AMD has published HIPRT 3.2, the latest version of its ray tracing library for HIP, and the release notes name the integrated graphics of its upcoming Zen 6 "Medusa" processors outright. Under new features, AMD lists Medusa support for three GPU architecture IDs: gfx1170, gfx1171, and gfx115E. HIPRT is mostly used by rendering apps like Blender, so this says nothing about gaming performance. Still, the link between these IDs and Medusa has so far rested on compiler patches and leaks, and now the Medusa name sits right in AMD's own changelog.

The gfx117x IDs belong to RDNA 4m, which first showed up in LLVM back in February. AMD added GFX1171 and GFX1172 targets in March, merged GFX1171 into Mesa 26.3 in August, and pushed RDNA 4m firmware later that month. Despite the name, RDNA 4m sits in the GFX11 family with RDNA 3 and 3.5, plus a few RDNA 4 style additions like FP8/BF8 conversion and new matrix instructions. The gfx115E entry is the more curious one. GFX115x covers the RDNA 3.5 graphics in Strix Point and Strix Halo, so it looks like at least one Medusa chip could stick with RDNA 3.5. The same release fixes a bug that truncated architecture numbers containing a letter, gfx115E included.

AMD is bringing FSR 4.1 to even more graphics cards and gaming platforms; in a major announcement today by Jack Huynh, SVP & GM, Computing & Graphics, the company announced that it is bringing FSR 4.1 to Radeon RX 6000 series graphics cards powered by the RDNA 2 graphics architecture, which date back to 2020. AMD has developed and optimized FSR 4.1 for the entire range of RX 6000 series GPUs, from the top RX 6950 XT, down to the entry-level RX 6400.

Next up, the company announced that it is bringing FSR 4.1 to the integrated graphics solutions of its APUs, spanning all Ryzen AI Max "Strix Halo" and "Gorgon Halo" APUs, Ryzen AI 9 "Strix Point" and "Gorgon Point" mobile processors. FSR 4.1 is also coming to the Ryzen Z2 series processors for gaming handhelds. Lastly, the company announced that it has worked with Microsoft to bring FSR 4.1 to Xbox Series X and Xbox Series S consoles.

A report from Hardware Luxx suggested that AMD does not currently plan to bring its upcoming FSR 4.1 upscaling technology to RDNA 3.5-based integrated graphics. If accurate this would leave Ryzen AI 300 and Ryzen AI 400 series APUs without support for the latest iteration of FSR. Back in May AMD previously confirmed that FSR 4.1 which utilizes INT8 processing, will arrive for discrete RDNA 3 (Radeon RX 7000) GPUs in July, while RDNA 2 (Radeon RX 6000) desktop graphics are scheduled to receive support by early 2027. This exclusion affects silicon families that are using the RDNA 3.5 graphics architecture. More precisely Strix Point, Strix Halo, Krackan Point and upcoming refreshes like Gorgon Point and Gorgon Halo. These APU configurations integrate high-performance iGPUs such as the Radeon 890M and the top-tier Radeon 8060S, both capable of running the required INT8 operations.

However, AMD's Frank Azor has since pushed back on reports that RDNA 3.5 has been ruled out. Responding on X, Azor stated that "no such decision as being reported and implied here has been made," while adding that AMD is not ready to discuss future product plans at this time. While this does not confirm FSR 4.1 support for RDNA 3.5 graphics, it suggests that AMD has not made a final decision regarding the architecture's inclusion in the technology roadmap.

AMD is preparing to launch its "Medusa Point" APU in early 2027. However, more benchmarks are emerging to showcase what the actual SoC can do as AMD and its OEM partners test the chip. In the latest Geekbench v6 run, AMD's 10-core, 20-thread "Zen 6" chip appeared with the AMD Engineering Sample number 100-000001713-21_N, achieving a 2,300 single-core and 13,002 multicore score while officially running at only a 2.4 GHz base frequency. In real-world operation, it ran within the range of 2.0-2.1 GHz during the benchmark. The most surprising factor is that this "Medusa Point" test system can match a 10-core, 20-thread AMD Ryzen AI 9 365 "Strix Point" APU that operates at more than double the frequency. When comparing the two, "Medusa Point" scores slightly lower in single-core performance, while the multicore score is surprisingly higher.

