Artemis (itself delayed by Artemis I preperation delays, but mostly the pre-Artemis I issues) just isn't a complete landing program in itself, so relies on SpaceX succeeding in creating its bit for the mission (or, I'd suggest as a very stop-gappy contingency, bringing Blue Origin in to replace them). And certainly SpaceX has nailed orbital insertion of cargo, booster-landing, manned-to-orbit(-and-safely-back) and sent a Tesla far
beyond the Moon, so it wasn't exactly a complete shot in the dark that they can engineer their piece of the puzzle.
But with
Spaceship edit: Starship![4] not yet demonstrating itself as even getting close to the Moon, before then having to show as 'solved' the landing and then unsupported take-off again, it's probably getting finger-bitingly close to various reserve deadlines in the planners' office. (And BO hasn't even gone so far as to
fail to launch its own option, for this trip.)
Artemis 2 is as proven workable as it can be, given that it's equivalent to Apollo (7
?/)8, with maybe six or seven[1] successful tests up to that point (and one very notable ground-failure), to put this "new race" into perspective with the old one.
Artemis 3 is planned to be (part of) the Apollo 11-equivalent mission, conflating three stages of Apollo (with 9=undocking/redocking tests in LEO and 10=all-but-landing rehearsal). Obviously SpaceX, once it proves it can launch Starship, has its own side-program of proving a Moon-tuned Starship can do the landing (automated, as it will probably mostly be anyway), take-off (likewise) and any necessary docking tests (but with Dragon/Dragon-Crewed and the ISS already in its back pocket on that issue), so it may not need (or be seen as cost-efficient practical) to still take the original project's exact steps.
Which might mean one or two mangled (or just un-relaunchable) Lunar Starships get left on the Moon in the gaps between Artemis missions, in trying to perfect (and man-rate) the landing phase in time for 3's crew to take the plunge. But it
does seems awfully close to next year, already, watching it all unfold from out here on the bleachers... (Apollo had its delays, and not just because of Grissom, et al, which shoved things back a year or two at times. And that was Space Race time with Space Race pressures.)
Apart from maybe something like Biden wanting to be sure that he was in the Whitehouse to do the famous phonecall (which is a thing his predecessor would have wanted even more[3]), I think the same haste isn't in play. It'd probably be nice for them to get back there before China (or anyone else!) reaches it for the first time, but "hope for 2026, plan for <2030" is a viable amount of slack if the budget can cover the delays at least as much as the alternative of throwing money at things to get them to happen sooner.
[1] Depending on how you count. But before Apollo 1 were three main tests for the hardware (but one not including the Apollo CSM), and after it were unmanned Apollo 4, 5*, 6 and manned 7*. Those marked * were not on Saturn Vs, but 5 was the first with the LM loaded, and 7 was the first crewed CSM flight (so, Artemis 2-like), so probably these definitely[2] count.
[2] By all other relevent measures anyway, even if they hadn't been given full Apollo numbers (even 'unused' ones). Like the various pre-1 tests weren't that I would count, as well as technical tests that I wouldn't.
[3] I'm tempted to go back into the archives and see if it was buttoned down as a schedule within his first term as something that would happen before the end of his presumed second term, so maybe pressure/hints about how nice it would be happened.
[4] I have been known to get SpaceX and Scaled Composites names mixed up! And there was a
SpaceshipTwo launch the other day, so maybe that will have caused my error there. ;P