Only the reususability part failed in the last test, the ascent phase was successful. Landing them is by far the hardest task.
Starship didn't achieve orbit, and the wonky 'kinda-almost' orbit it did achieve was massively out of control, and it was out of control even before the descent stage - it failed to re-enter with the heat shield facing the right way - it didn't even get the chance to fail at a soft landing. I'd also suggest that failure of the stages to land doesn't
directly affect the mission, failure to achieve stable orbit to dock with the orbital re-fueller is massively mission critical, and it has to do that about 10-20 times perfectly because of the ridiculous mission profile.
There's also little evidence the internal propellant transfer worked, and the unrelated-to-Artemis Pez dispenser cargo door appeared to fail once opened too, on top of repeatedly substandard telemetry performance that persists despite SpaceX having more than 5,000 of their own satellites in orbit.
It does look awesome on lift-off, I'll give it that.
NASA can pretty much scrap Artemis, it's not going to happen anytime soon with SpaceX absolutely lagging.
It was conceptually unproven from the outset and SpaceX winning the contract at all appears to be the result of shenanigans by somebody that helped SpaceX fudge their submission, before awarding them billions, and then quitting NASA to go and work for them (hence Bezos suing NASA).
Thing is, this is all stuff that probably is worth doing. NASA should have taken one single objective look at Musk's track record and put him on a 10 year results based grant for developing Mars transit technology. If between spanking the ketamine and doing Putin favours Musk manages to make any actual progress with the technology that will
actually be required for any Human mission to Mars, all the better.