This phenomenon could be attributed to the fact that the "Zen 6" cores in the "Medusa Point" APU are much better performing and more optimized for the workloads that Geekbench tests. The IPC improvement target from "Zen 5" in "Strix Point" to the newest "Zen 6" could be a high single-digit to low double-digit gain on average. It is likely that the combination of new instructions and IPC improvements is what is pushing "Medusa Point" so high. Since the new APU also appeared in firmware running AVX-VNNI in FP16 precision, we might be seeing these workloads getting accelerated thanks to the lower precision of the floating-point operations. For now, the situation remains a mystery, at least until more benchmarks are available in the coming months. AMD is expected to launch this new APU around CES 2027, so we still have a lot of time before official and third-party benchmarks are released.

A fresh leak from Geekbench has surfaced listing an unannounced AMD processor with the OPN (Ordering Part Number) code 100-000001713-31. As for the platform, that is listed as Plum-MDS1, a somewhat direct hint to Medusa Point, AMD's next-gen mobile APU family based on Zen 6. The chip is a 10-core, 20-thread with a 2.4 GHz base clock, though actual test results show it running closer to 1.3-2 GHz, unsurprising for an engineering sample this early in development. Each core gets 1 MB of L2 cache, and the L3 comes in at 32 MB, up from 24 MB on Strix Point and 16 MB on Hawk Point. This is 50% more cache than the current 10-core parts like the Ryzen AI 9 365. The test system had 32 GB of memory installed, type unspecified. Benchmark numbers aren't worth reading into at this stage, as the chip spent most of the test hovering around 1.39 GHz and barely peeked above 2 GHz, well below what a final retail part on a 3 nm process would run at. Too early to draw any conclusions from the scores.

Shipping manifests from Planet3DNow linked to the "Medusa" codename suggest a 4C+4D layout, standard and density-optimized cores, though that doesn't exactly explain the 10-core count seen in Geekbench. One theory is that two additional low-power cores sit in the IO die, bringing the total to ten. The part is expected to be a 28 W TDP mobile chip for the FP10 socket. Medusa Point is shaping up to combine Zen 6 CPU cores with a mix of RDNA 5 and RDNA 3.5 graphics, plus an updated NPU. A launch around CES 2027 fits the AMD usual cadence, meaning there is likely a long wait ahead; however, this Geekbench entry confirms that testing is already happening, whether at AMD itself or at one of its early hardware partners.

AMD announced its Ryzen AI 400 series "Gorgon Point" mobile APUs based on Zen 5 and Zen 5c cores at CES 2026 earlier this year, simultaneously showing off the AM5-based Ryzen AI 400-branded desktop APUs at the same show. Now, according to a new leak by Moore's Law is Dead on X, AMD will allegedly launch the new Ryzen AI 400-series desktop processors for AM5 in the first half of 2026. MLID adds that the launch may even happen as early as Q1. He also released what he claims to be a short promotional video of an AMD Ryzen AI 400 Pro processor in an AM5 socket to lend credence to his leak.

Exact specifications for the AM5 APUs are unconfirmed, but if they're anything like the laptop counterparts, it is expected that the new APUs will launch with the same silicon as Strix Point but with adjusted power management and increased clock speeds. This means we can expect to see the Ryzen AI 400 desktop processors top out with a 12-core (4× Zen 5; 8 × Zen 5c) CPU cluster tied to a 16 CU RDNA 3.5 iGPU, and a 60 TOPS NPU.

GPD, a well-known manufacturer of mini PCs with unique form factors, showcased its entire lineup of systems powered by AMD "Strix Halo" and "Strix Point" at CES 2026. Leading the lineup is the GPD WIN 5, a handheld device featuring the AMD Ryzen AI Max+ 395 processor. It offers up to 128 GB of LPDDR5X system memory running at 8,000 MT/s, supporting the 16-core/32-thread CPU and Radeon 8060S integrated graphics. The system provides up to 4 TB of SSD storage, with ports including USB4, USB 3.2 Gen 2 Type-C, USB 3.2 Gen 2 Type-A, and slots for microSD and Mini SSD for additional storage. It features a 7-inch 1080p display and an 80 Wh battery.

