AMD's Madhu Rangarajan, corporate VP of Compute and Enterprise AI, says the company never set out to build its Zen 6 EPYC 9006 "Venice" chips around agentic AI, a not-so-subtle dig at NVIDIA and Arm, which both pitch their latest server CPUs as made for that exact job. Rangarajan still called EPYC 9006 the best CPU for agentic workloads during a media session covered by Tom's Hardware. Chief Zen architect Mike Clark said much the same in a separate talk, admitting AMD didn't see agentic computing coming, though he added that the company always expected the CPU to keep a role in AI software. Here's what Rangarajan said:

We think that the EPYC 9006 CPU is the best CPU for the agentic era, and spoiler alert, we didn't magically know four years ago that agentic was going to be a thing and purpose-built something for agentic. The reality is an agentic workflow is a diverse set of tasks… and in order to execute those tasks well, you need compute that is able to handle a diverse set of workloads with a full portfolio.

AMD's Zen 6 just got one of its first rough per-clock comparisons against Zen 5, and the early numbers look good. Leaker HXL spotted a Geekbench 7.1.0 listing for the 256-core EPYC 9996, the flagship "Venice" chip AMD launched in July, scoring 2,847 points in single-core on a Dell PowerEdge R9815 server running Linux. User three twenty six on X then locked a Ryzen 7 9700X to 4.1 GHz, matching the EPYC 9996's maximum boost, and got 2,444 points on the same Geekbench version. That works out to roughly 16.5% higher single-core performance for Zen 6 at the same frequency. Apparently, Zen 6 came out ahead in all 15 single-core subtests, with gains ranging from 8.2% in File Compression to 25.7% in HTML5 Browser.

Of course, these scores only give us a general idea about the gains. The EPYC 9996 uses the dense Zen 6c cores, not the full Zen 6 cores expected in desktop Ryzen chips, so it's hard to say how closely these numbers will carry over. The two systems are also wildly different, with a server platform on one side and a 9700X running DDR5-6000 CL34 under CachyOS Linux on the other, and the EPYC listing has since been pulled from the Geekbench database. Still, a double-digit per-clock gain from the dense core variant is quite encouraging. AMD has already confirmed that "Olympic Ridge" desktop CPUs will stay on AM5, though they reportedly won't arrive until 2027, so we can expect more comprehensive performance leaks well before then.

It seems as though AMD is readying a new mobile APU that bears a striking resemblance to the existing Ryzen Z2 Extreme. The AMD Ryzen AI 7 HX 455 was recently leaked by Golden Pig Upgrade on Weibo, with the post suggesting that the new CPU would launch with eight cores and 16 threads and a 16 CU iGPU, while the 16 CU iGPU will reportedly be identical to the existing Radeon 890M on the Ryzen AI Z2 Extreme, which has proven itself capable of very decent gaming performance in our in-depth review of the ASUS ROG Xbox Ally X.

The exact core breakdown is unknown, but in a response to a question about the CCD layout, the leaker alludes to a similar layout to the Z2 Extreme, which features three Zen 5 cores and five Zen 5c cores across two CCDs, although the new APU will have a significantly larger cache to work with—24 MB L3 instead of the Z2 Extreme's 16 MB—signalling that it will be a cut-back version of the existing Ryzen AI HX 470 APU, which features four Zen 5 cores and eight Zen 5c cores.

Today, AMD is introducing the next AMD Ryzen AI Halo Developer Platform, powered by the AMD Ryzen AI Max+ PRO 495 processor.

With up to 192 GB of unified memory, support for 300-billion-parameter and larger models, and a curated software experience, the new platform gives developers and creators the freedom to build, experiment, and run more AI locally. Choose your models, frameworks, operating system, and workflows, then put them to work on your terms.

AMD may be preparing a significant update to its Ryzen Z-Series processors designed for handheld devices, with the latest Z3 processor bringing substantial changes over the previous generation. At the core of everything will be the CPU architecture. For the standard Ryzen Z3, AMD plans to use four "Zen 6" cores and two "Zen 5" low-power cores. The Ryzen Z3 Extreme will share the same CPU core configuration, resulting in a total of six cores and twelve threads in the CPU cluster. Alongside this, there will be a cluster with an integrated graphics solution, featuring up to 12 Compute Units of RDNA 4m. This version reportedly derives from RDNA 3.5 but includes RDNA 4-style ALUs for matrix operations that power technologies like FSR 4.

