AI PC Processors: The future laptop CPU battle of AMD, Intel, and Qualcomm

Introduction

In the coming months, you'll be swamped with new laptop processors from AMD, Intel, and Qualcomm. Strix Point, Lunar Lake, x86 vs. Arm, integrated GPUs, NPUs for AI workloads, and on-package memory are just some of the things coming to laptops in the next few months.

This article will explain the different laptop processors from mostly AMD and Intel, but Qualcomm is entering the Copilot+ PC ecosystem with its Arm-based Snapdragon X series processors, which also need to be talked about.

AMD has already hit the ground running with its new Zen 5-based Ryzen AI 300 series "Strix Point" APUs, which feature heavily beefed-up RDNA 3.5-based integrated graphics, and a low TDP which is fantastic for thin-and-light, but still powerful laptops for both work and gaming. We can't forget the new XDNA 2-based NPU for AI workloads, providing the best AI PC experience so far.

Intel has its upcoming Core Ultra 200V series "Lunar Lake" processors launching in the coming months, with Lunar Lake being the first CPU in Intel history to be outsourced to TSMC. TSMC will make Intel's new Lunar Lake CPUs on its N3B process node, with 50% more GPU performance from the next-gen Xe2 "Battlemage" GPU, 4x the AI performance of Meteor Lake, and power efficiency for thin-and-light laptops.

Qualcomm was first to market with a Copilot+ PC with its Snapdragon X processors led by the Snapdragon X Elite, offering fantastic Arm-based PC performance with 45 TOPS for AI workloads, and a great mix of CPU and GPU power. However, its Arm-based architecture means you need to live inside of the Windows on Arm ecosystem, which is a dark, dark world compared to using regular x86-based Windows.

AMD has its benefits with Strix Point, Intel has its benefits with Lunar Lake, and Qualcomm has its ben.... uh, well, yeah. Strix Point joins Zen 5 CPU goodness with RDNA 3.5 GPU goodness with a low TDP and tons of design options with the ability to scale higher than 32GB of RAM limits on Lunar Lake, with Strix Point laptops rocking up to 64GB of RAM. Strix Halo is coming in early 2025, and its rumored to support a whopping 128GB of RAM.

• Lunar Lake = 32GB RAM supported
• Strix Point = 64GB RAM supported
• Strix Halo = 128GB RAM supported (rumor)

ASUS Zenbook S16 OLED laptop

	Today	Yesterday	7 days ago	30 days ago
	$1779.00	$1779.00	$1779.00	$1779.00	Buy
* Prices last scanned on 10/4/2024 at 12:41 am CDT - prices may not be accurate, click links above for the latest price. We may earn an affiliate commission from any sales.

AMD's Ryzen AI 300 "Strix Point" APUs

16

AMD's introduction of the Ryzen AI 300 series, codenamed "Strix Point," marks a significant step in AMD's moves towards a large market of Intel's for decades now: laptops. AMD hasn't had a great mobile processor, but its new Strix Point APUs change that considerably. They're an astonishingly good APU with an NPU that is ready for AI workloads of the future.

The Skinny on Strix Point

AMD has its new Ryzen AI 300 series "Strix Point" APUs made at TSMC on its 4nm process node, with an unknown transistor count. Inside, we've got up to 12 cores and 24 threads based on the new Zen 5 CPU architecture, a beefed-up RDNA 3.5-based Radeon 890M integrated GPU, an upgraded XDNA 2-based NPU for AI workloads, and a low TDP of up to 54W.

I've already done some preliminary testing on an ASUS Zenbook S16 powered by an AMD Ryzen AI 9 HX 370 "Strix Point" APU and have walked away rather impressed indeed, further impressed when the power consumption at full load is a mere 33W.

The upgraded RDNA 3.5-based Radeon 890M is impressive, while not beating a mid-range discrete graphics card, it provides ample amounts of performance for 1080p 60FPS+ in AAA gaming, while I could even play Overwatch 2 (with FSR 2.2 enabled, and details on low) at 90FPS+ which wasn't bad at all... more impressive when you consider -- once again -- we're only using 33W of power here.

16

TDP (Thermal Design Power)

The Ryzen AI 300 series, particularly models like the Ryzen AI 9 HX 370, are designed with varying TDPs to fit different laptop designs, from ultrathin to high-performance gaming laptops. You'll see most designs with a 28W to 33W TDP, like the ASUS Zenbook S16 with the Ryzen AI 9 HX 370, which uses 33W under full load. AMD has a 54W TDP on the flagship Ryzen AI 9 HX 370 APU, but it is configurable between 15W and 54W depending on the design.

