Quick summary: AI mentioned 119 times Aritificial Intelligence mentioned 20 times. That gives 139 AIs over 100 minutes of keynote. 1,39 AIPM. And yes, I used AI to count all the mentions of it.
Wait fot review, coz he said not possible with all the new AI features. Could mean only tested under those new rendering technologies. So we shall see.
Not possible without AI is some slimy marketing tactics all to say you need to use GeForce specific features on compatible platforms to get the full performance and it’s not actually rendering everything, only spitting out AI fill in graphics. Frame gen inserts, upscaling, full AI created shaders.
They should really start making new generation monitor modules that kind of like the original G-Sync modules were, only the new generation ones have an efficient GPU built in that specializes in only frame generation of the already completed frames that are sent to the monitor. Then have the monitor double or even quadruple the framerate output by the GPU and do it all on local to the monitor processing so we don't exceed the HDMI/DP cable bandwidth. With no real rendering of geometry, or shader computation or any of that, simple interpolation generation of frames the GPU should be capable of doing a 4x frame gen with a fairly small GPU core wise.
Bravo to Jensen and the entire Nvdia team for this presentation, and more importantly for the AI-future to come. This is incredibly exciting, yet scary at the same time...what AI tech can bring goodness to humanity, or harm it if we are not careful.
@@juiceskyrous Well hello there retard. How the flying fuck can a 12GB card perform like an 24GB card... Sure it can on 1080p but no one is buying an 4090 to play at 1080p!!!!! I want you to run Cyberpunk or some other high end graphics game @ 4k with no DLSS or other upscale crap running. Now tell me if a 5070 is even close to a 4090... Now if you run all the fake frame crap then maybe you get close at 1080p but never in higher resolution because you run out of Vram idiot... Nvidia and the stupid fan boys never can see a lie even if it slapped them in the face.
Thanks for sharing such valuable information! Could you help me with something unrelated: My OKX wallet holds some USDT, and I have the seed phrase. (alarm fetch churn bridge exercise tape speak race clerk couch crater letter). What's the best way to send them to Binance?
I'm a French 13yo boy and I got my high school diploma last year. I started learning machine learning before looking for a university. I own a 6 year old laptop with NVIDIA GPU and I need a lot of creativity to optimize my code. I look forward to get such an amazing supercomputer. BTW I also need creativity because no budget
"In his 2050 CES keynote speech, Jensen Huang mentioned that quantum computers would not become a reality for another 20 years. However, Korea's LK99 is set to finalize its room-temperature and atmospheric pressure superconductor patent by 2025 and release pilot products within the next two years. Given this, it seems Jensen Huang has made a significant miscalculation. Quantum computers and room-temperature and atmospheric pressure superconductors are directly related, and he is overlooking the fact that the commercialization of room-temperature and atmospheric pressure superconductors will happen much sooner than he predicts. The patent for room-temperature and atmospheric pressure superconductors has already been registered in Eurasia and South Africa. By mid-January, the foundational patent will be completed in Korea, followed by PCT patent registration, and patents will also be registered in the US, Japan, and Europe. Meanwhile, the zero-resistance test results will be announced with the guarantee of the Korea Institute of Basic Science (KBSi). The Quantum Energy Research Institute in Korea has already produced a prototype power transmission line using room-temperature and atmospheric pressure superconductor technology, and they are just waiting for the patent registration announcement. Jensen Huang is completely unaware of the development status of Korea's superconductor technology and delivered an outdated lecture. As a result of this lecture, the stock prices of the quantum computer sector in the US market plummeted, showing a temporary adjustment in response to Jensen Huang's speech. However, as the commercialization of room-temperature and atmospheric pressure superconductor technology becomes imminent, both Jensen Huang and the general public will come to realize this. If Jensen Huang does not drastically revise his business plans, businesses including NVIDIA will surely suffer. Misleading information like this has negatively impacted the US stock market, but it must be corrected with accurate information acknowledging that Korea's room-temperature and atmospheric pressure superconductors are on the verge of commercialization. Quantum computers utilizing room-temperature and atmospheric pressure superconductors will become commercially viable within three years. Thus, the US stock market should not waver but continue its upward trend. This accurate information represents the fair functioning of a market economy in capitalism."
