NVIDIA CES 2023 Special Address Live Blog (8am PT/16:00 UTC)
by Ryan Smith on January 3, 2023 10:57 AM EST11:00AM EST - Thanks for joining us for another CES
11:00AM EST - Kicking things off this year, we have NVIDIA's "special address" presentation
11:00AM EST - Starting things off is Jeff Fisher
11:01AM EST - Expect this to be fast and furious, covering several of NVIDIA's business segments
11:01AM EST - First off, gaming
11:01AM EST - Recapping the Ada Lovelace Architecture
11:02AM EST - Ada introduced the hardware needed for DLSS 3, NVIDIA's frame interpolation technology
11:03AM EST - Look for NVIDIA's next DLSS 3 update later this quarter
11:03AM EST - 50 released or soon-to-be-released DLSS 3 titles
11:04AM EST - Now running a demo real of DLSS 3 games
11:04AM EST - And plugging new games that will have DLSS 3 support, including People Can Fly's Witchfire
11:05AM EST - As well as The Day Before, with RT and DLSS 3
11:07AM EST - As well as DLSS 3 support for Nexon's Warhaven
11:08AM EST - Still plugging games. Now on to Throne and Liberty as another game being announced to add DLSS 3 support
11:08AM EST - (YouTube's heavy compression really doesn't do these games justice)
11:09AM EST - Still going, Atomic Heart will be out in February with DLSS 3 support
11:10AM EST - Now on to GeForce Now
11:11AM EST - NVIDIA has continued to expand the service through additional partnerships
11:11AM EST - Ada Lovelace GPUs are coming to GFN
11:11AM EST - GFN will be using RTX 4080 Superpods
11:12AM EST - Available as part of the new Ultimate membership tier
11:12AM EST - Availability starting in select datacenters in late January, expanding beyond that throughout the quarter
11:13AM EST - The Ultimate tier can stream at up to 4K. $20/month
11:13AM EST - Hardware announcement! RTX 4070 Ti
11:14AM EST - This is the previously cancelled RTX 4080 10GB, which is now back as a lower tier card
11:14AM EST - Available on January 5th with prices starting at $799
11:15AM EST - The RTX 30 series will make up the rest of NVIDIA's desktop stack for the time being
11:15AM EST - Now on to mobile and laptops
11:15AM EST - The Ada Lovelace architecture is coming to laptops
11:15AM EST - RTX 40 Series laptops
11:16AM EST - NVIDIA is claiming up to 3x the power efficiency and 4x the performance (using DLSS 3, I'm sure)
11:16AM EST - Max-Q is back for this generation as well
11:17AM EST - Ada Max-Q laptops will use ultra low voltage GDDR6, as well as what they are terming "tri-speed memory" which can go to additional low-power voltage states
11:18AM EST - Introing RTX 4070, 4060, and 4050 laptops
11:19AM EST - Laptop prices start at $999, available February 22nd
11:19AM EST - Oh, RTX 4090 and 4080 as well
11:19AM EST - Starting at $1999, available February 8th
11:19AM EST - No mention of which specific GPUs are being used here
11:20AM EST - And that's gaming. On to NVIDIA Studio news
11:20AM EST - Talking up the impact of AI on content creation
11:21AM EST - As well as RTX acceleration for other creation applications
11:22AM EST - And of course, it wouldn't be an NVIDIA presentation without an Omniverse mention in there somewhere
11:22AM EST - Now rolling a lifestyle video showing how Omniverse can be used in content creation
11:23AM EST - And making sure to name drop all of the applications that can export to Omniverse via USD
11:24AM EST - And a quick Toy Jensen cameo
11:26AM EST - Audio2Face for Blender, NVIDIA's AI Toybox, RTX Remix, and more
11:26AM EST - And finally, NVIDIA's Avatar Cloud Engine (ACE)
11:27AM EST - Simplifies and semi-automates the creation of avatars
11:28AM EST - NVIDIA has also updated its impressive Canvas program with 360 degree support
11:29AM EST - Also announcing RTX Video Super Resolution. Using AI upscaling to improve web video quality
11:29AM EST - For Chrome and Edge (so, Chromium) browsers
11:29AM EST - And back to Jeff. It's already time for the recap!