Clock Speeds

Clock speeds for the Ryzen AI 300 series are competitive, with the Ryzen AI 9 HX 370 reaching up to a 5.1GHz maximum boost clock, showcasing AMD's focus on delivering high-frequency performance alongside efficiency.

16

Zen 5 Architecture

The Zen 5 architecture represents AMD's latest leap in CPU design:

• Core Count and Threads: Strix Point APUs feature up to 12 cores (a mix of high-performance Zen 5 cores and efficient Zen 5C cores), handling up to 24 threads, which is a significant upgrade from previous generations limited to eight cores in mobile segments.
• Performance Improvements: Zen 5 brings IPC (Instructions Per Cycle) improvements, better power efficiency, and enhanced performance per watt. Zen 5 aims to provide substantial gains in both single-threaded and multi-threaded workloads.
• AI and Machine Learning: With Zen 5, AMD has integrated the XDNA2 architecture, which is their third-generation neural processing unit (NPU), offering up to 50 TOPS (tera operations per second) of AI compute performance, significantly outpacing previous iterations and competitors in AI acceleration.

16

RDNA 3.5 Architecture

The integrated graphics in the Ryzen AI 300 series are based on the RDNA 3.5 architecture:

• Graphics Performance: Up to 16 compute units (CUs) in the top model, which is an increase from the previous 12 CUs, promising better gaming and content creation performance on integrated graphics.
• Efficiency and Features: RDNA 3.5 delivers large improvements in performance per watt, new rendering techniques, and possibly better support for modern graphics APIs and features tailored for gaming and professional applications.

16

Additional Features and Market Positioning

• AI Integration: The focus on AI with the XDNA2 NPU positions these APUs as leaders in AI-capable laptops, especially for applications leveraging Microsoft's Copilot+ features.
• Competitive Edge: AMD claims these chips outperform competitors like Intel's Meteor Lake and Qualcomm's Snapdragon X Elite in various benchmarks, particularly in AI and integrated graphics performance.
• Laptop Ecosystem: Manufacturers like MSI are already integrating these APUs into high-end laptops, with a big push into gaming, content creation, and professional use with Strix Point APUs.

AMD's new Ryzen AI 300 series "Strix Point" APUs mark a gigantic change for the company in the laptop market, which has been dominated for decades by Intel. AMD now has an APU on the market that can easily play 1080p 60FPS in AAA titles while using 28W. That's an incredible feat.

The NPU side of things means that any Strix Point laptop you buy will have AI workloads taken care of, while providing oodles of CPU and GPU performance in a thin-and-light laptop design.

Intel's Core Ultra 200V Series "Lunar Lake" CPUs

16

Intel's Lunar Lake, branded under the Core Ultra 200V series, represents Intel's latest endeavor in mobile computing, particularly aimed at enhancing AI capabilities and power efficiency. Lunar Lake is a move by Intel to better align itself with Apple and its M-series processors by having on-package memory.

Normally, you'd have a CPU -- no matter if it's a laptop or desktop CPU -- and RAM would be a separate part that is connected to the motherboard, with interconnects between it all. There's latency involved in that, so Intel's latest move is using on-package memory with Lunar Lake, 16GB or 32GB of RAM that sits right beside the CPU itself. This means the RAM is not upgradeable (a big con for some people) but makes for an exciting CPU release in the months ahead.

Here's an in-depth look at what these new CPUs bring to the table.

Semiconductor Process and Architecture

• Process Node: Intel is tapping TSMC and its N3B process node for Lunar Lake CPUs, with it being the first time in Intel history that it is outsourcing its consumer desktop CPU line.
• Architecture: Lunar Lake introduces the Lion Cove Performance cores and Skymont Efficient cores, a hybrid architecture designed to balance high performance with energy efficiency, similar to its desktop counterparts but optimized for mobile use.

TDP and Power Efficiency

• TDP Range: The series varies from 17W to 30W, with the flagship Core Ultra 9 288V model having both PL1 and PL2 set at 30W, showcasing Intel's focus on power efficiency for longer battery life on new Lunar Lake laptops.