After much anticipation, Intel has finally unveiled its "Panther Lake" Core Ultra Series 3 mobile processors, built on the 18A node. This new mobile platform aims to maximize performance across various areas, with a significant focus on integrated Arc graphics. The "Panther Lake" features a new Arc B390 GPU with 12 Xe cores, delivering up to 77% more performance compared to the previous generation "Lunar Lake" Arc 140V iGPU. Intel's performance claims are supported by comparisons to non-Intel graphics, such as the AMD Radeon 890M. In these comparisons, Intel's solution outperforms AMD's "Strix Point" iGPU by about 73% with upscaling technologies and is 82% faster at native resolution. For handheld devices relying on iGPU performance, Intel's "Panther Lake" appears to be the leader in gaming, at least for the time being.

Interestingly, Intel is also developing a separate handheld-focused "Panther Lake" SoC, likely featuring a new Arc B SKU, rumored to be named Arc B380. This version is expected to be more power-efficient, possibly through special binning of the dies, maintaining the full 12 Xe cores configuration at slightly reduced clock speeds to achieve similar performance with a much lower TDP. According to Intel, many handheld manufacturers, including Acer, MSI, and GDP, are showing interest in this SoC for their designs. However, we can also expect some full laptop SoCs to be used in handheld devices as well. Intel also claims that the new Arc B390 is comparable to NVIDIA's GeForce RTX 4050 laptop, which is a discrete chip.

AMD's not-yet-launched "Ryzen AI 400 series" seems to be nearing finalized form, given an uptick of leaks across the latter half of 2025. Yesterday morning, an ASUS Vivobook S 15 (M5650GA) test platform posted further "Gorgon Point" details within the Geekbench browser database. The sort of next-gen ultra-thin laptop was driven by a Ryzen AI 9 465 APU. This particular SKU also turned up via a CrossMark entry, late last month. Looking at fresh Geekbench 6.5 scores, the evaluated ASUS device achieved overall 2780 (single-core) and 12001 (multicore) digits. When compared to a direct predecessor, the unreleased mobile chip only holds an advantage in single-core stakes.

It is possible that the Ryzen AI 9 465 mobile processor was put through its paces under unfavorable conditions—e.g. running on immature drivers and firmware. The direct forebear, Team Red's readily available Ryzen AI 9 365 "Strix Point" mobile processor holds a 474 point advantage in multicore scenarios. Industry observers reckon that the "Gorgon Point" generation will deploy with improved onboard XDNA 2 NPUs, albeit combined with familiar "Strix Point" credentials. The forthcoming Ryzen AI 9 465 SKU seems to share its predecessor's base specifications: 10 cores/20 threads, maximum 5.0 GHz boost/turbo clock, and a Radeon 880M iGPU. Amusingly, Geekbench has identified the leaked part as belonging to the "Strix Point" mobile family—it is widely believed that "Gorgon Point" will emerge as a mild refresh over the "previous" generation.

AMD is gradually releasing the AGESA PI 1.2.8.0 BIOS update, which primarily aims to enhance memory compatibility. However, users who have installed this BIOS version are reportedly facing issues. A member of the Chiphell forum installed the latest AGESA PI 1.2.8.0, hoping for improved system optimizations and memory tuning, only to encounter significant problems. The user is unable to access the CMOS screen, as the system crashes immediately upon each attempt. Furthermore, the computer fails to boot into the operating system, and the system boot repair utility consistently crashes when used.

The distribution of the update appears to have been limited and quietly handled by some manufacturers. MSI published BIOS files for several X870E and B850 models that reference AGESA PI v1.2.8.0, and an AMI BIOS image labelled 7E59v1A65 with a release date of 2025-12-03 and a file size near 15.9 MB lists improved memory compatibility in its release notes. Community tracking shows the new BIOS available only for a few SKUs, including the MPG X870E Edge Ti WiFi, MPG X870E Carbon WiFi and MAG B850M Mortar.

