Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/computex.22840/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Computex

P

Paul2

Well-known member
So the big date of the year is coming. Gentlemen, what are your bets for the biggest news coming from Computex?

Main Themes
COMPUTEX is a leading global exhibition focused on AIoT and startups.
The expo will continue with the position of “AI Next”,
featuring the latest tech trends: AI & Robotics, Next-Gen Tech, and Future Mobility.
Recognized for its adaptability to industry changes.
It has established itself as a premier platform for showcasing technological innovations.

www.computextaipei.com.tw

Join the World's Largest AI Exhibition!

COMPUTEX has evolved into the world’s leading platform for AIoT and startups, showcasing the latest trends in technology. This premier exhibition highlights innovation and collaboration, driving the global tech ecosystem. With a commitment to connecting visionaries, COMPUTEX fosters...
www.computextaipei.com.tw www.computextaipei.com.tw
 
It will be AI everything like the rest of the conferences I have attended this year.

Jensen Huang will be there of course and Lisa Su. They are big Taiwan supporters.
 
The tech media has a competition where they count who said the word "AI" most of the times and AI/minute 🤣.
 
Last edited:
Definitely some new "bad" marketing coming at Computex :)

Nvidia marketing is cooking early (p)reviews of the RTX 5060 by not providing drivers in advance of launch to reviewers, unless they agree to release a preview under very strict guidelines. Those guidelines include making sure the card doesn't run out of VRAM (8GB) so it doesn't appear worse than RTX 3060 (12GB), and allowing 50 series to do AI multiple frame gen while the older cards are software/driver locked from doing so.

1747578307724.png
 
Xebec said:
Definitely some new "bad" marketing coming at Computex :)

Nvidia marketing is cooking early (p)reviews of the RTX 5060 by not providing drivers in advance of launch to reviewers, unless they agree to release a preview under very strict guidelines. Those guidelines include making sure the card doesn't run out of VRAM (8GB) so it doesn't appear worse than RTX 3060 (12GB), and allowing 50 series to do AI multiple frame gen while the older cards are software/driver locked from doing so.
Click to expand...

1747587669551.png
 
NVIDIA CEO Jensen Huang delivers a live keynote ahead of COMPUTEX 2025 in Taipei, unveiling what’s next for the era of AI factories. Don’t miss this glimpse into the future of accelerated computing—May 19 at 11:00 a.m. Taipei time.


 
1747785869834.png


NVIDIA founder and CEO Jen-Hsun Huang said on the 20th that Moore's Law has ended. NVIDIA has used three methods to rapidly advance artificial intelligence (AI) chips and industries, and has driven the development of AI chips at a very fast pace every year. In the future, only the sky can limit the speed of development of the AI industry.

Foxconn Chairman Liu Yangwei delivered a keynote speech at the 2025 Computex and invited Jensen Huang as a mystery guest. Given that the development speed of AI chips and industries is very fast every year, does Huang see that the development of AI chips and industries will be restricted?

Jensen said, "Moore's Law is over because we see the limitations of the physical world, as well as rising chip costs and energy efficiency issues, and the speed of processors has not increased rapidly. Therefore, Nvidia uses different methods to develop. The first is to integrate multiple chips together, which is 3D chip packaging."

The second method is NVLINK. Through this network connection technology, all chip functions can be integrated into a huge chip. When all chips can be integrated into a huge chip, the performance will be very powerful. Third, Nvidia uses mechanical systems and liquid cooling systems to place the entire system in a server cabinet. Of course, there is also the most important software, that is, on these systems, there is Nvidia's software architecture and algorithms.

He said that through the above methods, Nvidia reconstructed the AI model, of course, especially the network part, which no longer needs to transmit information back and forth. In Nvidia's model, information can be transmitted directly within the network.

He emphasized that currently Nvidia can double its growth rate every six months, but in the future it may only double its growth rate every three months. Therefore, everything from chips, systems, data centers to operating systems must be recreated. Therefore, in the future, only the sky can limit the development of AI.
 
Xebec said:
Lip-Bu Tan is certainly unlike any other Intel CEO ever. Communications style, experience, leadership methods, and industry connections.
Click to expand...
They had it before they shot themselves in the foot 😅
Xebec said:
This is probably what Intel needed more than anything else -- should be interesting to see how Intel moves forward. I just hope the Board doesn't slow him down too much.
Click to expand...
agreed
 
You must log in or register to reply here.
Back
Top