16

Clock Speeds and Cores

• CPU cores, no Hyper-Threading: Intel will have no Hyper-Threading for its upcoming Lunar Lake CPUs, with its new Lion Cove and Skymont CPU architectures featuring no HT support. The CPU cores will be split between P-Cores for performance and E-Cores for efficiency.
• Clock Speeds: The Core Ultra 9 288V boasts up to 5.1GHz for P-cores and 3.7GHz for E-cores, indicating a significant emphasis on single-thread performance alongside efficient multi-threading capabilities with its 8-core setup (4 Performance + 4 Efficiency cores).

AI and NPU Capabilities

• AI Performance: Lunar Lake CPUs are set to deliver up to 48 TOPS, with the flagship model reaching even higher, aiming to meet and exceed Microsoft's Copilot+ standards for AI processing. This is boosted by an integrated NPU alongside XMX units for enhanced AI computations.

16

Integrated Graphics

• Battlemage GPU: Featuring the new Arc 140V and Arc 130V graphics, these GPUs are tailored for better gaming and content creation performance on mobile platforms. The Core Ultra 9 and 7 series get the more powerful Arc 140V, which provides a large 50% performance improvement in gaming compared to Arc Alchemist inside of Meteor Lake processors.

On-Package Memory

• On-package Memory: Lunar Lake processors come with integrated LPDDR5X memory, with options for 16GB or 32GB, reducing power consumption by minimizing data travel distance within the system. This design choice not only boosts efficiency but also performance by having memory closer to the CPU cores.

Market Positioning and Release

• Release Date: Intel has scheduled the launch for September 3, 2024, aiming to make a significant impact at the IFA 2024 in Berlin, showcasing their commitment to leading AI-capable laptop processors.
• Competitive Edge: With Lunar Lake, Intel aims to directly compete with Arm-based solutions like Qualcomm's Snapdragon X Elite and AMD's Ryzen AI 300 series, focusing on superior power efficiency and AI processing capabilities.

16

Conclusion

Intel's Lunar Lake processors under the Core Ultra 200V series are poised to set new benchmarks in mobile computing with their focus on AI, on-package memory, and a balanced approach to performance and efficiency. This move signifies Intel's strategic, if not Apple-like, pivot towards not just raw performance but smart, efficient computing made for the next generation of AI-driven applications and devices.

It will be exciting to see how a thin-and-light Intel Core Ultra 200V series "Lunar Lake" laptops will fare against thin-and-light AMD Ryzen AI 300 series "Strix Point" laptops. Strix Point systems are using an average of 33W to 35W under load, with my personal ASUS Zenbook S16 with an AMD Ryzen AI 9 HX 370 APU inside under full load consumes 33W of power, inside of a beautifully thin laptop. Impressive so far.

But, with a 50% performance uplift with its next-gen Xe2-based Battlemage GPU, upcoming Lunar Lake laptops are going to be a real nice proposition against AMD's new Strix Point APU-powered laptops hitting the market already.

Another thing that is exciting for Lunar Lake is that Intel is using on-package memory, which sits next to the CPU itself. Technologically, this is just awesome to see (and will have its own benefits) but the drawbacks are not being able to upgrade your RAM inside of any Intel Lunar Lake laptop, period. 16GB or 32GB, that's your option for RAM with Lunar Lake.

Qualcomm's Snapdragon X Processors

16

The Qualcomm Snapdragon X series represents Qualcomm's latest venture into high-performance computing for laptops, aiming to challenge the dominance of Intel, AMD, and Apple in the mobile and PC markets with its Arm-based architecture. Here's a detailed tech breakdown:

Architecture and Core Design

The Snapdragon X series introduces Qualcomm's new Oryon CPU cores, designed specifically for high-performance computing tasks. Unlike traditional Arm designs, which might differentiate cores for performance and efficiency, Qualcomm's approach with Oryon is to have all cores as "performance cores." This design philosophy aims to deliver high performance across the board, potentially reducing the need for complex core management seen in hybrid architectures.

16

• Snapdragon X Elite: Features up to 12 cores, with a maximum clock speed of 3.8 GHz, and can boost up to 4.2 GHz on select cores. This configuration is tailored for top-tier performance, competing directly with high-end laptop processors from Intel and Apple.
• Snapdragon X Plus: A slightly downscaled version with 10 cores, still aimed at performance but with a focus on power efficiency, suitable for a broader range of devices where battery life might be as critical as raw performance.

