Intel's Raja Koduri Refutes Rumors About Company Cancelling Arc Graphics

[Dragokar]

No, it won't.

You are pretty sure about your guessing, but that is fine for me.

 

[lexluthermiester]

Whatever you say, man.

Oh ,no, it's not what I say, it's called common sense and good business tactics, something many here seems to have no grasp of.

No, I don't, and neither does anybody else on the forum.

Ask me I care(hint, no).

Personal attacks and name calling only make you look unhinged

Oh? What personal attack?

and childish.

Well it would seem I'm in good company then, wouldn't YOU say so? Hmmmmm?

If you want to call people out for being ridiculous it's a good idea to make sure you don't look ridiculous first.

Irony, far out..

I'm surprised the mods continue to put up with your inflammatory behavior.

Unlike your opinions, I'm not attacking anyone personally. I'm attacking a specific mentality and school of thought. If you are taking that personally, YOU have the problem. See to that.

You are pretty sure about your guessing, but that is fine for me.

It's not a guess.

 

[95Viper]

Stay on topic.
Stop the back-and-forth personal bickering/insulting/arguing.

 

[Luminescent]

If Intel reads this, most people don't need the absolute best, a well priced midrange GPU that plays most games is good enough.

 

[GunShot]

Rumors are just that, rumors!

Aka, IOW, it seems just like a regular liberal/progressive average day, eh?!

 

[Bwaze]

That is not only a silly notion but makes no sense considering that Intel has already started ramping up stock production.

We have absolutely no way of knowing what's the production run.

We can buy Intel A380 for couple of months now. Only in China, though, and even there it's not widely available. And reviews aren't favourable.

Graphics card sales have plummeted. No cryptomaniacs, and gamers aren't jumping on slightly discounted cards from 2020, when the new generation is coming right now. Where they even are discounted, in Europe prices are still above MSRP mainly. And, of course, recession is looming.

And Intel is competing in low and mid 2020 range. Even if it will be available only in late 2022, early 2023. So, a range nobody's buying any more.

 

[ZoneDymo]

Completely.

You're missing some context. Opinions based on flawed perceptions of fact does not make for factual information. It's called spin and only knuckle-draggers and tree-swinging poo-flingers will buy it..

ok so for some reason you have the correct perception of facts? what makes you think that and can you support those claims?
Because a simple google supported the march claim where intel was on track with Optane development and was even talking about the next version sooo that checks out.
And intel cutting parts of their business was an article on this very website, you can probably find if if you just scroll down.

So go ahead then, make your case.

 

[lexluthermiester]

So go ahead then, make your case.

I needn't bother. Reality tends to explain itself. However, I can't help you with historical context. If you fail to understand, that's on you.

 

[LemmingOverlord]

I feel split between two grade-A types of bullshit.

1. On the one side, MLID and TweakTown saying they have insiders talking about late night meetings and cancelling "discrete" Arc (meaning no gaming cards, just DC)
2. On the other side, we have Raja, who (to me) is a notorious word-mincer, who once again is feeding us half-truths.

 

[R0H1T]

Well it's not like they haven't done this just in the past decade ~

1) They spent upwards of ~10 billion dollars on mobile SoC's & then bailed on with literally billions in losses.
2) SSD's yup they bailed on that as well.
3) Optane, which isn't SSD but still a major loss (for the industry)
5) G modem ~ remember that, me neither

For anyone thinking Intel wouldn't abandon a project after sinking billions in there either doesn't know their past or is just too ignorant to see the (possible) future!
I probably could find a few others but why bother

 

[Tom Yum]

Is it just me or is it slightly odd that Raja is using past tense rather than present tense in his statement: 'we are still in first gen and yes we had more obstacles than planned to overcome, but we persisted". They are still in first gen, but they persisted rather than persist, which either implies that all the issues are resolved (no evidence of that found) or they no longer persist, but did in the past. Just the way he worded his response seems like he is hedging, he could have easily flat out said 'no the rumours are false', but instead went straight to the motive behind them which is a rather politician thing to do when you are evading a question.

 

[R0H1T]

It's above his pay grade, he can't refute things which he doesn't have 100% control over. It's probably a hot topic Inside Intel™ right now!

 

[aQi]

Nope Nope Nope

If i were intel, i would never quit at this stage.

 

[Dimitriman]

I do believe they may cancel discrete graphics if they feel it won't make money.

Alchemist can't beat a 3070/6700XT, and now will have to deal with 2nd hand market as well. Max sales price $350
Battlemage could double the performance and it still will not beat mid range 7600XT/4060 next gen cards. Max sales price will again be $350
Celestial gets to compete with 8000 series and 5000 series from AMD and Nvidia respectively. I don't think the scenario will change.

Intel completely missed the target by a good 2 years and now will be catching up to the mid range for the next 6 years of product cycles. It does not look good from an investor perspective, and the investors and shareholders are the only people that matter in the end.

This doesn't mean Arc is gone completely, I do think Intel will still find a market in notebooks.

 

[AusWolf]

You’re right. They won’t pull out after spending millions. They will pull out after spending billions! Just look at Itanium. But at least Itanium had a good run for awhile. Not so sure about these GPUs.

Arc Alchemist is basically ready. It would be a mistake not to release it and earn back at least a fraction of their investments.

 

[GunShot]

I feel split between two grade-A types of bullshit.

