00:00Envy has unveiled the latest RTX 5000 with Blackwell DSS 4 architecture and refills 2 at CES 2025 Las Vegas.
00:09I'll summarize everything in just 10 minutes. Let's get started.
00:15They've finally unveiled the desktop graphics cards that we've talked about so much in Leaks and Rumors,
00:21which, well, are just that, Leaks and Rumors. Okay?
00:25Well, it can finally be said that they all have GDDR7, that much acclaimed and long-awaited memory.
00:32Specifically, they've unveiled four models, the 5070 will have 12GB of GDDR7,
00:38the 5070T will have 16GB of GDDR7, the 5080 with 16GB of GDDR7 and, watch out, the 90 with 32GB of GDDR7.
00:49Things to note about these configurations.
00:51The 5080 has a speed of 32GB per second on these modules. Watch out, the fastest of all,
00:59but the 5090 has a memory bus, finally, finally. 512-bit, which was what was needed in these
01:06configurations, already 20, 24, 32GB. Also some things to consider.
01:14What performance can we expect from the desktop RTX 5000?
01:18What they have released is a 50.90 graphics card that can even double the performance of the 4090.
01:26But the next or most outrageous thing that has had a lot of impact on the RTX 5070 networks and a
01:32performance supposedly equivalent to the 4090. This is crazy.
01:37And by the way, they claim this with the Ryzen 7 9800X 3D within the setup, they have tested to sign this.
01:46Not bad at all, guys. The last thing I want to comment on is very important about these RTX 5000 from NVIDIA.
01:52Are they going to consume more? Or at least that's what the TDP indicates. It is true that the TDP does
01:57not directly indicate how much it consumes, but we already have experience in this. And when the TDP goes up,
02:02the final cosmos goes up. I don't want to give you any away, but the TDP of the RTX 5090 is 575 WS.
02:13They have introduced DLSS 4, NVIDIA's famous technology for intelligent upscaling using deep learning.
02:22It has been updated and has introduced multi-frame generation. We come from DSS 3 with the famous
02:28frame generation that doubled the FPS compared to the RTX 3000. Well, multi-frame generation is capable
02:34of extracting three additional frames from a single one, which is absolutely incredible. Hence, well,
02:40we see performance graphics cards like the RTX 5090 and the 4090 where the first one doubles the
02:47performance of its predecessor. Incredible, guys, incredible. In fact, I think that DLSS 4 has been
02:54the possibility or the key that opens the door for the RTX 5070 to be equivalent to the 4090 in
03:01performance. I guess 1080p because if we go into 4K, I highly doubt that the 1070 will, hey, reach the
03:08FPS of this wonderful GPU that we've had in PC components with 24GB GDR6X. Amazing, I don't think
03:16it's a graphics card anymore, but something superior. It's a work tool for all those data scientists or
03:22people who work in artificial intelligence. Something I didn't like about DLSS 4, guys,
03:29is that it's only compatible or supported by the RTX 5000, or at least that's what's understood from
03:35what Nvidia has put out. I don't really like this, what can I say, especially for those users who bought
03:41an RTX 4000 Super, a 4070, 4080 Super, 4070T Super last year and, hey, you're stuck with DLSS 3.
03:51I don't like it. I understand Nvidia, I respect them, come on, everyone who works at Nvidia is a genius,
03:59but hey, they could have extended that to RTX 4000 Super users, I think.
04:03By the way, a special mention should be made of the AI model they presented, Transformer,
04:12that's what it's called, in Nvidia Envy for DLSS 4. It does all the frame generation in less time,
04:19uses overamp memory, and only runs once. It's a model that optimizes the entire process so that we gain FPS.
04:27Along with DSS 4, they have presented Nvidia Reflex 2, the evolution of the original Reflex,
04:39which is a low latency technology to prevent latency caused by DLSS frame generation in esports or
04:45competitive games, in this case multi-frame generation, because we're talking about DLSS 4 to
04:51prevent us from getting killed, saying, hey, but I clicked and the shot didn't go off when I clicked,
04:57etc. Well, that additional latency introduced by friend generation is mitigated through Nvidia Reflex.
