at this point just focus on OEM. DIY market? come back here when Arc reaching at least into the third generation.
Considering Intel's been known for their mediocre integrated graphics for so long, they likely want a wider release to get tech enthusiasts talking about their hardware in an attempt to turn that viewpoint around. Limiting the hardware to OEM systems probably isn't the best way to achieve that. By selling cards on the DIY market, they can provide clear price points for them, and with Intel's typically competitive pricing when entering a new market, there's a good chance that the feedback will be positive, assuming the cards don't have any major drawbacks. With an OEM-only release, the perceived value of the cards would be more vague.
While that's kinda true, I'd rather they release them when they're ready with decent support rather than release them half-baked. Whether or not they do so I guess we'll find out in August.
Yeah, it's likely better for them to focus on giving a better first impression than being "on time" with a product that might still be rough around the edges. If they're planning on being a major player in the market in the long-term, then something like missing the back-to-school shopping season hardly matters compared to making sure everything is as polished as possible to avoid negative feedback about unfinished features and driver issues, things people will remember hearing about long after the initial launch. This also gives them the chance to have better supply, and potentially a wider selection of cards available at launch.
Arc is very quickly becoming i740 all over again. Their first attempt at discrete gpus was late and underperformed. The same thing is happening with Arc in light of Lovelace and RDNA3 launches looming in the fall.
The cards will underperform compared to what? As long as they offer good performance per dollar within their price ranges they have the potential to be successful. To the vast majority of the market, it doesn't matter all that much if they have anything to compete with the highest-end thousand-plus dollar cards from their competitors. And at least currently, most cards on the market "underperform" compared to where they should be at their price levels. A 3050 still costs as much or more than the 2060 did when it came out over three years ago, and isn't as fast, so it underperforms. A 3060 costs roughly as much as the 2070 SUPER did nearly three years ago, while not being as fast, so it underforms. Even the best current "sale" prices for Nvidia's lineup are still close to 50% above their announced launch prices, so those cards are all arguably a poor value right now. I suspect the situation may be better in a few months, but I don't think Intel should have too much trouble competing with those cards. They are likely willing to sell them at a relatively low cost to help them make inroads into the market, and it will probably still be months later before new competition arrives on the market.
Facebook
Twitter
Instagram