Compared to the previous Ryzen Z2 Extreme's eight-core CPU cluster and 16 CU RDNA 3.5 integrated graphics cluster, the new solution focuses on extracting extra performance from the new "Zen 6" cores and optimizing for RDNA 4m. This ensures the software stack is mostly new, without needing any backporting to older RDNA generations. The difference between the two Z3 variants will be in power limits, with the Z3 catering to the 15-watt segment, while the Z3 Extreme will be optimized for a 25-watt power envelope.

AMD launched the Gorgon Halo Ryzen AI Max+ Pro 495 in May 2026, and the powerful new APU seems to finally be making its way into the wider mini PC and laptop market, with Colorful being the latest brand to launch new workstations based on the new AMD silicon. Colorful has launched two new systems under its Colorfire brand: the Lingchuang Mini Pro mini PC, with the AMD Ryzen AI Max+ Pro 495, and the Ryzen AI Max+ 395-powered Lingchuang K16, a new mobile workstation laptop.

Colorful hasn't revealed much about the new mini PC other than that it can be configured with up to 196 GB of LPDDR5X-8533 unified memory, 160 GB of which can be assigned to VRAM to feed the Radeon 8065S. Pricing has not yet been announced, but configurations will start at 96 GB for the Lingchuang Mini Pro. The Lingchuang K16, meanwhile, has been listed on Chinese storefront, JD.com, with a price of ¥34,999, which converts to around $5,218 at the time of writing.

Acemagic showcased an interesting solution at its booth at IFA 2026 in Berlin, featuring the latest mini-workstation equipped with AMD's top-tier Ryzen AI Max+ PRO 495 "Gorgon Halo" APU. Housed in a compact two-liter chassis measuring 158.5 x 158.5 x 81.5 mm, it boasts AMD's 16-core, 32-thread "Zen 5" CPU, capable of boosting up to 5.2 GHz, alongside Radeon 8065S integrated graphics. The system supports up to 192 GB of LPDDR5X memory running at 8,533 MT/s, which is more than sufficient for local processing of large language models, gaming, or creative work. Local file storage is managed by an M.2 2280 PCIe 4.0 x4 NVMe slot, supporting SSDs up to 4 TB. The APU also includes an NPU with a performance of 55 TOPS in INT8, bringing the total system power to about 131 TOPS at full speed.

For connectivity, this mini-workstation features two USB4 ports, dual 2.5 GbE Ethernet ports, the latest Wi-Fi 7, Bluetooth 5.4, and display outputs via HDMI 2.1 and DisplayPort 2.1. Additionally, there is an OCuLink interface with PCIe 4.0 x4 lanes for potential expansion if the user requires more GPU power or simply a PCIe connection on the outside.

At IFA 2026, GEEKOM revealed that the company has successfully interconnected multiple miniature PCs into a single local supercomputer capable of running inference with ease. At the company booth, GEEKOM demonstrated how it linked four of its A9 Mega mini PCs into a local inference cluster. The GEEKOM Private AI setup uses USB4 to connect these four mini PCs directly, eliminating the need for a dedicated rack or proprietary networking hardware. The result is essentially a miniature supercomputer that can sit on a desk, keeping AI workloads and data entirely local.

Each node is based on AMD's Ryzen AI Max+ 395, featuring 16 "Zen 5" CPU cores with integrated Radeon 8060S graphics and up to 128 GB of unified LPDDR5X memory running at 8,000 MT/s. Storage is provided by two PCIe 4.0 x4 M.2 slots, supporting drives up to 4 TB. The compact 171 x 171.1 x 70.1 mm system also includes dual 2.5 GbE, Wi-Fi 7, Bluetooth 5.4, USB4, and HDMI 2.1 connectivity. The processor has a 140 W TDP and an NPU rated at up to 50 TOPS, which means the total mini rack system will draw less than 1,000 W at peak load, making this setup extremely power efficient.