05:08What's new in Nvidia Reflex 2?
05:14Well, something called Frame Warp, and they've come up with a system to further reduce latency
05:19based on how we change cameras when we move in the game, right?
05:22We're in a shooter and suddenly we move the mouse a little because the enemy appears, right? Well,
05:30all those camera movements have introduced a predictive artificial intelligence model, right?
05:37To reduce latency by up to 75%. I think that's spectacular, they've demonstrated it with several
05:44demos and so on. And hey, it's not bad at all, it's not bad at all, but I have the good news here and that is
05:51that it's compatible with the RTX 5000 at first, but then it will be extended to all RTX generations.
06:02And what will the starting prices of the RTX 5000 be?
06:07We're talking about IDD starting prices, okay? This means that there will be more expensive models,
06:12especially the custom ones from Asus, Gigabyte, MSI, and Zota. They will all have their higher-end models
06:19that will cost more. Let's start with the 5070, with a starting price of 659 euros. The 70T has a
06:27starting price of around 899 euros. The price of the RTX 5080 remains the same compared to the 4080,
06:35at 11,199 euros. And it is true that the RTX 5090 is the one that has increased the most in terms of
06:42price. And we can find it from 2300 euros, almost 2400 euros for friends. Well, it is a high price,
06:51it is a low price. I think Nvidia hasn't done anything wrong, and it's true that I would love
06:55for the 5080 to cost less than the 70TTI 2. But hey, that's how technology works and we all know that
07:01when a new technological, cutting-edge product comes out, etc. You're not just buying a graphics
07:07card from Nvidia, above all, you're also buying a software suite that's behind it, very, very
07:12powerful. And I have to say that I was pleasantly surprised, especially the 16GB 5070T, the original
07:214070T only had 12GB, they had to release the S model to offer 16GB. For me, in this sense, very good.
07:30Be careful for those of you who think that the RTX 4000 will be cheaper now that the RTX 5000 is coming
07:38out. This doesn't have to work that way. We've already seen that Nvidia tends to progressively
07:44phase out older models to make way for the most recent ones. All brands do this, guys, it's not
07:50just Nia. I think the 4060, the 4060T will be available, there will still be good stock until the
07:565060 comes out. Even with the 5060 coming out, we've already seen that the 4060 and the 3060
08:04have been coexisting a bit and well, I think the price could be regulated a bit, dropping below 300
08:10euros, even at some point in 2025. I don't know, I'm speaking based on my own beliefs. Uh, I don't have
08:17any information about any type of PC components, but as you know, the 4070T, the 4080 and the 4090 have
08:25already stopped producing completely and the stock is minimal, if there is no more, okay, first-hand.
08:31Now, if you get them from other second-hand channels, etc. Well, that's up to each person to
08:36decide what they want to do, but I can tell you in advance that the 4070 TS, the 4070 Super,
08:43and the 4070 may be discontinued. In fact, if I remember correctly, the 3060T dropped,
08:50leaving only the 3060. You've been warned.
08:57Launch of the NVIDIA RTX 5000. When will they be? Well, I have to say that on January 30th for the
09:040, 50, 90 and for the rest of the 50, 70 in February, they will arrive in February. Good news for me,
09:11I thought they were going to announce, present, and launch only two models, the 5080 and 5090,
09:17and they were going to leave the 5070 for spring, summer. But hey, four entry-level models, also
09:23spaced out in prices. We'll have the 4060 in the low-end range, around 300 euros, which isn't bad
09:30at all, hey, I think an 8GB GPU from SS3, etc. Not bad at all. And could there be a 5060 later this year?
09:38Well, maybe, we don't know. So subscribe to this channel, give it a like if you like this content,
09:45and follow us because we'll discuss all the news here. Bye, bye, bye, bye, bye.
Comentarios