Qualcomm is ready to push its Snapdragon X2 Elite and Elite Extreme SoCs into a sector that was previously dominated by x86 SoC makers, and that is gaming. The company now claims that its SoC can run more than 90% of the most played games available on the market on the initial launch. However, this is not the only striking claim, as Qualcomm has benchmarks that prove that its Adreno X2 iGPU is 50% faster than the iGPU inside Intel Core Ultra 9 288V "Lunar Lake" and 29% faster than the iGPU on AMD Ryzen AI 9 HX 370 "Strix Point," under the same graphics settings. This is of course, no small win, and Qualcomm expects to win hearts of gamers with its next-generation platform.

The new chip delivers frame rates up to 2.3x higher than its predecessor while supporting many popular Windows games. Its architecture uses four processing blocks that give the flagship X2 Elite Extreme an edge over competing integrated graphics. Each block efficiently processes graphics work before passing it to shader processors organized around two micro-units. These units each contain 128 KB of memory feeding 128 ALUs. The design supports multiple precision formats, employs a more efficient execution model, and can process two instructions at once to maximize its 128-ALU groups. In total, the GPU has eight shader processors and 2,048 ALUs running at up to 1.85 GHz, backed by 21 MB of on-chip memory, 2 MB of cache, and 228 GB/s of memory bandwidth on the flagship model.

AMD is gearing up to refresh its "Strix Point" APUs with the upcoming "Gorgon Point" generation, set for release in 2026. Recent online benchmarks have provided the first glimpse of this new APU. A submission to SiSoftware reveals an AMD Ryzen AI 9 HX 470 model. The "Gorgon Point" APU offers a boost in single-threaded performance while maintaining the same CPU core and iGPU architecture. It features up to 12 cores and 24 threads based on "Zen 5" or "Zen 5c," an iGPU based on the RDNA 3.5 graphics architecture, and an NPU based on XDNA 2. The cache configuration is listed as 12×1 MB of L2 and 3×16 MB of L3 cache. Although the core architecture remains largely unchanged, the CPU, iGPU, and NPU receive performance enhancements over the existing "Strix Point" chip.

According to leaked partner slides, the main difference is a slight increase in boost clock speeds for the new APU generation. In SiSoftware run, th Ryzen AI 9 HX 470 was spotted running 5.25 GHz boost frequency, which is a modest improvement from the previous 5.1 GHz, while maintaining a 28 W default TDP. The NPU throughput has been moderately increased to deliver over 55 AI TOPS. AMD has enabled full NPU performance across all tiers of Ryzen AI SKUs based on this chip. AMD's own benchmarks of the "Gorgon Point" APU are also shown below, as revealed by an earlier leak from a confidential presentation in China.

We recently reported on a rumor that AMD would soon be launching desktop APUs based on its Krackan Point architecture that has until now only been featured in mobile parts. At first, it seemed as though Strix Point and its more powerful AMD Radeon 890M iGPU would be limited to the mobile side of things, however a well-known leaker, @9550pro on X, has recently pointed out that AMD's AGESA update version 1.2.7.0 does make mention of Strix Point in its code. This observation is backed up by further evidence shown in a post by @xgfancz on X, who pointed out that the list of architectures in the latest 1.2.7.0 AGESA beta update makes clear reference to "STX," presumably "Strix Point," alongside "PHX," which would be Phoenix Point, AMD's current Ryzen 8000G series of desktop APUs.

All of this suggests that AMD Ryzen 9000G desktop APUs may launch with Strix Point architecture instead of being limited to Krackan Point, as was previously the assumption. Stepping up from Krackan Point to Strix Point would mean that iGPUs in the Ryzen 9000G series may feature up to 16 CUs and RDNA 3.5 instead of RDNA 3 in the current APUs, which could mean gaming and iGPU performance right around the level of the likes of the AMD Ryzen Z2 Extreme now making its way to swathes of gaming handhelds. The AMD Ryzen 9000G desktop APUs aren't expected to launch until closer to the end of this year, and even if they do, there's no telling if they will launch at the same time as what is essentially the 8000G refresh we have seen in rumors before.