At IFA 2026, GEEKOM showcased its latest GeekBook X16 Pro, designed to be a thin and light powerhouse. At the heart of this system is AMD's latest "Gorgon Point" APU. The showcased unit features an AMD Ryzen AI 9 HX 470, which includes four "Zen 5" cores and eight "Zen 5c" cores, totaling 12 cores and 24 threads. Paired with this CPU cluster is AMD Radeon 890M integrated graphics for a solid experience in 3D rendering, gaming, and local AI performance. The system includes 32 GB (8 GB x 4) of soldered SK hynix LPDDR5 RAM running at 8533 MT/s and a 2 TB Lexar SSD.

The entire laptop body is crafted from aerospace-grade magnesium alloy in a precision-milled CNC unibody design. It hosts a 16-inch 2560x1600 IPS screen with a 120 Hz refresh rate, 100% sRGB coverage, and integrated eye protection features. Inside, the system is powered by a 77 Wh high-capacity battery, providing up to 17 hours of standby time, while charging is handled by a 65 W GaN PD Fast Charging adapter. Wired connectivity includes a single 3.5 mm audio jack, a USB 3.2 Gen 1 Type A, a USB 3.2 Type C, a USB4 running at 40 Gbps transfer speed, and an HDMI 2.1 port. Wireless connectivity features Wi-Fi 6E and Bluetooth 5.4, while audio is managed by stereo speakers with DTS:X Ultra Audio.

The opening keynote at IFA 2026 was held by AMD and our team on the ground was there to see it. During the event, AMD announced the Threadripper Halo Station, a liquid-cooled workstation built around the 96-core Threadripper PRO 9995WX CPU codenamed "Shimada Peak" and up to four AMD Instinct MI350P accelerators. The system is designed around one specific use case: run AI models with more than one trillion parameters locally, without a cloud connection. Jack Huynh, AMD's SVP and GM of Computing and Graphics, described it as a new class of workstation bringing supercomputer-class compute to individual users and developers. The Threadripper PRO 9995WX Zen 5 CPU has 96 cores and 192 threads, 5,4 GHz boost frequency and supports up to 2 TB of DDR5 system memory.

Going further down the specifications, each MI350P card features 144 GB of HBM3E running at 4 TB/s, offering a bandwidth 14 times higher than any LPDDR5X variant. At IFA 2026, AMD presented a configuration based on two MI350P PCIe cards for a total of 288 GB of HBM3E memory. Users can upgrade the system to four cards and 576 GB of HBM3E. This amount of accelerator memory is enough, says AMD, to hold a trillion-parameter model entirely on-device. Each Instinct MI350P card has a TBP of up to 600 W and is liquid-cooled independently, as is the CPU. Complete system specifications, availability and pricing have not been announced.

AMD's update to the Strix Halo Ryzen AI Max+ 395 is the Gorgon Halo Ryzen AI Max+ Pro 495, which launched with 16 Zen 5 cores, a total of 32 threads, and AMD's Radeon 8065S iGPU. So far, Framework has been the biggest OEM to tease the new Halo APU, but it seems as though GMKtec will soon launch a new Gorgon Halo mini PC in the Evo-X5 Pro.

GMKtec's official X account teased the new mini PC, which the brand says will launch at IFA Berlin 2026, on September 5. Aside from this, not much is known about the mini PC itself, however, the GMKtec Evo-X3 launched not too long ago, stacked to the gills with pro-tier ports, a beefy cooling solution with a 140 W peak TDP, and triple cooling fans. If the Evo-X5 doesn't stray too far from its predecessor, we can expect an OCuLink port, HDMI 2.1, at least one USB4 port, and 2.5 G LAN—or faster—packed into an aluminium chassis. We can also expect to see the price of the mini PC climb to the $4,000-range, especially for the top-end variant with 192 GB of RAM.

We are thrilled to officially unveil our next-generation compact, high-performance workstation: the Acemagic F9A (Ryzen AI MAX+395). Designed as a 2L-class mini AI workstation, it combines the flagship AMD Ryzen AI platform with Oculink expansion, dual 40 Gbps USB4 connectivity, and a 4-unit digital microphone array—all housed within a studio-inspired, CNC-machined aluminium unibody chassis. Packing desktop-grade workstation performance into a footprint the size of two water bottles, this product is tailored for AI developers, video creators, and professionals deploying local AI large models, meeting their needs for powerful performance and a clutter-free workspace.

