In the face of competitor launches like Qualcomm’s Copilot+ PCs, AMD has unveiled its next-generation Ryzen AI 300 chips. These processors, built on the Zen 5 architecture and codenamed Strix Point, are specifically designed for laptops and prioritize AI performance over anything else. With Computex currently taking place, we’re witnessing a whole host of reveals, and will be throughout the week. Let’s see what AMD has in store for us with these new laptop chips.
AMD prioritizing AI
One of the major headlines from Computex 2024 is the Ryzen AI 300’s impressive AI capabilities. Thanks to its new XDNA2 architecture, these chips boast a whopping 50 TOPS (trillion operations per second) – a somewhat significant jump compared to the 45 TOPS offered by Qualcomm’s Snapdragon X family. Ryzen AI 300, codenamed Strix Point, goes even further, promising superior performance for 16-bit applications through a novel block FP16 data type.
While the specifics of how this works remain unclear, it is suggested that it will improve the ability to achieve faster processing speeds while maintaining the accuracy of the FP16 format. A possible game-changer for developers working with 16-bit on laptops?
Ryzen AI 300 uncertainties
Despite the impressive specs, there are some uncertainties surrounding the Ryzen AI 300. Unlike competitors, AMD hasn’t disclosed the total package TOPS figure yet, which combines the AI processing power of the CPU, GPU, and NPU. This makes it difficult to compare the overall AI performance with other similar laptop processors. Additionally, the headline-grabbing 50 TOPS rating only applies to the two high-end Ryzen AI 9 models announced, not the whole range.
Lower-tier Ryzen 5 and 7 variants might have less powerful NPUs, potentially falling below the 40 TOPS threshold required for Copilot+ certification. This compatibility issue is further compounded by the fact that Copilot+ optimization currently favors Qualcomm processors, meaning it won’t be available for Ryzen AI 300 chips on launch day. This situation is also likely to affect Intel’s Lunar Lake launch.
| Specs | Ryzen AI 9 HX 370 | Ryzen AI 9 365 |
|---|---|---|
| Cores / Threads | 12 / 24 | 10 / 20 |
| Max Boost Speed | 5.1 GHz | 5 GHz |
| L2 + L3 Cache | 36 MB | 34 MB |
| TOPS (NPU) | 50 | 50 |
| Graphics | Radeon 890M | Radeon 880M |
Naming scheme change
Another noteworthy change is AMD’s move towards a simplified naming scheme – they have decided on Ryzen AI 300 for the name. Additionally, the suffixes like U and H that previously indicated power consumption are no more. Instead, the Ryzen AI 300 series will offer a wider power range of 15W to 54W, with the actual power level depending on how the manufacturer implements the chip.
While products equipped with Ryzen AI 300 processors are expected to hit the market in July, the initial rollout will likely be limited. AMD hasn’t provided a timeline for the release of the rest of the Ryzen AI 300 family, leaving some questions unanswered about the availability of lower-tier models for budget-conscious consumers.
Overall, the Ryzen AI 300 series presents a compelling option focusing on AI performance and a simplified naming scheme. However, uncertainties around NPU power in lower models and compatibility with Copilot+ at launch might give potential buyers a pause for now.