We have seen previous rumors claim that AMD would launch Ryzen 9000G series APUs for its AM5 platform sometime soon, although there has been no official announcement from AMD, yet. Most recently, though, data miners combing through AMD's AGESA 1207 code have managed to find evidence of new Krackan Point desktop APUs. Krackan Point has been spotted in MSI BIOS updates with references to the 00B60Fxx microcode, and the changelog for ASUS X870 beta BIOS versions mentions "new APU support," which could only really be Ryzen 9000G. These new APUs are said to be based on Zen 5, but previous leaks said that they would instead be a rehash of the 8000G APUs, making them a Zen 4 refresh. At any rate, it will likely still be a while before launch, since this is the first time the code has been spotted in desktop BIOS versions.

Until now, Krackan Point APUs have been limited to mobile applications, with the AMD Ryzen AI 5 340 and Ryzen AI 7 350 both coming to laptops earlier this year. Until now AM5 owners looking for an APU have been limited to something like the AMD Ryzen 7 8700G, which launched all the way back in January 2024. If the laptop silicon is anything to go by, Krackan Point will bring AMD's RDNA 3.5 to the desktop, although with iGPUs falling short of the powerful Strix Halo 8060S. We will likely see APUs top out at the same Radeon 860M iGPU as the laptop variants, or potentially step up to the Radeon 890M iGPU, where the Ryzen 8000G APUs topped out at the 780M. The 890M would bring 256 more shaders into the mix for a decent performance uplift, although this seems to be a GPU configuration reserved for Strix Point. Current speculation also suggests that CPU core counts will be limited to eight and six cores for the Ryzen 7 and Ryzen 5 CPUs, respectively.

During the IFA 2025 in Berlin, AMD hosted a media discussion, where one of the topics was the ongoing x86 vs. Arm debate. AMD has argued that x86 processors, both their own and Intel's SKUs, can offer competitively long battery life in the notebook form factor while maintaining the decades-old x86 software ecosystem. In a now-deleted article from ComputerBase, AMD claimed that the value of the overall package drives power efficiency and battery life, not the instruction set behind the CPU cores. As seen with designs like Intel's "Lunar Lake" and AMD's "Strix Point," x86 designs can deliver significant performance in a power-efficient package, including the GPU and even memory. Below is AMD's quote for ComputerBase, machine translated to English.

AMDSpeaking of efficiency, the topic of Arm vs. x86 came up here, as it does every year. AMD, almost like Intel recently, declared that the myth that x86 can't be efficient was refuted last year. Both AMD Ryzen and Intel Core can offer extremely long runtimes in notebooks, while accessing the entire x86 ecosystem. Ultimately, the overall package doesn't offer an advantage for Arm.

A pair of leaked partner roadmaps shared by hardware leaker @momomo_us suggests modest change for laptop processors through 2026, followed by a clearer shift in 2027. The documents indicate that AMD will largely be on autopilot next year, with a Zen 5-based refresh called Gorgon Point, which appears to reuse the same core counts, the same 4 nm node, and most of the same NPU and graphics building blocks as Strix Point. That implies only a slight performance improvement, likely due to tighter tuning rather than a new architecture. As a result, high-end mobile parts, such as the Fire Range and the Strix Halo APU, may see little to no refresh in 2026. Zen 6 is coming in 2026 for desktop and server, but for the mobile segment, another year of waiting is on the table.

The leak becomes more interesting for 2027, where two Zen 6 families appear: Gator Range and Medusa Point. Gator Range is positioned as the future elite mobile family and is likely to replace Fire Range; however, the roadmap provides no in-depth technical details beyond the move to Zen 6. Medusa Point is described more concretely. It is said to utilize the FP10 package, be manufactured on a 3 nm process, and ship as a monolithic APU that combines different core types and features stronger iGPU. The listing mentions four full Zen 6 cores, four compact Zen 6c cores, two low-power cores, and roughly eight RDNA / UDNA EUs for on-chip graphics. If true, Medusa Point would target premium laptops with a balance of CPU compute and capable iGPU for gaming and local AI processing. As always with leaked roadmaps, expect changes and wait for official confirmations from AMD.

