Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Nvidia feels the need to counter the pending launch of the Radeon RX 5600 XT with a marketing stunt
Last week, AMD announced the upcoming Radeon RX 5600 XT, a further cut down bin of the same Navi 10 die as the Radeon RX 5700 XT and Radeon RX 5700. AMD is claiming that it is faster than a GeForce GTX 1660 Ti, and will cost $280. Probably not coincidentally, that's exactly what a GeForce GTX 1660 Ti costs.
Nvidia's next card above the GTX 1660 Ti in their lineup is the GeForce RTX 2060. If you go to New Egg right now, the cheapest new GeForce RTX 2060 is $320 after promo code, or $335 without the promo code. The Radeon RX 5600 XT will presumably be slower than the RTX 2060, and also cheaper.
So what Nvidia has now done is to counter by launching one particular SKU of GeForce RTX 2060, the EVGA GeForce RTX 2060 KO, for $280. Or at least, it's listed on New Egg, but not yet available to buy. They're not slashing the price on the RTX 2060 more generally to $280. They're just offering unknown quantities of one particular SKU for $280, while all other RTX 2060s will remain much more expensive.
So why is that a marketing stunt? If you want an RTX 2060 for $280, why should you care if there are many SKUs available or just one? It's not like EVGA is some dodgy board partner that no one has ever heard of.
Let's go back in time to October 2010. At the time, AMD was way ahead of Nvidia in every gaming efficiency metric you can think of, plus most of the ones that you can't. Performance per dollar, performance per watt, performance per mm^2, or anything else you like. AMD's GPUs also had the advantage that they actually worked, while Nvidia was struggling with yields that were astonishingly awful--much worse than anything that either major vendor has struggled with at any time since. And AMD was pricing their products aggressively to claim market share, which put Nvidia in a tough spot. AMD launched the Radeon HD 6870 for $240, and the Radeon HD 6850 for $180, which really were the best gaming options at those prices. And because their Fermi architecture was so bad, Nvidia didn't have a good way to counter.
So what Nvidia did was to bin out the very best GF104 chips and give them to EVGA to make a new SKU, the EVGA GeForce GTX 460 FTW. It would be a very heavily overclocked GeForce GTX 460, with a clock speed of 850 MHz. For comparison, the next highest factory overclock of any GTX 460 was 810 MHz. That was a large enough overclock to be barely faster than the new Radeon HD 6870, which was not true of other factory overclocked GTX 460s. And Nvidia would offer it for the same $240 as the Radeon HD 6870. Nvidia seeded them to reviewers and demanded that the Radeon HD 6870 reviews use this special GTX 460 as the comparison point.
Good deal, right? Well, assuming that you didn't care at all about power consumption. Or reliability from such an enormous factory overclock.
Or availability. There were very few of the EVGA GeForce GTX 460 FTW ever made. The stock available for launch sold out that day. Over the course of the next month, the SKU would flicker in and out of stock as a few cards trickled in here and there and subsequently sold. Within a month, the SKU would be discontinued and disappear forever.
But tech review sites didn't go back and modify their Radeon HD 6870 reviews to note that the special GTX 460 barely existed and you couldn't actually buy it at $240 or any other price. On some sites, instead of the conclusion being that the Radeon HD 6870 was clearly the best option at its price point, they instead said that the special GTX 460 was better on a price/performance basis. And they left that conclusion in place for the period of more than a year that the Radeon HD 6870 was a sensible purchase and the EVGA GeForce GTX 460 FTW was not available at all.
So that was a marketing stunt. It sure looked like it on launch day, and it became abundantly clear that that's all it was in hindsight. But it was a marketing stunt that worked. It likely led some people who didn't read the reviews that carefully to see that an EVGA GeForce GTX 460 FTW was a little faster than a stock Radeon HD 6870, not find that particular SKU in stock, and figure that some other GeForce GTX 460 must also be equivalent to it.
For comparison, there were two different cards that Nvidia branded as GeForce GTX 460. One had a stock clock speed of 650 MHz, and the other of 675 MHz. Those are not at all close to 850 MHz, so a "normal" GTX 460 was much, much slower than the special version that Nvidia wanted in comparisons. Actually, there were three different desktop cards branded as GTX 460, but the third didn't launch until about a year later, so it wasn't relevant to this stunt.
