Still reading through the review, but 1080p consistency is all over the place and there are too many games that struggle to call this a true 1440p card.
I suspect driver immaturity is to blame for a lot of this, and potentially the Arc 750 might be the card to get if you play the longer game and wait for performance improvements via drivers.
Honestly, that's not bad for their first real dGPU. If it had released in 2020 as initially planned, Intel would have made a killing but this 18 month delay sure has hurt its prospects. Here's hoping they don't give up because the second gen is likely to be much better than this - they're likely learning
a lot from their first attempt's mistakes.
The thing that stops me from buying one is that although it's clear the performance is there in many titles, there are plenty of instances where it's outperformed by half-decade old cards that you can pick up on ebay for $100 now. Until those drivers are good enough that you can expect proper utilisation and performance in at least 90% of games, you can't look at its average performance, because you'll almost certainly have several games you want to play where it runs like a dog.
The average performance is somewhere between a 6600 and 6600XT but you'll only care about performance when there's not enough of it, at which point you'd have been much better off buying an RX 6600 for $224.99:
When a $224 card (it's definitely not hard to find sub-$250 RX6600 cards somewhere on any given day) is consistently providing a superior overall experience, it's really hard to justify paying $290 or $350 for an Arc card. You're effectively paying now for
maybe 6700XT performance 12 months down the line but if you're only worried about performance 12 months* from now, buy an RX7600 or RTX4060 instead, which will presumably run circles around Arc, the 3060, and the 6600-series.
* - a number I pulled out of my ass as a very rough estimate of how long I think it'll take Intel to deal with the worst-performing games through driver updates
