Note: the following is a modified version of the transcript Jim used for his recent video on Nvidia, FreeSync, and Intel “financial horsepower.” This transcript has been revised for an article format and is not a 1:1 copy.Special thanks to the member of our Discord server who prepared this, Anubias.
Alright guys, how’s it going?
On the 6th of January, just before the tech world arrived in Las Vegas for CES 2019, Nvidia announced they were “expanding the G-SYNC ecosystem” with their “G-SYNC Compatible” monitor branding. For which they will test monitors, with those that pass their validation tests certified as “G-SYNC Compatible” and enabled by default in the GeForce driver.
At launch, 12 monitors were validated as G-SYNC compatible, and this is what I had to say to about it, in a video three days later:
Whenever Nvidia gets involved in anything like this, it is about two things: they want to be paid, and they want their brand splashed all over the good stuff. How long before FreeSync is no longer the brand on these higher-end monitors? Replaced by “G-SYNC compatible”. I’m gonna end this part of the video with a surprise to many of you by saying: AMD, this is what you deserve! You haven’t done enough quality control here and you constantly leave yourself at the mercy of Nvidia’s games. I can’t remember exactly what video it was, but a couple of years ago I said that when Nvidia is finally forced to adopt FreeSync, they will do it in way that pushes their brand over AMD’s. That’s exactly what they’re doing here, and as usual, AMD will just play the victim and they will do nothing to prevent it.
Let’s take a closer look at this list of 12 monitors for which Nvidia announced initial G-SYNC Compatible certification.
Using the Wayback Machine website we can compare the Amazon listings for some of them.
Starting at the top with the Acer XFA240, we can see that in November last year, it was branded as FreeSync Technology monitor (highlighted with red, below) in several distinct places on the page, including the FreeSync badge. If we compare the exact same monitor on Amazon today, we can see now that the FreeSync branding has been replaced by Nvidia’s G-SYNC compatible brand (highlighted with green, below), including the “Nvidia G-SYNC” badge. Not a single mention of FreeSync anywhere on the page.
Taking another look at the monitors list, let’s check what Asus are doing with the MG278Q – that’s my main monitor in fact. So once again, in the Waybach Machine, back to December 2019, we see again that AMD FreeSync Technology branding is used in multiple places on the page. And again, on the exact same page on Amazon today, MG278Q is G-SYNC Compatible. Similarly, all instances of AMD FreeSync have been changed to Nvidia G-SYNC Compatible.
Moreover, the exact same story can be found with the help of Wayback Machine on Asus’s web page for that particular monitor. And as a final kick in the teeth, they are now also displaying “G-SYNC Compatible” on the monitor picture. The same monitor, now with Nvidia logo on it. You can still see the red color around the base of the monitor. Red for Radeon, green for GeForce – of course. Except now, as I said it would happen back in January, the red is branded with green.
The exact same thing can be observed for other monitors on that list from all manufacturers: Asus, AOC, BenQ. These were all FreeSync monitors before the G-SYNC Compatible program began. Now they are all branded as either G-SYNC Compatible or “adaptive sync”.
I dunno about you guys but I’m starting to get a feeling that in order to be certified as G-SYNC compatible, ditching the FreeSync branding for plain old “adaptive sync” could be one of those unwritten requirements.
Now that was just some monitors from that list, but it doesn’t end there because now any adaptive sync monitor that Nvidia certifies is promoted as G-SYNC Compatible.
I’ll leave it up to you to decide quite why the brand FreeSync is being removed wholesale and any reference to AMD together with it, but clearly this is going to have a deleterious effect on AMD’s brand. And as I said back in January, AMD won’t do anything to stop it.
As if being Nvidia’s punching bag wasn’t enough, AMD’s even larger competitor – Intel – are back doing what they do best. But before we move on to that, let’s have a quick look at this recent article over at CRN – “Intel Slashes Price of i9 X-Series Processors By Up To 50 Percent” where Intel executive Frank Soqui told them that:
“[The price cut] is not related to AMD. It’s in service of my customer and oh, by the way, I believe AMD’s going to be the one responding to this versus us responding to them.”