cant wait for my 5090. Just hope I can buy one here in Austria. already built my 9800x3D pc when AMD released that CPU. Also bought a 32" 4k 240hz QD-OLED in August, in anticipation of the RTX 5090. please let me (and other waiting customers ofc lol) pre order a FE model of the 5090, here in Europe. It cant be that scalpers buy up everything and then sell this thing for 5000€ or anything. So please offer pre orders, so you can built exactly the amount the ppl need. I would pay the whole thing immediately now, if that means, I can get one.
the only thing Im a bit worried is the fact that 575w is extremely close to the 600w limit of the 16pin connector, so I dont know if this works out. Even with the updated 12vhpwr connector "with shorter pins" Im not so sure, since there def were cases where it appeared that the connector was fully seated and it still burned up. I think GN and Igorslab have some nice info on these things. I dont think Nvidia would push this single connector so hard, if it didnt work but still. a saving grace is obviously the fact that a card can draw like 75w through the PCIE slot and my motherboard (x870e Carbon Wifi, MSI) has an extra 8 pin connector on the bottom to deliver power to the GPU, so thats another 150w on top. Meaning the GPU can get around 200w from the motherboard, so "only" 475w on the 16pin. But yeah, I dont know. Will still buy it ofc.
This is absolutely amazing!! However I'm wondering what's going to happen to jobs, homes, and the 99% of the human race??? Who can afford such a machine?
Don't use clips made by drivers that might have decided wrong in that scenario to train ai. You might build a rogue ai on the road that often breaks the law 😅
Too bad it didn't have 4x the RT Tflops instead of the 2x it got, so that ray tracing would actually be useful with only a small performance hit then. Nvidia really milking the slow technology drip.
4090 for 1599 USD. Oh yeah ? In which AI dimention is this a reality ? 5090 for 1999 USD ? 😂 Sure thing. It will cost a fucking grand more than that on the market thanks to scalping that nvidia is completely fine with. 5070 has 4090 performence for "only" 549 USD. Not in this dimention. Never ever going to be a reality. I bet all those fancy audio issues in the keynote were made by AI as well. The future is now guys.
I'll believe it when I see reviews from Gamer Nexus or similar. He was specific, it's through DLSS... so it might be a play of marketing words. Like 5070+DLSS = 4090 no DLSS.
I have a 4090 and a 4070 Super. That 4070 can do more with DLSS than my old 3080ti could. However - without DLSS it's not powerful at all. That's how they're playing this.
@@11nvaderZim Yeah, big claims, but without having equivalent specs in memory and bandwidth, it looks like to reach those heights it will take games that take advantage of those new compression bells and whistles he mentioned during the presentation. If it can take any game and run at 4090 levels, even without DLSS, that would be a surprise and quite impressive.
Very important reminder ray tracing with denoiser applied once again because the GPU can't handle it otherwise AND with DLSS to compensate even further is not really full ray tracing. GPUs are just not ready for this.
Skipping this series easy. Not worth at all to trade in the 3080 12GB I have. I don't use the DLSS and other frame upscale crap simply because it makes the game look very bad and graphics are all over the place. Tried DLSS on Cyberpunk... fucking hell that got turned off FAST since it really messes with lighting and everything else. So next GPU might be AMD for the very first time for me and I been with team green since GeForce 256. That stops now. Prices are also so insane it just to laugh at.
Jensen Huang in a nutshell So you all know now that we have established ourselves as a Leading AI tech, here's your not so improved next gen RTX gpus with synthetic frames from lower resolutions so you all stop complaining about VRAM limitations cause we don't give a crap about gamers as there are tonnes of corpos out there lining up for our industry grade GPUs. Be happy with fake frames and the statictics presented for comparison with last gen and forget about raw performance & vram coz we don't talk about it here. With more upcoming unreal engine 5 or 5.5 titles, VRAM limitations are gonna cause problems.