GPU and Graphics

The Snapdragon X series incorporates a Qualcomm Adreno GPU, which, according to Qualcomm, can offer up to 4.6 TFLOPS of performance in the X Elite model. This GPU performance is claimed to match or exceed that of some of the best-integrated graphics solutions from competitors while consuming significantly less power. This makes it particularly appealing for gaming on laptops where thermal and power constraints are significant.

AI and Machine Learning Capabilities

One of the standout features of the Snapdragon X processors is their AI acceleration. Each chip includes a Hexagon NPU (Neural Processing Unit) capable of up to 45 TOPS (tera operations per second). This NPU is designed to handle AI tasks locally, reducing dependency on cloud computing for AI-driven applications, which is crucial for privacy, latency, and offline AI processing.

Memory and Connectivity

• Memory: Supports up to 64GB of LPDDR5X memory, which not only increases the amount of data that can be processed but also does so at speeds up to 8448 MT/s, significantly reducing latency and increasing bandwidth.
• Connectivity: The Snapdragon X series integrates the Snapdragon X65 5G Modem-RF System, promising faster and more reliable 5G connectivity. This is complemented by support for Wi-Fi 6, 6E, and 7, ensuring devices can leverage the latest wireless technologies for high-speed data transfer and low-latency communication.

16

Performance Claims and Real-World Expectations

Qualcomm has made bold claims regarding the performance of the Snapdragon X series, suggesting it can outperform or match the likes of Apple's M2 Max and Intel's Core i9-13980HX in certain benchmarks while consuming less power. However, real-world performance can differ due to various factors:

• Benchmarking: Initial benchmarks like Geekbench show promising results, with the Snapdragon X Elite outperforming in multi-threaded tasks against some of its competitors. However, single-threaded performance, often critical for many applications, shows a mixed bag where Apple's chips might still hold an edge.
• Efficiency: The power efficiency claims are significant, potentially offering longer battery life for similar or better performance, which could revolutionize mobile computing.
• Software Compatibility: While Windows on Arm has improved, not all software is optimized for Arm, resulting in most (literally) of your games NOT working on a Snapdragon X processor. Not just games but regular software as well. It will have to be an Arm version of the software, or it will be emulated, resulting in far worse performance than an AMD or Intel CPU.
• No review samples: Qualcomm didn't send out many review units of its Snapdragon X-powered Copilot+ AI PC laptops, and we're sure it's because they'd be torn apart.

Ecosystem and Adoption

The Snapdragon X series has already found its way into various devices from leading manufacturers, signaling strong industry support. This adoption is crucial for building an ecosystem where developers can optimize their software for these new processors, enhancing performance and compatibility over time.

16

Conclusion

The Snapdragon X processors represent Qualcomm's big push into the Copilot+ PC world with high-performance chips focusing on power efficiency, AI acceleration, and raw performance. Internal benchmarks are all good, but real-world performance is what consumers want. The Arm-based architecture is great for battery life, but most regular users avoid it. They have a place in the market, but they're being made out like they're regular, new AI PC laptops, but they're not, and we're making that clear to you.

Qualcomm has a compelling AI PC processor, but the Snapdragon X will pale in comparison to an AMD Strix Point or Intel Lunar Lake laptop in virtually everything. Maybe not crazy 20+ hour battery life, but from a standpoint where everything will just work, AMD and Intel will win that battle.

Final Thoughts

The laptop market is about to explode with options, starting with AMD's new Ryzen AI 300 series "Strix Point" APUs inside of laptops right now. Intel is weeks away from unveiling its new Core Ultra 200V series "Lunar Lake" processors, finding homes in new thin-and-light laptop designs.

AMD and Intel will have laptop processor rivals that are almost equal between Strix Point and Lunar Lake, with Qualcomm not even in the same league as it's using an Arm-based architecture. I have found my experience having on the Strix Point APU to be rather surprising, with the APU capable of running games like Black Myth: Wukong on high graphics at 1080p 30FPS, impressive for 33W.

16

Intel's entry into the market with its upcoming Core Ultra 200V series "Lunar Lake" processors will shake up the laptop market again, and then we're riding directly into CES 2025 which is in January, only 4 months away now. We should see even more new laptop designs, mixed with NVIDIA's new GeForce RTX 50 series Laptop GPUs, mixed with AMD and Intel's new processors.

You'll be safe in AMD's new Ryzen AI 300 series or Intel's new Core Ultra 200V series CPUs, just be mindful of anything powered by Qualcomm if you're a gamer or use a particular set of software that wouldn't run, or would be too slow to use, on a Snapdragon laptop.