11:30AM EST - Oh, that was just for graphics. Now on to NVIDIA robotics
11:32AM EST - Sounds like NVIDA is going to be focusing on Isaac here, their simulation platform for training robotics AI models
11:33AM EST - Over 1000 companies are using some part of the Isaac platform
11:33AM EST - Recapping the availability of Isaac from cloud services
11:34AM EST - And Isaac will soon be able to handle multiple robots in a single instance
11:36AM EST - Announcing the next release of Isaac today. New conveyer building tool, radar/lidar support, and more
11:36AM EST - "There's something for everyone in the new Isaac sim"
11:37AM EST - And now on to self-driving cars
11:38AM EST - Recapping NVIDIA's Drive platform, as well as the DGX and OVX systems it's backed by
11:39AM EST - As well as talking up the benefits of Drive centralizing many of the microcontrollers used throughout a car
11:40AM EST - Meanwhile, the shift to EVs is giving car designers a chance to redesign their systems, as well as to incorporate new hardware and features (e.g. self-driving computers)
11:41AM EST - NVIDIA is announcing a partnership with Foxconn, who will be building EVs based on the Drive Hyperion platform
11:41AM EST - (I had no idea Foxconn made cars. But it makes sense; they make just about everything)
11:42AM EST - GeForce Now is coming to cars
11:42AM EST - GFN will be available in internet-enabled cars
11:43AM EST - Several Drive partners are integrating GFN, including Hyundai and Polestar
11:44AM EST - Mercedes-Benz will be using Omniverse for assembly and production planning
11:44AM EST - Now rolling a short video on the subject
11:45AM EST - NVIDIA's tech is also being used to design and simulate internal physical design concepts
11:47AM EST - And that's a wrap on NVIDIA's CES 2023 keynote. Thanks for joining us!
9 Comments
View All Comments
Samus - Tuesday, January 3, 2023 - link
The 4070Ti is faster than the 3090Ti while having 33% less cores of near-equal capability and 1/3rd the non-RT tflops? So basically this is only true in edge cases where DLSS3 trickery allows it to excel, at the expense of a consistent gaming experience.Seems like nVidia is relying on software this generation to push the numbers.
shabby - Tuesday, January 3, 2023 - link
Magic bro... magic!TEAMSWITCHER - Tuesday, January 3, 2023 - link
"near-equal capability" is a statement you pulled out of your ass.ET - Tuesday, January 3, 2023 - link
NVIDIA said "faster than 3090 Ti", and I'd say that this is clearly meant when using DLSS 3.We'll have to wait for reviews to see how it performs when not using DLSS 3. Which could very well be "slower than 3090 Ti".
catavalon21 - Tuesday, January 3, 2023 - link
That. Well, somewhere I suppose.techjunkie123 - Tuesday, January 3, 2023 - link
You don't need to guess about this.....Nvidia had already shown benchmarks for the 4080 12gb. Now rebranded as the 4070 Ti. It is indeed slower than or comparable to the 3090 Ti at best without frame generation.cycomiko - Wednesday, January 4, 2023 - link
So, does their continued use of 30 series in the remainder of the desktop space result in a cost reduction, or are they just continuing to profit excessively off a generation old tech?Gothmoth - Wednesday, January 4, 2023 - link
what a sh*tshow.3x the perfomance. nvidia must think everyone is braindead.
799$... lol.
someone should investigate nvidia and amd for price fixing.
this is crazy.
Nfarce - Thursday, January 12, 2023 - link
"someone should investigate nvidia and amd for price fixing...this is crazy."No, what is crazy is people like you who think that the overlord big government should step in and set prices. Nothing would kill innovation faster. You think that money they make doesn't go into R&D research to future generations of GPUs? If people like you were in charge of things, we'd still be running on $325 GTX 970 video card tech at 22fps in modern DX12 games at 1080p.
Finally, nobody price fixes anything. If the market buys, the price is right. And clearly with Nvidia's demand over the past two decades of GPUs, there is PLENTY of demand.