Frank Soqui, vice president and general manager of Intel’s Desktop, Workstation and Channel Group, said in an interview with CRN that [the] semiconductor giant has not embarked on a processor pricing shakeup of this magnitude ever before. He said Intel decided to lower the prices for its X-series processors after observing customers who held back from upgrading due to the $1000-plus price tag of the previous generations.
“That gap is so big for the people at the top end of our mainstream roadmap [with processors like the Core i7], they can’t cross that chasm,” said Soqui, a 37 year Intel veteran. “So we brought those prices down, so that those people could move up and enjoy a higher performance.”
I don’t know about you, but I for one can detect BS from a very long distance away… and this still reeks of it even across the Atlantic.
The reality of pricing is that it depends almost entirely on the competition and also on the ability to supply the market. Intel has been struggling with the latter for two years so you maybe think this rather savage price cut seemed like an unlikely move.
On the other hand, even at nearly half the price of the previous generation, these are still not exactly stunning value CPUs. They will not appeal to gamers as Intel’s HEDT lineup is not particularly good at gaming. We saw that already with Skylake-X, and Cascade Lake-X won’t be much of an improvement either. These chips are not going to get near the i9-9900K or even the Ryzen 9 3900X.
Intel may feel that their Core i9-9900K would be the choice for gamers anyway and certainly it remains the fastest gaming CPU, albeit very slightly.
But the high-end desktop space is not about gaming, it’s about productivity. Intel’s new lineup now offers 72 PCIe lanes on all SKUs. Again, it is interesting to see the effects of strong competition on Intel. However, I wonder how previous generation Skylake-X owners feel about that. The new chips also have support for 256GB of DDR4-2933 so obviously, depending on your particular workload, these chips could end up destroying any of the Ryzen chips.
The real problem for Intel isn’t actually Ryzen, of course; it’s Threadripper. AMD will be premiering with a 24-core Threadripper which will absolutely put all of these Intel chips to the sword on all fronts. With 33% more cores and likely having higher clock speeds than the 10980XE, AMD’s Threadripper could be almost 50% faster in some workloads. The question now is, whether it will be AMD who will start segmenting their products with PCIe lanes and memory channels.
Intel priced the Cascade-Lake X parts as high as they could do without being completely embarrassed by Threadripper. They’ve left just enough room for AMD to charge up to $1500 for the 24-core part. However, I feel that $1300 is more likely. AMD could crush Intel’s entire lineup like a bug at $1000 but they won’t do that. Remember, this is AMD’s low-effort Threadripper chip. Twenty-four cores is realistically as low as they’ll go with Threadripper this generation and the only thing stopping them releasing 48 and even 64 core parts is time, that is – manufacturing enough parts that aren’t being sold as EPYC.
And that’s another clue to Intel’s pricing, as EPYC is putting the heat on Intel’s data center sales. It may well be the case that Intel expects these lower core-count Xeon chips to struggle badly against the entry level EPYCs. And in fact, even last generation Naples SKUs were beginning to put pressure on those.
Looking at Anandtech’s table of Cascade Lake-AP SKUs, there’s an 18 core there for almost $4K, $2.5 and half thousand, and some more lower down closer to the $2000 mark. At these prices they are un-buyable against some of the EPYC chips.
So I would suggest that there’s little to prevent Intel from offering vastly better high-end desktop value than we have seen from them in years. They just aren’t competitive and they will likely have a bunch of unsellable server silicon coming soon.
Let me remind you of Frank’s version of that analysis:
That gap is so big for the people at the top end of our mainstream roadmap [with processors like the Core i7], [that] they can’t cross that chasm,” said Soqui, a 37 year Intel veteran. “So we brought those prices down, so that those people could move up and enjoy a higher performance.
Again, I’ll let you decide which is more likely.
I’ll end by repeating what Frank said about AMD being the ones who will need to react:
I believe AMD’s going to be the one responding to this versus us responding to them.