1. On the one side, MLID and TweakTown saying they have insiders talking about late night meetings and cancelling "discrete" Arc (meaning no gaming cards, just DC)
2. On the other side, we have Raja, who (to me) is a notorious word-mincer, who once again is feeding us half-truths.

You know, this sad practice that some writers try to use their platform to destroy a company e.g. the ill writers over there at WCCFTech (Chris, Nathan, etc.), that they all HATES Acti/Blizz so much that they've TOTALLY lost all integrity for its site and they have been pushing a smear campaign against Acti/Blizz since early 1H21 on EVERY Acti/Blizz headline.

I mean... I'm definitely no Acti/Blizz fan but what these fellas do over there at WCCF, etc., and there are even worst publishers out there that WCCF just look over and/or PRAISE that pubs that has the same rumours/scandals as AB, just shows us that, it is NOT about what is "right & balanced" but... what it is about is... THEM and THEIR AGENDA!

 

[SSGBryan]

Exactly. Why would anyone in their right mind cancel a multimillion USD project before fruition?

Edit: I love the picture with the loosely connected power plugs!

I see the sunk cost fallacy is alive and kicking.

 

[zlobby]

I see the sunk cost fallacy is alive and kicking.

At some point it would be cheaper to just let it go and move on. Then again intel have money to burn for office building heating, so not sure what are going on over there.

 

[SSGBryan]

At some point it would be cheaper to just let it go and move on. Then again intel have money to burn for office building heating, so not sure what are going on over there.

That is probably what is going to happen - ARC will move into the data center for machine learning, and that will be that.

From reading the financials over the past few years, intel is not doing as well as they should be, hence the exiting of money losing projects.

 

[trsttte]

As they say, die with the lie...

I'm not into the "it will be cancelled" bandwagon but it's hard to deny things aren't going well and with the amount of times this is being mentioned it just seems more and more likely for there to be some truth to it - latest one i think is that consumer discrete will be dumped and only integrated and server continue. That seems dumb but might be the way forward for at least a couple generations until everything (like cpu and manufacturing) gets their footing.

 

[Minus Infinity]

I worked for a company that also categorically denied it was shutting down and said such rumours were a disgrace. 1 month later the company shutdown permanently and the last 40 surviving employees from a once 300 workforce were given their marching orders. These denials are a dime a dozen and 90% of the time they are a sure sign Alchemist in in deep crap. They might as well try to get Battlemage out long before RDNA4 and Lovelace's successor ship.

 

[zlobby]

I worked for a company that also categorically denied it was shutting down and said such rumours were a disgrace. 1 month later the company shutdown permanently and the last 40 surviving employees from a once 300 workforce were given their marching orders. These denials are a dime a dozen and 90% of the time they are a sure sign Alchemist in in deep crap. They might as well try to get Battlemage out long before RDNA4 and Lovelace's successor ship.

Logic is sound. IF they had something to brag about, they would have bragged about it yet. We've seen far less significant things blown out of proportion by marketing teams.

 

[AusWolf]

I worked for a company that also categorically denied it was shutting down and said such rumours were a disgrace. 1 month later the company shutdown permanently and the last 40 surviving employees from a once 300 workforce were given their marching orders. These denials are a dime a dozen and 90% of the time they are a sure sign Alchemist in in deep crap. They might as well try to get Battlemage out long before RDNA4 and Lovelace's successor ship.

I see the logic in what you're saying. Sometimes company leaders want everybody to stay calm on board the sinking ship. I'm still on a "wait and see" approach.

Alchemist is pretty much complete, and the development of Battlemage is already underway. Intel (as well as AMD and Nvidia) plan for multiple generations ahead. Shutting down Arc would mean shutting down the work of multiple R&D teams. After so much money spent, I don't think it'd be worth it. People always come up with the Itanium example, but let's not forget that it involved finished products. Seeing that your investment isn't returning enough profit, and cancelling it before seeing if it actually returns any are two different things. One is a nunsuccessful business venture with net loss, the other is just money down the toilet.

 

[thestryker6]

Intel very much doesn't need to deliver high end performance, just a good value. Current "budget" prices for nvidia are absurd still and there won't be budget 4xxx any time soon. AMD is doing better here, but I think that's a place Intel could very well compete in since while AMD's drivers are pretty solid now they still hold that stigma so people may be willing to take a chance on Intel if the price is right. I think the biggest problem facing the first generation is actually the required ReBar support as this kills Arc as an inexpensive upgrade for older systems.

It's entirely possible that Intel will kill off the consumer GPU division, but given that they're going to be making future architectures for HPC and IGP it doesn't really make a lot of sense unless they're losing money from specifically the desktop card market. I think it'll mostly come down to the driver situation and how much time/money the software side is going to cost them along the way.

For those citing Optane as precident that's a pretty weak one: they sold off their half of the only fab that made it, then Micron stopped making it and sold the fab entirely. Intel never spun up a fab for production and Gelsinger isn't a fan of memory type products (dumb chips) so the writing was on the wall it was only a matter of time. Now I very much think it's a huge loss for everyone since further R&D might have gotten prices down close enough to NAND for consumers and more businesses to justify the premium (even 2x the cost may have worked, but P5800x is about 6-7x the cost of similar sized enterprise NAND).

 

[wahdangun]

Exactly. Why would anyone in their right mind cancel a multimillion USD project before fruition?

Edit: I love the picture with the loosely connected power plugs!

Like larabe ? Or optane