By looking at which pools mine P2A spends, we can guess when a pool roughly upgraded their getblocktemplate node to Bitcoin Core version v28.
| pool | first spend | first creation | total P2A inputs | total P2A outputs |
|---|---|---|---|---|
| MaraPool | 867278 (2024-10-25) | 881521 (2025-01-30) | 51 | 40 |
| ViaBTC | 883626 (2025-02-13) | 883626 (2025-02-13) | 7 | 4 |
| Foundry | 896194 (2025-05-11) | 850457 (2024-07-03) | 9 | 10 |
| F2Pool | 898061 (2025-05-23) | 850307 (2024-07-02) | 6 | 16 |
| AntPool | 900676 (2025-06-10) | 867352 (2024-10-25) | 14 | 17 |
| SpiderPool | 907615 (2025-07-28) | 907610 (2025-07-28) | 17 | 13 |
| Braiins | 850000 (2024-06-29) | 0 | 2 | |
| SecPool | 896193 (2025-05-11) | 0 | 1 | |
| MiningSquared | 894977 (2025-05-03) | 0 | 1 |
As per https://bitcoincore.org/en/lifecycle/, v28 was released on 2024-10-02. It seems like MaraPool upgraded fairly quickly. ViaBTC upgraded sometime in early Feburary 2025, about four month after the release.
On 2025-04-14, v29 was released. In May and June, Foundry, F2Pool, and AntPool upgraded to v28. SpiderPool upgraded sometime in July. This matches what I’ve heard elsewhere: some pools are lagging behind on purpose.
Braiins, SecPool, MiningSquared don’t have much hashrate, so we don’t know if they would mine a P2A spent if they see one. Similar with the remaining pools not appearing here.
The same can be done with Ephemeral Dust to see when the upgraded to v29 as described in https://github.com/0xB10C/mainnet-observer/issues/101.