Minisforum today released the N5 Pro AI NAS barebone. The unit comes with five 3.5-inch caddies with SATA 6 Gbps backplanes, and provision for three M.2-2280 NVMe SSDs. The NAS also comes with a PCI-Express 4.0 x16 slot (electrical x8), so you can install a graphics card, a DPU, or an enterprise-grade NIC. In terms of its own network connectivity, the N5 Pro AI offers a 10 GbE interface, and a secondary 5 GbE interface. The N5 Pro AI NAS is powered by an AMD Ryzen AI 9 HX Pro 370 "Strix Point" processor with a 12-core/24-thread CPU, a powerful iGPU with 16 CU, and a 50 TOPS NPU. The NAS comes with two DDR5 SO-DIMM slots. The barebone variant lacks any installed RAM, but comes with a basic 128 GB NVMe SSD with the MinisCloud OS. You can opt to have 48 GB ECC DDR5 or even 96 GB ECC DDR5. There is also a value N5 AI NAS (non-Pro), which comes with an AMD Ryzen 7 255 processor, and 16 GB of DDR5 memory. Other interesting features include support for OCUlink, 40 Gbps USB4, and Docker single-key installation. The N5 Pro AI NAS starts at USD $1,019, while the N5 AI starts at $645.

AMD announced that it is now world's fastest in nearly all CPU form-factors. The company took Gold and Silver places in the latest Top500 supercomputers list, driven by a combination of 4th Gen EPYC "Genoa" CPUs and Instinct MI300A AI GPUs. It also has the world's most powerful AI PC processor with the Ryzen AI Max "Strix Halo" processor. With AMD enabling OEMs to combine the chip with up to 96 GB of unified 256-bit LPDDR5X memory, "Strix Halo" is the only PC processor capable of accelerating 128B-parameter AI models locally.

AMD has the world's fastest desktop processor for both gaming and productivity in the Ryzen 9 9950X3D "Granite Ridge-X", which is significantly faster than the Intel Core Ultra 9 285K in both gaming and productivity. It also has the world's fastest processor for laptops with the Ryzen 9 9955HX3D "Fire Range-X," which is essentially a mobile-friendly BGA package of the 9950X3D. With their high core counts, higher memory channel and PCIe lane counts, 5th Gen EPYC "Turin" processors dominate the server space in performance per socket. Ryzen semi-custom chips dominate gaming consoles. Ryzen Threadripper 9000 "Shimada Peak" is the undisputed king of HEDT and workstation processors. Perhaps the only form-factors AMD isn't dominating are mobile processors in the U- and H-segments, with the "Strix Point" silicon falling behind "Arrow Lake-U" and "Arrow Lake-H" in CPU performance.

It looks like AMD is refreshing its Ryzen 8000G desktop APUs as the newer Ryzen 9000G series, these chips will be powered by the older "Zen 4" microarchitecture, and not the latest "Zen 5," suggests a credible rumor on ChipHell. It was anticipated that the Ryzen 9000G series would be based on the 4 nm "Strix Point" monolithic silicon, or its updated variant, "Krackan Point." The chip features a dual-CCX CPU component with one CCX packing four "Zen 5" cores sharing 16 MB of L3 cache, and the second CCX featuring eight "Zen 5c" cores sharing an 8 MB L3 cache. The "Zen 5c" cores operate at lower average clock speeds than "Zen 5," and hence don't need large amounts of last-line cache. "Strix Point" also features a large RDNA 3.5 iGPU with 16 compute units, and most importantly, a powerful NPU with 50 AI TOPS, which meets Microsoft Copilot+ requirements. This is not going to be the case for the Ryzen 9000G, if the rumor is true.

If based on the older 4 nm "Hawk Point" silicon, the 9000G series will feature a single CCX CPU with eight full-size "Zen 4" cores sharing a 16 MB L3 cache; and an RDNA 3 based iGPU with 12 compute units, but unfortunately a much slower NPU with only 16 TOPS of throughput, which doesn't meet Copilot+ requirements. Perhaps the only things differentiating the 9000G series from the 8000G series are minor speed bumps, around 200 MHz increases in CPU clock speeds across the board, and a 100 MHz increase in iGPU engine clocks. All is not lost, as the non-socketed mobile Ryzen AI 300 "Strix Point" processor can be had in plenty of MoD (mobile-on-desktop) mini PCs, and we can see from recent product launches that PC OEMs are all over the Ryzen AI Max "Strix Halo" chip, putting it into prebuilt desktops.

