Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/nvidia-x-mediatek-%E2%80%9Cai-pc%E2%80%9D-soc-to-debut-by-h2-2025-will-witness-integration-by-lenovo-dell-hp-asus.21186/
        )

    [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
)

NVIDIA x MediaTek “AI PC” SoC To Debut By H2 2025, Will Witness Integration By Lenovo, Dell, HP & ASUS

fansink

fansink

Well-known member
NVIDIA & MediaTek are set to unveil their "highly-anticipated" AI PC chip by the second half of 2025 and will be adopted by major OEM players.

NVIDIA & MediaTek's Upcoming AI PC SoC Is Set To Change Market Dynamics, Challenging AMD, Intel & Qualcomm
Rumors surrounding a custom chip being made by MediaTek for the AI PC markets have been swirling for quite some time now.

With NVIDIA's inclination towards MediaTek in recent days, it was very evident that both firms were proceeding towards collaborating in the AI PC markets, likely creating a new chip. Leaker on Weibo, "Mobile phone chip expert," has now claimed that the NVIDIA x MediaTek AI PC chip is expected to be released by the second half of next year, and interestingly, the chip has already reached the tape-out stages.

NVIDIA has been setting itself up to enter the custom chip business, and what better way there is than to capitalize on the growing AI PC markets, which has taken the industry by storm. NVIDIA has hands-on when it comes to bringing its graphical capabilities onto a chip, and MediaTek, which is a renowned mobile chipmaker, will see its first venture into the PC industry with this chip, but it is undoubtedly going to be a competitive end-product, competing with the likes of Qualcomm's Snapdragon X Elite, and Intel's Lunar Lake mobile chips.

While we are currently unaware of the specifications, it is rumored that the chip will employ TSMC's 3nm process, based upon the ARM architecture, and NVIDIA is likely going to feature its next-gen iGPU onboard, which will take the standard of the upcoming SoC to new heights. NVIDIA and MediaTek are cooperating in the automotive segment as well, within the MediaTek Dimensity Auto lineup of chips, so it won't be wrong to say that the partnership of both firms has worked out quite well, and Team Green will make an impact over the AI PC markets.

Interestingly, the leak also mentions that firms like Lenovo, Dell, HP, and ASUS have shown interest in NVIDIA's new project, and we can expect them to integrate the SoC in their lineup immediately after launch, which shows how big of a venture this actually is for the AI PC markets.

Consumers will finally have the option to opt out of the conventional "vendors" in the industry, which is why NVIDIA x MediaTek has undoubtedly got us hyped up.


wccftech.com

NVIDIA x MediaTek "AI PC" SoC To Debut By H2 2025, Will Witness Integration By Lenovo, Dell, HP & ASUS

NVIDIA & MediaTek are set to unveil their highly-anticipated AI PC chip by the second half of 2025 and will be adopted by major OEM players.
wccftech.com wccftech.com
 
I've been wondering how Microsoft is going to deal with all of these proprietary and different NPUs. It looks like they're programmed through drivers, which may mean that every Windows AI PC vendor will need separate software support from Microsoft. This could mean that Microsoft is providing strict interface requirements that the NPU drivers have to meet (Microsoft is not publishing the documents generally), so at least the basic functionality is the same, but it still looks like a mess for MSFT QA and field support. So far I'm counting four distinct and different NPUs for Windows: Intel, AMD, QCOM, and MediaTek-Nvidia. Apple execs must be having a good laugh at the way Microsoft's Windows life is getting complicated.

It should also be interesting to see how Windows virtualization works on a Mac (e.g. Parallels) when local inference processing becomes more common.
 
blueone said:
I've been wondering how Microsoft is going to deal with all of these proprietary and different NPUs. It looks like they're programmed through drivers, which may mean that every Windows AI PC vendor will need separate software support from Microsoft. This could mean that Microsoft is providing strict interface requirements that the NPU drivers have to meet (Microsoft is not publishing the documents generally), so at least the basic functionality is the same, but it still looks like a mess for MSFT QA and field support. So far I'm counting four distinct and different NPUs for Windows: Intel, AMD, QCOM, and MediaTek-Nvidia. Apple execs must be having a good laugh at the way Microsoft's Windows life is getting complicated.

It should also be interesting to see how Windows virtualization works on a Mac (e.g. Parallels) when local inference processing becomes more common.
Click to expand...
Many people called NPU is a waste of silicon, :). Probably for two reasons: 1) it is almost impossible to access NPU as a devolper, 2) MSFT has not rolled out very useful AI PC features/apps yet.
 
kevin01 said:
Many people called NPU is a waste of silicon, :). Probably for two reasons: 1) it is almost impossible to access NPU as a devolper, 2) MSFT has not rolled out very useful AI PC features/apps yet.
Click to expand...
Microsoft has rolled out privacy nightmares with NPU so i am glad i don't have Copilot+ pcs
 
kevin01 said:
Many people called NPU is a waste of silicon, :). Probably for two reasons: 1) it is almost impossible to access NPU as a devolper, 2) MSFT has not rolled out very useful AI PC features/apps yet.
Click to expand...
Not a complete waste, apparently, at least for Apple. The claim is that Apple uses the Neural Engine cores for Siri, Face ID, image enhancement, and augmented reality with Vision Pro. If true and the benefits substantial, this could mean Apple is years ahead of the competition. I've been curious about details for a while, but haven't found any. Just marketing stuff so far.
 
blueone said:
Not a complete waste, apparently, at least for Apple. The claim is that Apple uses the Neural Engine cores for Siri, Face ID, image enhancement, and augmented reality with Vision Pro. If true and the benefits substantial, this could mean Apple is years ahead of the competition. I've been curious about details for a while, but haven't found any. Just marketing stuff so far.
Click to expand...

This is true. Multiple NPUs since we use our face to unlock the phone. Maybe before that when we used our thumb. Apple is definitely ahead of the game when it comes to NPUs.
 
Daniel Nenni said:
This is true. Multiple NPUs since we use our face to unlock the phone. Maybe before that when we used our thumb. Apple is definitely ahead of the game when it comes to NPUs.
Click to expand...
Windows uses Windows Hello with AI/IR sensor iirc it's not just Apple many companies uses AI for different things we use it's just they are busy chanting AI AI they don't tell us about the actual good applications we use in our day to day life like Phone Cameras
 
You must log in or register to reply here.
Back
Top