AMD’s Hardware AI Ryzen ‘Phoenix’ CPUs Set to Soar in Razer’s Blade 14 Laptop

Red Team: Your wait for AMD’s “Phoenix” to rise is almost over. AMD first pulled the curtain off its new Phoenix line of Ryzen 7000 laptop processors back at CES 2023, and now laptops with these chips are almost here. AMD recently gave us the rundown of its Ryzen 7040HS Series, which will appear in cutting-edge laptops soon, notably the latest refresh of the Razer Blade 14.

The main differentiator between the Phoenix chips and their fellow Ryzen 7000 “Dragon Range” line is the presence of hardware-based AI acceleration. Read on for the CPU speeds and feeds, as well as the explanation of this onboard AI capability.

AMD Ryzen 7040HS: ‘Phoenix’ Soon Takes Flight

Before getting into the fancy AI features, let’s go back to CPU basics with a refresher of this announcement. The 7040 Series chips are built on AMD’s latest architecture, “Zen 4,” which has shown tremendous performance on both desktops and laptops. 

The few Ryzen 7000 systems we’ve seen so far have been impressive, as shown in our Dragon Range testing (linked above), but we haven’t yet tested any laptops yet with the Phoenix class. You can read more about the shared Ryzen 7000 architecture here, and you can also find a rundown of AMD’s recently revised chip-naming conventions for better digestion of the CPU number salad.

AMD’s Ryzen 7040 HS line comprises three chips: the Ryzen 5 7640HS, the Ryzen 7 7840HS, and the Ryzen 9 7940HS…

During a media briefing, AMD detailed the advantages and performance expectations of these processors. The chip maker claims to see an average 13% improvement on instructions per cycle (IPC) with Zen 4 over Zen 3, thanks to the larger cache, superior branch prediction, and a more efficient queue. AMD also ran through several benchmark samples, some of which you’ll see in these images, but we won’t run through all specific claims. Of course, we plan to test these chips ourselves when they become available. 

(Credit: AMD)

AMD also showed moderate frame-rate gains while gaming, including while playing on battery. These chips are built on RDNA 3 graphics architecture; the Ryzen 9 and Ryzen 7 CPUs include Radeon 780M integrated graphics (IGP), and the Ryzen 5 chip features the Radeon 760M IGP.

(Credit: AMD)

Of course, plenty of the laptops using these chips will include their own discrete graphics, but these IGPs are still relevant for when the discrete GPU isn’t in use, especially for saving battery when unplugged. Such laptops will include gaming and content-creation machines from Acer, Asus, HP, Lenovo, and Razer.

AMD’s ‘Ryzen AI’ Advantages

AI is the hot topic of the day, but we’re aware of the distinction between your favorite chatbot and support for AI features built into chip hardware. Phoenix chips include the world’s first dedicated AI engine on an x86 processor, dubbed Ryzen AI, no small feat for a laptop chip, which will support a range of applications and even Windows features.

AMD outlined how modern neural network solutions differ from what we traditionally know as AI from the last few decades—hard-coded bots, programmed by humans. Neural networks share data between their artificial neurons, training the AI by machine learning. Ryzen AI uses local distributed memory that has higher bandwidth and requires less capacity, to offload AI tasks onto for improved efficiency.

(Credit: AMD)

The generalized advantages of local AI mean you enjoy zero latency versus online data transfers, less time waiting for the cloud, and increased security (since your data is not using the cloud). Better still, your local content won’t be assimilated into a network model, and long-term cloud and subscription costs are reduced and eliminated.

Likely unknown to many users, Windows already touts a few built-in AI features, but you need a system equipped with an AI engine to view them. Some of these, in particular, are especially useful on laptops. For example, you’ll find AI-based camera improvements like background blur, automatic framing, and eye-contact-correction options. If you’re using your webcam for frequent video calls, these will be appreciated, and can be applied universally rather than per-application.

At launch, AMD is working with Microsoft and a smaller group of developers—you’ll find support for ONNX and TensorFlow model formats. The roadmap for 2023 is that, by Q4, support will be added for PyTorch, and the features will extend to XDNA. So for now, the actual application of the AI features will be the developer zone.

The Ryzen-Ready Razer Blade 14

I mentioned a list of partner systems that would make use of these chips, and Razer was on hand during the meeting, too. The laptop maker presented a full rundown on its latest Blade 14, which will be one of the flagship laptops utilizing this new silicon. You’ll find several models of the portable Blade 14, all of which will run on the most-powerful Ryzen 9 7940HS CPU despite its compact size.

(Credit: AMD; Razer)

The starting price is $2,399—definitely a premium product for enthusiasts—and ramps up to $2,799 for higher configurations. That base model nets you the Ryzen 9 chip, a 140W-configured GeForce RTX 4060 GPU, 16GB of memory, and a 1TB SSD. The top model comes with a 140W RTX 4070 and 32GB of memory, and you can order it in black for $2,699.99 or white for the top $2,799.99.

Razer’s tiny powerhouse also features a 16:10 QHD+ display with a 240Hz refresh rate, vapor chamber cooling, and expandable DDR5 memory up to 64GB. As always with a Blade system, this is packaged into a sleek all-metal build with an RGB keyboard. The new Blade 14 is just 0.71 inch thick and weighs 4.05 pounds. Pre-orders are open today, and it will be available on June 20. We’re looking forward to testing this unit, as well as other Phoenix-based machines, as they come available in the coming weeks.