Start of presentation: 23:12
RTX Introduction: 38:37
Pricing: 42:38
Ur my favorite person rn
Ngl this whole presentation brings a good vibe to the start of the year
RTX: 38:37
Price: 42:38
Nice
I love that he asked for a sip of a beer halfway through 😂
Keynote Starts at 23:17
Quick summary:
AI mentioned 119 times
Aritificial Intelligence mentioned 20 times.
That gives 139 AIs over 100 minutes of keynote. 1,39 AIPM.
And yes, I used AI to count all the mentions of it.
I love how he had me going for just a bit there when he said it was "Windows...." and then " WSL". LOL!
that was amazing... he even managed to avoid saying LINUX ffs 🤣
Understandable why AMD's RDNA4 is a no-show then.
Nvidia crushing everything.
Where did the first guy learn to say nvidia?
if 5070 really = 4090 it's massive bang for the buck
That 12gb VRAM is gonna hold it back so much though
Wait fot review, coz he said not possible with all the new AI features. Could mean only tested under those new rendering technologies. So we shall see.
@@edenfreeman5171 That's my guess. There are many unknowns and he stressed it was all due to AI/DLSS.
Not possible without AI is some slimy marketing tactics all to say you need to use GeForce specific features on compatible platforms to get the full performance and it’s not actually rendering everything, only spitting out AI fill in graphics. Frame gen inserts, upscaling, full AI created shaders.
Yes
They should really start making new generation monitor modules that kind of like the original G-Sync modules were, only the new generation ones have an efficient GPU built in that specializes in only frame generation of the already completed frames that are sent to the monitor. Then have the monitor double or even quadruple the framerate output by the GPU and do it all on local to the monitor processing so we don't exceed the HDMI/DP cable bandwidth. With no real rendering of geometry, or shader computation or any of that, simple interpolation generation of frames the GPU should be capable of doing a 4x frame gen with a fairly small GPU core wise.
Bravo to Jensen and the entire Nvdia team for this presentation, and more importantly for the AI-future to come. This is incredibly exciting, yet scary at the same time...what AI tech can bring goodness to humanity, or harm it if we are not careful.
So raster basically unchanged from previous generations, possibly improved slightly due to GDDR7? However more synthetic frames are created?
Anything about better Linux support?
When people think a 5070 can beat a 4090 😂😂😂😂
@@juiceskyrous Well hello there retard. How the flying fuck can a 12GB card perform like an 24GB card... Sure it can on 1080p but no one is buying an 4090 to play at 1080p!!!!! I want you to run Cyberpunk or some other high end graphics game @ 4k with no DLSS or other upscale crap running. Now tell me if a 5070 is even close to a 4090... Now if you run all the fake frame crap then maybe you get close at 1080p but never in higher resolution because you run out of Vram idiot... Nvidia and the stupid fan boys never can see a lie even if it slapped them in the face.
interesting guy) nice presentation)
Thanks for sharing such valuable information! Could you help me with something unrelated: My OKX wallet holds some USDT, and I have the seed phrase. (alarm fetch churn bridge exercise tape speak race clerk couch crater letter). What's the best way to send them to Binance?
I'm a French 13yo boy and I got my high school diploma last year. I started learning machine learning before looking for a university. I own a 6 year old laptop with NVIDIA GPU and I need a lot of creativity to optimize my code. I look forward to get such an amazing supercomputer. BTW I also need creativity because no budget
It looks like announcement of iPhone and iPhone 4, many wow factors to shine the whole world.
"In his 2050 CES keynote speech, Jensen Huang mentioned that quantum computers would not become a reality for another 20 years. However, Korea's LK99 is set to finalize its room-temperature and atmospheric pressure superconductor patent by 2025 and release pilot products within the next two years. Given this, it seems Jensen Huang has made a significant miscalculation. Quantum computers and room-temperature and atmospheric pressure superconductors are directly related, and he is overlooking the fact that the commercialization of room-temperature and atmospheric pressure superconductors will happen much sooner than he predicts.