AMD's flagship desktop processor until the Ryzen 9 9950X3D2 Dual Edition upstaged it, the 9950X3D, is now selling in the US at an all-time low price of $569. This amounts to an 18.5% drop from its launch price of $699.99. The $569 price can be found on both Amazon (US) and Newegg. The 9950X3D is a 16-core/32-thread Socket AM5 desktop processor based on the "Zen 5" microarchitecture, and powered by 3D V-Cache technology on one of its two 8-core chiplets.

The 9950X3D2 Dual Edition features 3D V-Cache on both its 8-core chiplets. In our testing, the 9950X3D was found to be about 3% slower than the 9950X3D2 at gaming (1080p, averaged), and about 2.5% slower averaged across all our application benchmarks; but it holds significant leads in our Physics Simulation, console emulation, AI ML, and rendering benchmarks.

Intel's next-generation Core Ultra Series 4 "Nova Lake" processors will mark an official return of AVX-512 to the company's client processor lineup. The 512-bit SIMD instruction set will be featured on both the "Coyote Cove" P-cores, and "Arctic Wolf" E-cores, according to Jay Kihn. It is a requirement for both core types to feature an identical ISA, without which thread migration between the two core types will result in runtime errors arising from ISA mismatch.

The last Core processor to feature an official implementation of AVX-512 was 11th Gen Core "Rocket Lake," in which AVX-512 workloads would notoriously cause the 14 nm processor to overheat. Intel would deprecate AVX-512 on all succeeding client microarchitectures, mainly because their respective E-core types didn't support it. Intel is expected to offer a client-relevant set of AVX-512 instructions. Meanwhile, AMD introduced AVX-512 support with "Zen 4," and would go on to improve AVX-512 performance with "Zen 5." The reentry of AVX-512 on Core could advance client AI inferencing applications.

AMD, through its official documentation portal, has confirmed the arrival of the next-generation Ryzen Threadripper TR6 "Mustang Peak" platform, along with some intriguing details about the upcoming CPU lineup. Based on the "Zen 6" CPU architecture, AMD is incorporating the latest I/O standards for these workstation powerhouses, including PCIe 6.0 and DDR5. As the "Zen 6" CPU cores are manufactured on TSMC's 2 nm semiconductor node, we expect AMD to increase the core count over its previous "Zen 5" Ryzen Threadripper workstation platform and deliver a significant performance boost due to advancements in node and microarchitecture. While specific details like TDP, clock speed, core counts, and segmentation are still lacking, we have learned that AMD is retiring its TR5 socket in favor of the new TR6 socket that will accommodate the platform.

With the introduction of PCIe 6.0, which doubles the bandwidth of PCIe 5.0 per lane configuration, AMD needs to add more pins to handle the increased bandwidth. Additionally, with varying core counts and clock speeds, the CPU socket likely requires more electrical capacity to support the new platform, necessitating an overhaul. If readers recall, the previous TR5 socket used 4,844 pins. For the TR6 update, it is possible that the pin count could exceed that number.

According to an investigation by Ars Technica, AMD has quietly removed a memory encryption feature from its consumer-focused Ryzen processors after years of inclusion. This feature, known as Transparent Secure Memory Encryption (TSME), is used to encrypt a system's memory content and prevent possible physical exploits that unencrypted memory data could be vulnerable to. Reportedly, AMD introduced TSME about a decade ago and began shipping it across all of its processors, including regular Ryzen, professional Ryzen Pro, Threadripper, and EPYC solutions. However, a random test on a Ryzen 7 9700X CPU, which uses "Zen 5" IP, has highlighted the absence of TSME support in consumer chips. In various exchanges between AMD engineers and some security professionals, the investigation led to a single response from AMD stating that "TSME is a security feature only applied to PRO CPUs as part of AMD PRO Technologies."

AMD has reportedly removed the feature quietly without prior notice. Discovering that the feature is missing is incredibly difficult on Windows-based machines, but still possible on Linux-based OS installations. Under a Linux-based OS, checking for firmware security revealed a missing encrypted RAM flag, despite TSME being enabled in the BIOS previously. It turns out that starting with the AGESA 1.2.7.0 build and in future AGESA versions, AMD has systematically removed the feature from its consumer Ryzen processors. After further investigation and inquiries to AMD, there is no information on whether the AGESA firmware is disabling this feature or if newer CPUs will ship with TSME disabled at the silicon level. All we know so far is that the feature is no longer present after years of being included on the consumer side.