Apparently that worked so well that Nvidia decided to do it again. They're not slashing prices on the GeForce RTX 2060 across the board. I'd have no problem with that. Instead, it's just a single SKU, the EVGA GeForce RTX 2060 KO. You can bet that Nvidia is pushing reviewers to include that particular SKU in their reviews of the Radeon RX 5600 XT. And it's a pretty good bet that the SKU will only exist in small quantities, and then quickly vanish once the stunt has served its purpose.
Nvidia's next card above the GTX 1660 Ti in their lineup is the GeForce RTX 2060. If you go to New Egg right now, the cheapest new GeForce RTX 2060 is $320 after promo code, or $335 without the promo code. The Radeon RX 5600 XT will presumably be slower than the RTX 2060, and also cheaper.
So what Nvidia has now done is to counter by launching one particular SKU of GeForce RTX 2060, the EVGA GeForce RTX 2060 KO, for $280. Or at least, it's listed on New Egg, but not yet available to buy. They're not slashing the price on the RTX 2060 more generally to $280. They're just offering unknown quantities of one particular SKU for $280, while all other RTX 2060s will remain much more expensive.
So why is that a marketing stunt? If you want an RTX 2060 for $280, why should you care if there are many SKUs available or just one? It's not like EVGA is some dodgy board partner that no one has ever heard of.
Let's go back in time to October 2010. At the time, AMD was way ahead of Nvidia in every gaming efficiency metric you can think of, plus most of the ones that you can't. Performance per dollar, performance per watt, performance per mm^2, or anything else you like. AMD's GPUs also had the advantage that they actually worked, while Nvidia was struggling with yields that were astonishingly awful--much worse than anything that either major vendor has struggled with at any time since. And AMD was pricing their products aggressively to claim market share, which put Nvidia in a tough spot. AMD launched the Radeon HD 6870 for $240, and the Radeon HD 6850 for $180, which really were the best gaming options at those prices. And because their Fermi architecture was so bad, Nvidia didn't have a good way to counter.
So what Nvidia did was to bin out the very best GF104 chips and give them to EVGA to make a new SKU, the EVGA GeForce GTX 460 FTW. It would be a very heavily overclocked GeForce GTX 460, with a clock speed of 850 MHz. For comparison, the next highest factory overclock of any GTX 460 was 810 MHz. That was a large enough overclock to be barely faster than the new Radeon HD 6870, which was not true of other factory overclocked GTX 460s. And Nvidia would offer it for the same $240 as the Radeon HD 6870. Nvidia seeded them to reviewers and demanded that the Radeon HD 6870 reviews use this special GTX 460 as the comparison point.
Good deal, right? Well, assuming that you didn't care at all about power consumption. Or reliability from such an enormous factory overclock.
Or availability. There were very few of the EVGA GeForce GTX 460 FTW ever made. The stock available for launch sold out that day. Over the course of the next month, the SKU would flicker in and out of stock as a few cards trickled in here and there and subsequently sold. Within a month, the SKU would be discontinued and disappear forever.
But tech review sites didn't go back and modify their Radeon HD 6870 reviews to note that the special GTX 460 barely existed and you couldn't actually buy it at $240 or any other price. On some sites, instead of the conclusion being that the Radeon HD 6870 was clearly the best option at its price point, they instead said that the special GTX 460 was better on a price/performance basis. And they left that conclusion in place for the period of more than a year that the Radeon HD 6870 was a sensible purchase and the EVGA GeForce GTX 460 FTW was not available at all.
So that was a marketing stunt. It sure looked like it on launch day, and it became abundantly clear that that's all it was in hindsight. But it was a marketing stunt that worked. It likely led some people who didn't read the reviews that carefully to see that an EVGA GeForce GTX 460 FTW was a little faster than a stock Radeon HD 6870, not find that particular SKU in stock, and figure that some other GeForce GTX 460 must also be equivalent to it.
For comparison, there were two different cards that Nvidia branded as GeForce GTX 460. One had a stock clock speed of 650 MHz, and the other of 675 MHz. Those are not at all close to 850 MHz, so a "normal" GTX 460 was much, much slower than the special version that Nvidia wanted in comparisons. Actually, there were three different desktop cards branded as GTX 460, but the third didn't launch until about a year later, so it wasn't relevant to this stunt.
Apparently that worked so well that Nvidia decided to do it again. They're not slashing prices on the GeForce RTX 2060 across the board. I'd have no problem with that. Instead, it's just a single SKU, the EVGA GeForce RTX 2060 KO. You can bet that Nvidia is pushing reviewers to include that particular SKU in their reviews of the Radeon RX 5600 XT. And it's a pretty good bet that the SKU will only exist in small quantities, and then quickly vanish once the stunt has served its purpose.
