NVIDIA, the Santa Clara, California-headquartered technology giant, is in a major boom cycle lately. With the superheated hype of artificial intelligence (AI) going forward at full steam, NVIDIA is one of the companies that’s in the forefront of this field of innovation and is fully plugging into it. In fact, thanks in part to AI, NVIDIA’s valuation had, for a while, even reached the rarified air of the corporate “trillion dollar club.”
With that kind of momentum behind the company, its CEO and founder Jensen Huang delivered a keynote at the massive COMPUTEX 2023 expo, which was held from May 30 to June 2, 2023 in Taipei, Taiwan. Huang brought plenty of AI news (and many other updates), including news addressing game visualization technology, one of their core markets.
In this nearly-two hour video on YouTube, along with many other announcements, the NVIDIA GeForce RTX 4060 Ti GPU was announced as being in mass production, which may be of interest for gamers looking for a more approachable entry point into the 40 Series RTX GPUs. The RTX 4060 Ti GPU starts at a relatively tame $399 MSRP, especially when compared to the very powerful and very expensive flagship GPU, the RTX 4090, with its $1,599 MSRP.
On the software development side, NVIDIA also announced what it calls the NVIDIA Avatar Cloud Engine (ACE) for Games. ACE is designed as a customizable AI model creation service for the cloud or hosted on local devices with non-player characters (NPCs) able to communicate with the player via AI large language model (LLM)-powered engagement. To see the demo example, give this link from the video a click and you’ll see how ACE works in action. “AI will be a very big part of the future of video games,” Huang says in summary. As either developers or players, that’s a prediction that may well come true.
After these keynote-leading announcements, numerous other data center and supercomputer-class GPUs and AI and big data-processing announcements were made, including around when to expect the next generation of the Grace Hopper
“superchip,” which is currently in the GH200 iteration and in full production. But what if that’s not enough juice for your needs? That’s when you might consider — wallet permitting — the DGX GH200, which is essentially 144TB of memory shared across 256 NVIDIA Grace Hopper GH200s.
This brief summary, of course, only scratches the surface of this densely-packed keynote address. To dig deeper, you can visit the official video, in whole or in part, that follows!
Enjoyed the video? Be sure to subscribe to our YouTube channel for more videos!
Whether you are an event coordinator, press, or general attendee, if you have an event video to share with us, please let us know and we’ll be happy to feature it!