AMD's upcoming Ryzen 10000 series of desktop processors, codenamed "Olympic Ridge" and based on the "Zen 6" CPU architecture, are rumored to introduce some interesting platform changes. According to the latest leak from X account "Gotou_3rd," AMD's upcoming AM5 platform is likely to replace its integrated graphics processing unit (GPU) with an integrated Neural Processing Unit (NPU), marking a significant shift from the current setup. Since the "Zen 4" based "Raphael" processors, part of the Ryzen 7000 series, AMD has included integrated graphics solutions with its desktop CPUs. This trend continued with the "Zen 5" based "Granite Ridge" Ryzen 9000 series, which also featured an integrated GPU for basic graphics output and boot functionality.

However, with the next-generation Ryzen 10000 series, AMD seems to be reallocating the silicon area for different purposes, as the role of integrated GPUs in desktop systems diminishes. Most PC builders pair a dedicated GPU with their DIY or OEM PC builds, using integrated graphics mainly for debugging when the discrete GPU encounters issues. Since the silicon space in the CPU can be repurposed, AMD might switch to an NPU to enhance local AI capabilities and offload some AI tasks from the GPU to the CPU's integrated NPU. This could mean the platform might become Copilot+ AI PC certified if the NPU achieves over 40 TOPS of compute power.

AMD today formally relaunched the Ryzen 7 5800X3D Socket AM4 desktop processor with 3D V-Cache, as the Ryzen 7 5800X3D 10th Anniversary Edition. This special edition SKU features a full-featured 5800X3D processor with the same OPN and specifications as the original 5800X3D that was discontinued in 2023; but comes with special branding on the PIB package, and includes a sheet of Carbice Ice Pad—a high-performance thermal pad. The package still lacks an in-box cooling solution. Based on the "Zen 3" microarchitecture, the 8-core/16-thread chip boosts up to 4.50 GHz, and features 96 MB of L3 cache.

AMD claims that the Ryzen 7 5800X3D is the fastest gaming processor on platforms with DDR4 memory. To illustrate, the company compared its gaming performance with that of an Intel Core i9-14900K "Raptor Lake Refresh," with both machines featuring 32 GB (2× 16 GB) of DDR4-3600 memory. The 5800X3D is shown posting an average of 10% higher gaming performance than the i9-14900K (DDR4), making the case for 5800X3D as the fastest gaming processor with DDR4 memory. Available from June 25, the Ryzen 7 5800X3D 10th Anniversary Edition is priced at $350. Next up, the company formally launched the Ryzen 7 7700X3D, a very interestingly priced gaming PC processor powered by the older "Zen 4" microarchitecture.

AMD in its Computex 2026 presentation, celebrated 10 years of the Socket AM4 platform that kickstarted the company's long march to competitiveness with Intel in the desktop PC processor market, and its eventual domination. Socket AM4 supports the original "Zen" and "Zen+," across the Ryzen 1000 and Ryzen 2000; and would go on to support Ryzen 3000 "Zen 2," Ryzen 4000 "Zen 2" APUs, and the vastly successful Ryzen 5000 "Zen 3" series, at which point it had achieved performance leadership over Intel. The company formally launched its "latest" product for AM4, a fresh production run of the Ryzen 7 5800X3D, which is the fastest processor for gaming on the platform.

In 2022, AMD debuted Socket AM5 alongside the Ryzen 7000 series powered by the "Zen 4" architecture. At the time, the company promised a similar cadence of at least two processor generations on the socket, leading up to 2027 (5 years in the market). The company would go on to launch the Ryzen 8000 series APUs, and its current Ryzen 9000 series powered by the "Zen 5" architecture. Its latest product line sees the introduction of Ryzen AI 400 series desktop APUs. Without naming any future processor generation or architecture, AMD announced that AM5 will have platform longevity till 2029. This would mean the introduction of at least one more microarchitecture, "Zen 6," and the product lines derived from it. Since AMD didn't announce "Zen 6" at Computex, the company likely won't debut it in 2026, instead, hold on to the current lineup, and possibly tease or announce "Zen 6" in 2027, with product launches expected that year. This would align with market availability of "Zen 6" for the following two years going into 2029.

