Intel Lunar Lake chips might be delayed, while competing Snapdragon X and AMD Strix Point are ready to go

Daniel Sims

Posts: 1,895   +52
Staff
Why it matters: The AI PC race began this year with new hardware focusing on powerful NPUs from numerous hardware manufacturers. However, Intel might have fallen a step behind. While the company technically launched the trend with its Meteor Lake processor lineup late last year, Strix Point and Snapdragon X have since leapfrogged it, leaving Chipizilla playing catchup.

Anonymous insiders told DigiTimes that Intel has internally delayed shipments of its Lunar Lake laptop CPUs, tentatively named Intel Core Ultra 200, to September instead of June. The move would put Chipzilla's latest AI PC processors months behind competitors from Qualcomm and AMD. The three companies recently unveiled their latest laptop processors intended to meet Microsoft's new Copilot+ PC standards. Copilot+ PCs aim to leverage the new SoCs for onboard generative AI workloads.

Microsoft defines a Copilot+ PC as one with an NPU capable of at least 40 TOPS (trillion operations per second). Intel's Core Ultra 100 (Meteor Lake) series, available now in variants of some recently released Copilot-branded products, falls just shy of that at 34 TOPs.

Meanwhile, notebooks using Qualcomm's Arm-based Snapdragon X SoCs, which advertise 45 TOPS, recently beganshipping. Microsoft's Copilot+ AI software features are currently exclusive to Snapdragon but will come to other AI PC processors in a free update late this year or early next year.

Qualcomm's strong showing pushed AMD and Intel to accelerate their AI PC plans, and Team Red plans to make its Ryzen AI 300 (Strix Point) lineup, boasting 50 TOPS, available starting July. Lunar Lake aims to respond with at least 40 TOPS and other features like on-package memory.

The back-to-school shopping season is another potential reason the delay could put Intel at a disadvantage. Qualcomm and AMD already released computers with the most advanced options just in time for the late summer shopping season. Students and parents are already browsing selections for the upcoming school year. However, Intel's Core Ultra 200 series will get a second chance during the 2024 holiday shopping season.

Apple might also join the fray late this year and early next year. The Cupertino giant reportedly plans to launch multiple new Macbooks featuring the recently introduced M4 Apple Silicon. The M4 SoCs feature NPUs capable of 38 TOPS and are currently only available on the latest iPads. However, new MacOS and iOS updates in late 2024 will introduce on-device AI functionality similar to Microsoft's Copilot suite.

Permalink to story:

 
Funny how this site advertises Mediocre Lake as having 34 TOPS, which is in fact the total TOPS of cpu, npu and iGPU, it's npu is a poor 11 TOPS, but compares it to the npu only TOPS of Qualcomm Oryon and AMD Strix. The total TOPS of these is far higher. Let's compare apples to apples shall we and not show bias.
 
This is sloppy reporting. The digitimes article claims Lunar Lake was delayed from June, which is a made up date. The official date was/is September.
 
This is sloppy reporting. The digitimes article claims Lunar Lake was delayed from June, which is a made up date. The official date was/is September.
I think they are talking about shipments to oems, not the public launch date. This could simply mean two things: some laptops get delayed and/or we are getting a very limited launch with proper product stocks coming later.
 
Funny how this site advertises Mediocre Lake as having 34 TOPS, which is in fact the total TOPS of cpu, npu and iGPU, it's npu is a poor 11 TOPS, but compares it to the npu only TOPS of Qualcomm Oryon and AMD Strix. The total TOPS of these is far higher. Let's compare apples to apples shall we and not show bias.
Why are you fanboying, though? If you didn't care about it, you wouldn't have written a comment about it. Who cares about less than 20 TOPS difference between CPUs when GPU's can do 800+ TOPS?
 
Why are you fanboying, though? If you didn't care about it, you wouldn't have written a comment about it. Who cares about less than 20 TOPS difference between CPUs when GPU's can do 800+ TOPS?
We need real actual benchmarks, not just numbers on a paper.
 
Last edited:
Back