Note: This feature was first published on 7 January 2025.
Apart from desktop and high-performance mobile processors, AMD also announced its “Strix Halo” Ryzen AI Max series of laptop processors. These new APUs are targeted at the enthusiast gaming and AI workstation markets and on paper at least, look set to deliver some wicked performances – all within the confines of slim and light laptop designs.
During its CES 2025 keynote, AMD claims the Ryzen AI Max series, powered by the Zen 5 architecture, will have “the fastest integrated graphics solution” in the Windows ecosystem. At the heart of the new chips is a disruptive memory architecture that allows for incredible flexibility and performance, particularly for memory-intensive workloads like gaming and AI. AMD has not held back with the specifications, equipping the new mobile processors with up to 40 RDNA 3.5 cores and up to 128GB of shared memory, aimed at not just gaming but also content creation and artificial intelligence applications.
Image: AMD
Image: AMD
The flagship of the new series, the Ryzen AI Max+ 395, is a monster of an APU, boasting 16 CPU cores and 32 threads (sounds familiar?), alongside a 40-core RDNA 3.5 integrated GPU. This APU also integrates AMD’s XDNA 2 NPU AI engine, delivering up to 50 TOPS of AI performance. In terms of raw processing power, AMD claims that the Ryzen AI Max series will outperform Intel's flagship Lunar Lake Core Ultra 9 288V by 1.4 times in gaming performance and deliver up to 2.2 times better performance in AI workloads compared to NVIDIA's top-of-the-line GeForce RTX 4090 graphics card, all while consuming just a fraction of the power – an incredible 87% less, to be precise.
The Ryzen AI Max+ 395 is accompanied by three other processors, the Ryzen AI Max Pro 380, Ryzen AI Max 385 and Ryzen AI Max 390 in varying configurations. See the specifications below.
Image: AMD
While AMD’s performance benchmarks are certainly eye-catching, particularly the claimed 1.4x gaming performance boost over Intel’s best, it is important to remember that these tests come from synthetic benchmarks, and we’ll have to do some real-world testing to see these processors’ actual performance. Gamers, however, will likely appreciate the fact that AMD's new chips are geared towards providing great frame rates at lower power consumption.
For content creators, the Ryzen AI Max series has some serious firepower for rendering tasks. AMD’s benchmark tests show that the Ryzen AI Max+ 395 outperforms the Apple MacBook M4 Pro by up to 84% in rendering workloads, thanks to its higher core counts and significantly more powerful integrated GPU. It also outpaces Intel’s Core Ultra 200V processors in 3D rendering by up to 2.6x on average. However, in the multi-threaded Cinebench 2024 test, the flagship Ryzen AI Max+ 395 shows a more modest advantage, outperforming Apple’s 12-core M4 Pro by just 2%, and falling slightly behind the 14-core M4 Pro by 3%.
Image: AMD
Image: AMD
Image: AMD
One of the key differentiators of the Ryzen AI Max series is the integrated memory architecture. AMD has designed the AI Max chips to use a unified coherent memory pool, with the ability to carve out dedicated memory for the GPU, allowing for up to 96GB of system memory to be allocated solely to graphics. This innovative memory management system eliminates the need for costly memory copies and dramatically improves memory bandwidth, leading to improved performance in AI workloads, gaming, and content creation. The Ryzen AI Max chips also support up to 128GB of total system memory, shared across the CPU, GPU, and AI engines, allowing all components to access memory as needed.
When it comes to AI performance, AMD compared the AI Max+ 395 to a desktop NVIDIA GeForce RTX 4090, and during the presentation showed how the Ryzen processor could outperform the RTX 4090 in AI token processing by up to 2.2x, while consuming just 55W of power compared to the 450W TDP of the RTX 4090. There’s some exceptional power efficiency being showed off here, especially in AI workloads. While AMD did not share specific benchmarks for its integrated XDNA 2 NPU AI engine, the company’s claims of up to 50 TOPS of AI processing performance are certainly intriguing, and this could be a key selling point for professionals working in fields such as AI, machine learning, and data analytics.
The Ryzen AI Max series isn’t just about raw performance; it’s also designed to be really efficient at power management. While it has a base TDP of 55W, the Ryzen AI Max chips can also be configured to run at up to 120W, allowing for more aggressive performance tuning in laptops with advanced cooling systems. This gives the chips an unprecedented versatility in a wide range of, from ultra-slim laptops to powerful mini workstations. The Core Ultra 288V in comparison, has a max TDP of 37W.
Image: AMD
AMD’s Ryzen AI Max processors will begin appearing in laptops and workstations starting from H1 2025 onwards. Brands like HP and ASUS are already planning to release laptops powered by the new chips, including the ZBook Ultra G1 compact workstation and the ROG Flow Z13 gaming tablet with a detachable keyboard. These systems are expected to offer an impressive mix of gaming performance and AI capabilities in a slim form factor, showing just how far AMD has come in terms of integrating powerful components into laptops without compromising on portability.
But there's more.
Image: AMD
AMD also announced the new Ryzen AI 200 and 300 series chips, which are aimed to make powerful AI capabilities more accessible to both consumer and commercial markets, as well as more affordable options for users looking to balance performance and cost.
We have already seen the top-of-the-line Ryzen AI 9 HX 370 and Ryzen AI 9 365, both of which appeared high-end in laptops last year (read about it in our ASUS Zenbook S 16 review). Following that release, AMD is expanding its Ryzen AI 300 range further with four new chips. These include models designed for both consumers and commercial clients. At the same time, AMD is rolling out the Ryzen 200 series, which offers a more cost-effective entry point for those who don't require the full power of the 300-series chips.
While the Ryzen AI 300 series is designed with AI integration in mind, featuring 50 TOPS of NPU (Neural Processing Unit) power, the latest additions bring slightly slower clock speeds and reduced core counts compared to their higher-end counterparts. Nevertheless, the 300 series processors still benefit from long-lasting battery life (AMD claims up to 24 hours), making them ideal for those who need a balance between performance and longevity. Like the Ryzen AI Max, these chips are powered by AMD's Zen 5 architecture, although AMD has been tight-lipped about the finer details of the graphics specs at launch.
Image: AMD
On the other hand, the Ryzen AI 200 series chips are designed for those seeking more affordable laptop and seems to go toe-to-toe with Qualcomm’s recently announced “affordable” Snapdragon X processors. With a focus on value, the Ryzen 200 chips pack in 16 TOPS of NPU power, which still provides solid AI capabilities but at a more accessible price point. These chips rely on AMD’s previous-generation Zen 4 cores and RDNA 3 graphics.
AMD is also catering to business users with Ryzen Pro versions of both the Ryzen AI 200 and Ryzen AI 300 processors. These professional-grade models will offer the same core features as their consumer counterparts but with additional enterprise-level support and security features. While the consumer versions of the new processors are expected to ship in the first quarter of 2025, business customers can expect the Pro models to arrive a few months later, in the second quarter.
We are only in the first month of 2025, but the laptop landscape already looks set to be a key battleground for AMD, Intel and Qualcomm.
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.