ZOTAC Technology, leader in innovation and high-performance hardware solutions, is joining COMPUTEX 2026, one of the largest, most anticipated technology exhibitions of the year. This year, ZOTAC celebrates its 20th Anniversary at the event by showcasing products both classic and new, as a nod to the Company's past and its outlook toward the future. As a member of the NVIDIA Partner Network (NPN) Program, the Company plans to illustrate its future product strategy: a full-range of computational offerings for every application, from Gaming, Local AI deployment, to Industrial and Enterprise-level high performance computing applications.

ZOTAC, as a brand historically focused on gaming, has come a long way since first starting to offer NVIDIA GeForce graphics cards, and was one of the first pioneers of the now ubiquitous Mini PC product category, known for novel designs that pushed compact computing to new levels. During the exhibition, ZOTAC will be looking back at the last 20 years, and present the Company's significant products that shaped the Company, while looking forward to new heights.

In addition to the Ryzen AI Halo, AMD today debuted the Ryzen AI Max 400 series of processors for AI development platforms. This is a refresh to the Ryzen AI Max 300 series "Strix Halo" processor. The chip combines up to 16 full-size "Zen 5" CPU cores with a large SoC die that has a powerful iGPU with 40 compute units powered by the RDNA 3.5 graphics architecture; a 50 TOPS-class NPU, and a 4-channel (256-bit wide) LPDDR5X unified memory interface. The series consists of three processor models, all three of these get AMD PRO features that rival Intel's vPro. There are consumer versions also planned for later this year.

Memory is the focus of this refresh. AMD updated the memory controllers on the "Strix Halo" SoC die to now support up to 192 GB of LPDDR5X memory. The previous Ryzen AI Max 300 series tops out at 128 GB. Here, users have the flexibility to manually partition the memory between the system and the iGPU, giving the latter up to 160 GB of video memory. AMD also took the opportunity to give the GPU a maximum boost frequency increase to 3.00 GHz, up from 2.90 GHz on the Ryzen AI Max 300 series; while the "Zen 5" CPU cores now boost up to 5.20 GHz, up from 5.10 GHz on the previous generation. The integrated NPU gets a 10% performance boost, now being capable of 55 TOPS.

AMD today released Ryzen AI Halo, a compact AI development computer designed by AMD to serve as a nearly full-stack machine for AI development, inferencing, and generative AI content creation. Ryzen AI Halo is designed to strike a unique balance of price, performance, and capabilities that rival Apple Mac Mini and NVIDIA DGX Spark. It is a compact computer that can be used as a standalone PC, or in small clusters talking to your workstation. An advantage it offers over DGX Spark and Mac Mini is that it's based on x86-64, and fully supports Windows. AMD is pricing the Ryzen AI Halo at $3,999. Pre-orders open in June 2026.

At the heart of the Ryzen AI Halo is the Ryzen AI Max+ 395 "Strix Halo" processor that combines 16 "Zen 5" CPU cores with an oversized iGPU powered by the RDNA 3.5 graphics architecture, featuring compute capabilities. There's also a nimble 50 TOPS NPU. Perhaps the most interesting aspect that brings this whole device together, is memory. The processor features a quad-channel (256-bit wide) LPDDR5x memory interface, and AMD has crammed 128 GB of memory into this box. Also helping things is a spacious 2 TB NVMe SSD. The large memory size, helps the processor run AI models with up to 200B parameters, with the right quantization. For reference, the current Mac Mini can only be configured with up to 64 GB of memory. Hardware is less than half the story, the bulk of AMD's engineering effort with the Ryzen AI Halo is on the software side. The company innovated a comprehensive software platform compatible with both Windows and Linux, which helps developers quickly deploy one or more Ryzen AI Halo boxes faster than it takes to set up DGX Spark, AMD says.