The patent for room-temperature and atmospheric pressure superconductors has already been registered in Eurasia and South Africa. By mid-January, the foundational patent will be completed in Korea, followed by PCT patent registration, and patents will also be registered in the US, Japan, and Europe. Meanwhile, the zero-resistance test results will be announced with the guarantee of the Korea Institute of Basic Science (KBSi). The Quantum Energy Research Institute in Korea has already produced a prototype power transmission line using room-temperature and atmospheric pressure superconductor technology, and they are just waiting for the patent registration announcement. Jensen Huang is completely unaware of the development status of Korea's superconductor technology and delivered an outdated lecture. As a result of this lecture, the stock prices of the quantum computer sector in the US market plummeted, showing a temporary adjustment in response to Jensen Huang's speech. However, as the commercialization of room-temperature and atmospheric pressure superconductor technology becomes imminent, both Jensen Huang and the general public will come to realize this. If Jensen Huang does not drastically revise his business plans, businesses including NVIDIA will surely suffer. Misleading information like this has negatively impacted the US stock market, but it must be corrected with accurate information acknowledging that Korea's room-temperature and atmospheric pressure superconductors are on the verge of commercialization. Quantum computers utilizing room-temperature and atmospheric pressure superconductors will become commercially viable within three years. Thus, the US stock market should not waver but continue its upward trend. This accurate information represents the fair functioning of a market economy in capitalism."
the day AI becomes self-aware and recognizes this jacket, it won't end well for Jensen... just saying ^^ #stayingalive 🕺🏼
I NEED 5060 !!!
Is CES not going to post the hundreds of other tech talks going on?
cant wait for my 5090. Just hope I can buy one here in Austria.
already built my 9800x3D pc when AMD released that CPU. Also bought a 32" 4k 240hz QD-OLED in August, in anticipation of the RTX 5090.
please let me (and other waiting customers ofc lol) pre order a FE model of the 5090, here in Europe. It cant be that scalpers buy up everything and then sell this thing for 5000€ or anything. So please offer pre orders, so you can built exactly the amount the ppl need. I would pay the whole thing immediately now, if that means, I can get one.
the only thing Im a bit worried is the fact that 575w is extremely close to the 600w limit of the 16pin connector, so I dont know if this works out. Even with the updated 12vhpwr connector "with shorter pins" Im not so sure, since there def were cases where it appeared that the connector was fully seated and it still burned up. I think GN and Igorslab have some nice info on these things. I dont think Nvidia would push this single connector so hard, if it didnt work but still.
a saving grace is obviously the fact that a card can draw like 75w through the PCIE slot and my motherboard (x870e Carbon Wifi, MSI) has an extra 8 pin connector on the bottom to deliver power to the GPU, so thats another 150w on top. Meaning the GPU can get around 200w from the motherboard, so "only" 475w on the 16pin. But yeah, I dont know. Will still buy it ofc.
1:40:00 Detroit become human is REAL
5000 series is amazing
Hi everyone
This is absolutely amazing!! However I'm wondering what's going to happen to jobs, homes, and the 99% of the human race???
Who can afford such a machine?
I need to see Tom’s hardware benchmarks asap
Some boomer will say that best gpu for GTA 6
when AI make GTA 6 NPC shoot you
I'll just keep using my evga 1080ti Hybrid
I came here only for Virtua Fighter.
🔥🔥 presentation! Nvidia will be the first multi trillion dollar company
26:46 No worries, as there were no crocodiles hurt while manufacturing that jacket...
Don't use clips made by drivers that might have decided wrong in that scenario to train ai. You might build a rogue ai on the road that often breaks the law 😅
Skynet is here
Too bad it didn't have 4x the RT Tflops instead of the 2x it got, so that ray tracing would actually be useful with only a small performance hit then. Nvidia really milking the slow technology drip.