Comments
For high-end performance to be available a year and a half later for $150 just doesn't happen. Top end cards are too expensive to build to do that. That would be kind of like seeing a brand new GeForce RTX 2080 for $150 later this year, and available in large quantities. That's why nothing remotely like it has happened since then.
But there were special circumstances here. AMD had built too many of their Cypress die. That die was now obsolete, due to the Barts die of the Radeon HD 6850 and 6870 that had launched the previous October. So they had a bunch of dies that their board partners didn't want.
Or rather, didn't want at the normal price. AMD cut a deal with Sapphire to take their remaining inventory that they wanted to get rid of off their hands cheaply. Sapphire passed along the savings to end customers. For about a month, if you wanted the best $200 card money could buy, the answer was to save some money and buy a Sapphire Radeon HD 5850 for $150, or later $140. It was a high volume SKU, and Sapphire sent it out for reviews, just to say, it's a perfectly normal card, and the price isn't because it's defective.
So could the EVGA GeForce RTX 2060 KO be another case of doing that? It's plausible enough that I feel the need to mention it. But it's unlikely, as the normal Radeon HD 5850 and 5830 were discontinued, and AMD already had their replacements on the market. Nvidia doesn't have a replacement for the GeForce RTX 2060, and it doesn't look like one is coming anytime soon. Unless Nvidia is going to discontinue some parts and vacate significant market segments, the GeForce RTX 2060 should stick around for quite some time.
Thus, the EVGA GeForce RTX 2060 KO is almost surely a low-volume marketing stunt. What I find interesting is that Nvidia feels the need to do it. Nvidia's lineup doesn't otherwise look like it is in a bad spot today the way it did in 2010. When AMD announced the Radeon RX 5600 XT, I didn't think it looked like a particularly compelling product. It looked to me like AMD offering a reasonable option at a price point missing from their lineup, and a way to get rid of some Navi 10 dies that can't meet Radeon RX 5700 specs. In contrast, the Barts GPU of the Radeon HD 6870 and 6850 was an important GPU from its era.
So why did Nvidia feel the need to repeat their marketing stunt for this particular launch? Maybe we'll find out next week.
I don't think there is any stunt whatsoever, EVGA in particular has a long history of tweaking cards.
It's two SKUs of a card that released in 2018, now available for "pre-order" in 2020. And that doesn't strike you as unusual in the slightest?
If Nvidia were cutting the MSRP on the RTX 2060 to $280, they'd announce that and be done with it. Or maybe they wouldn't make a public announcement, but would just let the cheaper cards show up at retail.
But that's not what is happening here. It's just two SKUs from one board partner. That's why it screams "marketing stunt". If it were a bunch of SKUs dropping by $50, I'd call it a normal price cut.
I have no problem with Nvidia slashing prices across the board. If they change the MSRP of the GeForce GTX 1660 Ti to $240 and the GeForce RTX 2060 to $280 and all of the SKUs drop in price, then expecting reviewers to compare a Radeon RX 5600 XT to a GeForce RTX 2060 at the same price would be totally legitimate. My objection is to fake prices on fake SKUs that you can't actually buy, and trying to get reviewers to pretend that that's the real price when writing up their comparison.
Also, the cards are already sold out on New Egg, which is exactly what you'd expect for a low-volume discount. It's kind of like a Black Friday "doorbuster" sale. That item you wanted at a huge discount would have been a good deal if you could get it at that price, but you can't.
The world is going to the dogs, which is just how I planned it!
The world is going to the dogs, which is just how I planned it!
But if they move the RTX 2060 to $300 and that's it, that would also be a very weird thing to do. The GeForce GTX 1660 Ti is still $280, and not much faster than the GeForce GTX 1660 Super for $230.
I'd expect to see Nvidia do something with the GTX 1660 Ti. That could mean cutting its price to $250 or so, launching a GTX 1660 Ti Super and deprecating the non-super version, or maybe something else. Because as things are priced now, there's basically no reason to get a fully functional TU116-based GPU. With a die size of 284 mm^2, as compared to 445 mm^2 for the TU106 of the RTX 2060, you'd think that Nvidia would want to price the smaller die cheaper to push people in that direction unless they're willing to pay a considerable premium for the larger die.
These days, assuming no impact on memory and os, upgrading the motherboard and cpu might be cheaper. Given the price of some graphics cards buying new memory could be included.