AMD has formally launched the EPYC 8005 series, codenamed "Sorano" after its announcement at MWC 2026 back in February, and now the full details are in. The new lineup succeeds the EPYC 8004 "Siena" family and sits between the high-end EPYC 9005 and the entry-level EPYC 4005, targeting single-socket low-power platforms with eight to 84 Zen 5 cores across a 95 W to 225 W TDP range. The flagship is the EPYC 8635P featuring 84 cores, 168 threads, 384 MB of L3 cache, and a 225 W TDP with a 1.6 GHz base clock. Compared to its 64-core EPYC 8004 predecessor, AMD claims 40% higher integer performance and 9.5% better performance per watt. Against Intel's 40-core Xeon 6716P-B at 235 W, it offers 2.1 times the core count (84 vs 40), a 10 W lower TDP (225 W vs 235 W), and 91% higher integer performance. AMD also claims 48% better integer performance per CPU watt per dollar versus Intel 72-core Xeon 6776P-B in a single-socket configuration.

The full lineup spans seven SKUs, from the 8-core EPYC 8025P at 95 W up to the 84-core 8635P at 225 W. All models are single-socket only and support six channels of DDR5-6400 ECC memory with up to 3 TB total capacity, 96 lanes of PCIe Gen 5, and full AVX-512 support. AMD is positioning the 8005 series around dense edge deployments and software-defined storage workloads. In those scenarios, core density, I/O bandwidth and power efficiency matter more than raw single-threaded performance.

AMD has announced six new Ryzen PRO 9000 Series processors for commercial desktops and workstations, with two of them bringing 3D V-Cache to the PRO lineup for the first time. All six are AM5 parts targeting managed business deployments rather than the consumer market. The top-tier model is the Ryzen 9 PRO 9965X3D featuring 16 Zen 5 cores, 32 threads, up to 5.5 GHz boost, 128 MB of L3 cache courtesy of 3D V-Cache, and a 170 W TDP. Think of it as the workstation equivalent of the Ryzen 9 9950X3D, just under the PRO branding with the associated manageability and security features. A non-X3D Ryzen 9 PRO 9965 sits alongside it with the same 16 cores and 170 W TDP but 64 MB of L3 cache. The 3D V-Cache also extends down to the Ryzen 7 tier with the Ryzen 7 PRO 9755X3D sporting eight cores, 16 threads, up to 5.2 GHz, 96 MB of L3 cache, and a 120 W TDP. Rounding out the lineup are the Ryzen 9 PRO 9955 (12 cores, 120 W), Ryzen 7 PRO 9755 (8 cores, 120 W), and Ryzen 5 PRO 9655 (six cores, 120 W), all with 64 MB or 32 MB of L3 cache depending on the model.

The new additions sit above the existing 65 W Ryzen PRO 9945, 9745, and 9645 parts, giving AMD a higher-power PRO tier for workstation builds that need more headroom. The latest additions to the AMD Ryzen PRO 9000 Series lineup are scheduled to launch in the second half of 2026, with systems based on the new processors expected to begin shipping during the same period. AMD also noted that Lenovo ThinkStation P4 workstation is currently planned for availability in Q3 2026.

AMD's upcoming APU refresh with the Ryzen AI Max 400 series is divided into "Gorgon Point" and "Gorgon Halo." Today, we see one of the first "Gorgon Halo" APUs appearing in online benchmark databases. The AMD Ryzen AI Max+ PRO 495 "Gorgon Halo" APU, featuring 16 cores and 32 threads based on the current "Zen 5" CPU architecture, has landed in the PassMark testing database. These cores can reach a boost frequency of up to 5.2 GHz, which is about a 100 MHz improvement over the current "Strix Halo" APU generation. Complementing the CPU setup is the RDNA 3.5 GPU architecture, now in the form of a Radeon 8065S, which appears to be an overclocked version of the current Radeon 8060S. This new Radeon 8065S iGPU runs at 3.0 GHz, while the current Radeon 8060S runs at about 2.9 GHz. No increase in cores is expected here, and the "Gorgon Halo's" integrated graphics should continue with the 40 RDNA 3.5 CUs.

In terms of performance, AMD has managed to achieve better efficiency thanks to the higher boost frequency. PassMark's comparisons now list the new Ryzen AI Max+ PRO 495 "Gorgon Halo" APU as about 4% ahead in multicore and about 3% in single-core benchmarks compared to the AMD Ryzen AI Max+ PRO 395 "Strix Halo" APU. Another significant aspect is the integrated memory configuration. With the previous "Strix Halo," the maximum memory configuration was 128 GB, while the latest "Gorgon Point" shows 192 GB of LPDDR5X memory, suggesting that AMD has updated its integrated memory controller to increase the maximum memory capacity.