AMD today introduced the Ryzen AI 5 330, a mainstream mobile processor model for Microsoft Copilot+ ready notebooks at mainstream price-points. The processor comes with a slim CPU complex, but a full-featured NPU to locally accelerate Copilot+ and other edge AI use-cases. The processor comes with a configurable TDP of 15 W to 28 W, making it fit into most mainstream notebook form-factors with minimal cooling needs.

As for specs, the Ryzen AI 5 330 is based on a cut down variant of the 4 nm "Strix Point" monolithic silicon. Its CPU complex is 4-core/8-thread, although AMD didn't specify what the mix of "Zen 5" and "Zen 5c" cores is. It comes with 12 MB of "total cache" (L2+L3), which means the L3 cache is 8 MB, given that both the "Zen 5" and "Zen 5c" core types come with 1 MB L2 caches. The CPU cores are clocked at 2.00 GHz, with 4.50 GHz maximum boost frequency. Meanwhile, the iGPU model is dubbed Radeon 820M, and is configured with 2 compute units for 128 stream processors. The NPU, however, is the fully fledged XDNA 2 unit with 50 TOPS. We expect notebooks powered by the Ryzen AI 5 330 to be priced well under the $500 mark.

Sapphire brought unique new hardware to Computex 2025. We begin our tour with the new Sapphire Edge AI line of mini PCs. First up, is the Edge AI 370, a mini PC measuring 117 mm x 111 mm x 30 mm (WxDxH). It is powered by AMD Ryzen AI 9 HX 370 "Strix Point" mobile processor, with two DDR5 SO-DIMM slots supporting up to 96 GB of memory. Storage comes from a 1 TB Gen 4 NVMe SSD. The CPU is configured with 12 cores (4x Zen 5 + 8x Zen 5c), while the maxed out RDNA 3.5 iGPU has 16 CU. The XDNA 2 NPU is clocked for 50 AI TOPS, and meets Microsoft Copilot+ requirements.

Next up, is the Sapphire X870EA Wi-Fi PhantomLink Edition motherboard. This is Sapphire's first high-end motherboard in years. The Socket AM5 motherboard is based on the flagship AMD X870E chipset, and offers premium connectivity and I/O features. It's also designed to visually match the company's latest Radeon RX 9070 XT NITRO+ graphics card, but its most striking feature is PhantomLink. This is a backside power delivery feature similar to ASUS BTF. The main PCI-Express 5.0 x16 slot has a trailing power delivery slot that relays power from a 12V-2x6 power input located next to the 24-pin ATX connector on the motherboard. This may not be a complete backside I/O motherboard, but at least keeps the graphics card free from any power cables sticking out. To use PhantomLink, however, you need a compatible graphics card, and Sapphire showed us the Radeon RX 9070 XT NITRO+ PhantomLink Edition.

Zotac is set to unveil the next generation of its Gaming Zone handheld at Computex in Taiwan later this month, following its initial entry into the portable gaming PC market last summer. This upcoming device will feature the powerful AMD Ryzen AI 9 HX 370 processor and, surprisingly, will run on a custom Manjaro Linux-based operating system specifically designed for handheld devices as reported by Liliputing. This represents a departure from using Windows or adopting Valve's SteamOS, which powers the popular Steam Deck.

The next-gen Zotac Gaming Zone will boast a 7-inch OLED display with a 120 Hz refresh rate. Inside, the Ryzen AI 9 HX 370 Strix Point chip makes a big leap in performance compared to the first version (the first-gen Zotac Gaming Zone featured an AMD Ryzen 7 8840U Hawk Point processor with an 8-core Zen 4 CPU and 12-core RDNA 3 graphics). It has 12 Zen 5 CPU cores and 16 RDNA 3.5 GPU compute units. It also comes with a better NPU to handle AI tasks. Other features include LPDDR5x memory, a PCIe 4.0 NVMe SSD slot (M.2 2280), Wi-Fi 6E and Bluetooth 5.2, two USB4 ports and a 48.5Wh battery.

One of the interesting long-term trends in the commercial market is the shift in how people are functionally using PCs. Over the past few years, AI has gone from a dinner party conversation piece to practical use cases. Hand-waved discussions of future benefits are now concrete benefits in the here-and-now, from writing and validating software code to shaping customer communication. While many of these early AI workloads were executed in the cloud, we are seeing new use cases that require new local PC capabilities - capabilities AMD has been working to develop as part of its larger AI PC initiative. We've seen AI deployment accelerate in PCs, particularly as Microsoft builds these functions right into the operating system via its Windows 11 Copilot+ PC program.