It's alot of stuttering in the presentation. 🤨
Euclides. 🙂👍👍👍💯🇺🇲🦅. Hello. You. 🙂👍💯
there are better image/sound quality of this video.
how much memory for the 5070 as it wasn't mentioned?
12 gigs, goto nvidia's site for specs.
2000 for a video card?
37:43 *transformers now inside a gpu*
36:22
It’s all fun & games til you know they’re end game for all of this…
Just make sure do not let chinese involved !!!
RayTracing Leather jacket now available at Nvdia online shop ... 🤪🤣😂
NAvidia… 😮
Micron MU 💼
yes it i undervalued, it is gonna rise morning!
4090 for 1599 USD. Oh yeah ? In which AI dimention is this a reality ?
5090 for 1999 USD ?
😂
Sure thing. It will cost a fucking grand more than that on the market thanks to scalping that nvidia is completely fine with.
5070 has 4090 performence for "only" 549 USD.
Not in this dimention. Never ever going to be a reality.
I bet all those fancy audio issues in the keynote were made by AI as well.
The future is now guys.
Wtf 5070=4090
I'll believe it when I see reviews from Gamer Nexus or similar. He was specific, it's through DLSS... so it might be a play of marketing words. Like 5070+DLSS = 4090 no DLSS.
@ i know, but wtf xd
I have a 4090 and a 4070 Super. That 4070 can do more with DLSS than my old 3080ti could. However - without DLSS it's not powerful at all. That's how they're playing this.
@@11nvaderZim Yeah, big claims, but without having equivalent specs in memory and bandwidth, it looks like to reach those heights it will take games that take advantage of those new compression bells and whistles he mentioned during the presentation. If it can take any game and run at 4090 levels, even without DLSS, that would be a surprise and quite impressive.
Frame gen trickery.
more blurry gaming incoming boys!
What card you've got ?
@@bunnyrabbitplays8979 4080 Super and I have to turn DLSS off on some game to remove the blur.
@@bunnyrabbitplays8979 4090
Very important reminder ray tracing with denoiser applied once again because the GPU can't handle it otherwise AND with DLSS to compensate even further is not really full ray tracing.
GPUs are just not ready for this.
Msrp 42:41
Nevidia
first Cant wait for my 5090
start buying kion!!!!
Skipping this series easy. Not worth at all to trade in the 3080 12GB I have. I don't use the DLSS and other frame upscale crap simply because it makes the game look very bad and graphics are all over the place. Tried DLSS on Cyberpunk... fucking hell that got turned off FAST since it really messes with lighting and everything else. So next GPU might be AMD for the very first time for me and I been with team green since GeForce 256. That stops now. Prices are also so insane it just to laugh at.
Your data and habits will pay. Forget any privacy leftovers ...
4090 dead
That's not gonna be raster
@@rockzors True, its dlss4 vs 4090 raster
You sound slow 😂
For the great pricing of 2 kidneys and an arm!
They're actually priced pretty competitively...
Curious to see how AMD and Intel responds.
Your next fascist state, powered by Nvidia! Yay!
Jensen Huang in a nutshell
So you all know now that we have established ourselves as a Leading AI tech, here's your not so improved next gen RTX gpus with synthetic frames from lower resolutions so you all stop complaining about VRAM limitations cause we don't give a crap about gamers as there are tonnes of corpos out there lining up for our industry grade GPUs. Be happy with fake frames and the statictics presented for comparison with last gen and forget about raw performance & vram coz we don't talk about it here.
With more upcoming unreal engine 5 or 5.5 titles, VRAM limitations are gonna cause problems.
1600 for 4090 lol he's delusional
🚀
I dunno man. 2000USD for a GPU is nuts. Rooting for Intel.
rtx, framegen, dlss, fsr etc. are all memes and this gen sucks fat balls. nvidia will NEVER have another 1080ti moment