With price cuts like what we’ve just seen, you’d think Intel would be a bit more humble or careful with their choice of words. But of course, they do have an ace up their sleeve… or more like an entire deck of them.
A couple of days ago this dropped into my email inbox. It is a leaked Intel slide from a sales meeting they held recently. If ever a slide outlined a game plan to deal with AMD, then this is it.
Intel’s scale advantage… financial horsepower.
A big pile of money on the left and a much smaller pile on the right. Underneath we see that Intel has set aside $3 billion dollars as “Meet Comp Discount” for 2019, which is 10 times more than AMD’s entire 2018 profit.
You see, Intel doesn’t just cut prices, they also offer very large… incentives to their partners under the guise of “meeting competition”. In short, when you can’t win with better products, you simply use your… financial horsepower to maintain your advantage instead.
Now you may recall that Intel did a lot of this in the past. If not, then you probably haven’t watched my “Intel Anti-Competitive” video yet and presumably will go do so after reading this. Back then, “meeting comp” was mostly about huge multi-billion dollar bribes to the likes of Dell and HP. Those bribes came with conditions like “you can only use 10% AMD products in your lineup, if you want to continue receiving”.
Meanwhile Intel got fined, but less than half of their “2019 meet comp” – it should be said, and they still haven’t paid it yet.
Now I’m not saying that they are doing anything illegal today. To be frank, they don’t really need to. The money that they offering just has to be enough to be worth it to the buyer to consider the Intel product over the AMD product. How this “meet comp” is paid is anyone’s guess. Perhaps it is just price cuts like we’ve seen with their HEDT lineup, perhaps it’s “contra revenue” like we saw with their failed Atom products or perhaps it’s something else.
What matters is that products are not allowed to compete on their merits. This stuff should be illegal! Period.
As I discussed in the previous video, Intel milked the desktop for almost a decade, offering quad cores at higher prices and tiny improvements year-after-year. We never saw any kind of “meet comp”. In fact, Intel never threw any of us a bone because they simply didn’t have to. We’re still not getting any bones today because AMD are doing their best to avoid direct confrontation on price. Whether that’s about to change with more production coming out of TSMC, is hard to say.
One huge issue here is this $3 billion, which looks like a huge number compared to AMD, is in fact nothing to Intel. They had 21 billion in net income last year. Should Intel really need to ramp up “meet comp” to 6 billion in 2020? Sure, they’ll do that. How about 9 billion? Sure, they will, if they need to. Intel has such deep pockets that they can buy their way out of any trouble. And they can certainly buy AMD into trouble, as well. In the last video I also talked about AMD not paying GlobalFoundries for 7nm. Maybe this is why. Had AMD agreed to buy $1 billion in 7nm wafers from GloFo every year for the next four or five years, Intel could have ramped up “meet comp” and left AMD with a bunch of unsold wafers again, even with vastly superior products.
If this situation doesn’t have you burning with a sense of injustice, then go away.
So in the end, what can AMD do?
Against Intel, I don’t know. They are just too big. Again in the last video I talked about AMD underpricing EPYC. However, with Intel able to meet comp at will, that’s probably why. On the other hand, Intel cannot match the performance of the fastest EPYCs, so for those companies that demand the very best, AMD is the only place to go and no amount of Intel kickbacks can matter. So should EPYC have been higher priced? I guess we’ll find out soon during the quarterly financials.
Against Nvidia? FreeSync is now effectively dead and AMD has no excuses for letting it happen. Or worse still, FreeSync is likely to become known as the cheap brand, further damaging Radeon’s reputation. It’s all too easy for Nvidia, and it’s easy because AMD does not appear to have a strategy for dealing with these kind of shenanigans. They should have a squad of people going around the monitor guys asking why exactly FreeSync is no longer mentioned on their monitors. And when I say “asking” I really mean something else. That’s one major difference between AMD on one side, and Intel or Nvidia on the other. The latter two would do be doing that. And that’s why AMD keeps getting their lunch money stolen by the other two. I’d be very surprised if anyone at AMD even realized this was going on.