycle)
In depth report - two of the UK's biggest fibre wholesalers have dampened expectations of imminent 25G and 50G PON upgrades, despite prep-work. Virgin Media O2-backed nexfibre sees no "imminent" case for commercial launch. Openreach also wants greater energy efficiency and to maximise older-gen...
www.telcotitans.com
This quote gives a giggle, as it reminds me of BT asset sweat.
I mean this is just good business sense, If I bought something that works for 95%+ of my customer base, am I going to replace it when it's only just started making me money ? There's no look for them on 25/50gig because they don't have the customer demand (from their point of view). It also would come with backhaul increases (potentially ££££ depending on where the bottleneck is) to support a handful of customers to max out their lines. It's not really a sensible strategy.
You have to remember we're a small section of the general population. Most people do not care, it just needs to be affordable, and work enough for things like Netflix.
Altnets pushing to newer/faster tiers will pick up those handfuls of customers where they have footprint. Once they've "made the market" the big players will be able to make some hardware swap outs and compete. The slow expensive bit is the initial deployment of fibre in the ground. Once that's done - It doesn't really matter what happens at each end of it.
Their take up seems way higher than the others.
Lots of ISPs can and do take BTOR FTTP Services, so you have big companies vying for those customers, vs a smaller altnet with a much more limited amount of smaller ISPs, with much smaller marketing budgets. Also people may not "trust" a new entrant vs going for someone like BT, Sky etc. Sky also have the addition of bundling with other services like TV. When doing comparisons if someone has Sky TV already, adding phone+broadband is usually a cost saver in comparison (or at least super competitive with an altnet)