In just two years, we've moved from an environment where the NPU was an unsupported, unknown proposition, to integrating NPU support into first and third-party applications. Commercial OEMs are expected to introduce AI PCs and Copilot+ PCs at a variety of price points throughout 2025, making local AI support more available and affordable. That's an important factor for corporate customers, many of which will be simultaneously negotiating the end of Windows 10 support and a historically large, pandemic-related system refresh wave. AI presents a further variable. The nature of the AI question has evolved over the last couple of years from "Does AI have a future in the enterprise?" to "What are the best practices for enterprise AI and AI PC deployment?" That's a significant shift, and it's worth unpacking in a little more detail.

PC hardware watchers continue to pore over official AMD repositories and adjacent databases, in the hopes of finding unannounced next-gen technologies. Olrak29 and InstLatX64 have presented their latest Team Red-related findings; apparently reaching across futuristic desktop, mobile, and workstation product families. As outlined and interpreted by VideoCardz, several of these next-gen branches are already somewhat "known" properties—namely AMD's allegedly Zen 5-based Ryzen Threadripper "Shimada Peak" 9000WX (workstation) processor series. Following almost two years of leaks, an official introduction is expected to happen during Computex 2025. The Ryzen 9000G "Gorgon Point" desktop (Zen 5 + RDNA 3.5) APU series has turned up again; now "fully" linked to the AM5 socket platform (not a big surprise). The two leakers have also uncovered another rumored AM5-bound product lineup—"Grado" chips could be based on existing "Granite Ridge" foundations, but elevated to commercial/enterprise levels. These speculated basic/entry-level "EPYC 4005" processors are floated as natural successors to currently available 4004 forebears (related to Ryzen 7000 "Raphael" architecture).

Olrak29 and InstLatX64 have also found multiple mysterious FP8 socket-related Ryzen AI Mobile SoCs. "Krackan2" could be a cheaper refresh of current "Krackan Point" APUs—Tom's Hardware proposes smaller designs that sport fewer cores, and not configured with NPUs. Kepler_L2 has weighed in on the matter of three listed "Gorgon Point" IPs—he reckons that the third variant ("Gorgon Point3") will be a spin-off (aka refresh) of a "Krackan2" design. As suggested by insider knowledge, Team Red's convoluted scheme points to "Gorgon Point" being the sequel to "Strix Point." An FF5-based "Soundwave" processor design has appeared alongside the aforementioned futuristic Ryzen AI Mobile chipsets—industry whispers propose that AMD will be leveraging Arm architecture within a lower product tier. InstLatX64 pulled additional compelling information from AMD's Technical Information Portal—providing further insight into Ryzen AI "Medusa Point" APUs (Zen 6 + RDNA 3.5) being dreamt up, with a matching "larger footprint" FP10 platform.

AMD is preparing to launch as many as six new laptop discrete GPUs this Computex based on its latest RDNA 4 graphics architecture. The lineup will be built on its 4 nm "Navi 48" and "Navi 44" chips. It will be led by the Radeon RX 9080M, its new flagship part that maxes out the "Navi 48," enabling all 64 compute units for a core-configuration resembling that of the desktop RX 9070 XT. This rumor comes from AllTheWatts! on Twitter, a reliable source with AMD leaks. The RX 9080M gets 16 GB of memory across the chip's full 256-bit memory interface.

Positioned below the RX 9080M will be the RX 9070M XT, configured with 48 compute units for 3,072 stream processors, 96 AI accelerators, 48 RT accelerators, 192 TMUs, and possibly 96 ROPs. It gets 12 GB of memory across a 192-bit wide memory bus. This is essentially the same core-config as the upcoming desktop RX 9070 GRE. A rather huge notch below the RX 9070M, which is based on the physically smaller "Navi 44" silicon, and has 32 compute units for 2,048 stream processors, 64 AI accelerators, 32 RT accelerators, 128 TMUs, and 64 ROPs. This chip gets 8 GB of memory across a 128-bit memory bus.