Customers evaluating AI infrastructure today rely on a combination of industry-standard benchmarks and real-world model performance metrics—such as those from Llama 3.1 405B, DeepSeek-R1, and other leading open-source models—to guide their GPU purchase decisions. At AMD, we believe that delivering value across both dimensions is essential to driving broader AI adoption and real-world deployment at scale. That's why we take a holistic approach—optimizing performance for rigorous industry benchmarks like MLperf while also enabling Day 0 support and rapid tuning for the models most widely used in production by our customers.

This strategy helps ensure AMD Instinct GPUs deliver not only strong, standardized performance, but also high-throughput, scalable AI inferencing across the latest generative and language models used by customers. We will explore how AMD's continued investment in benchmarking, open model enablement, software and ecosystem tools helps unlock greater value for customers—from MLPerf Inference 5.0 results to Llama 3.1 405B and DeepSeek-R1 performance, ROCm software advances, and beyond.

AMD and its board partners cleared "phase one" of RDNA 4 earlier on in March, with the launch of Radeon RX 9070 Series graphics cards. At the tail end of special introductory events, Team Red representatives—on both sides of the Pacific—teased a second quarter release of lower end Radeon RX 9060 Series models. A handful of AIBs have registered multiple custom Radeon RX 9060 XT 16 GB and 8 GB SKUs, so expectations have been set for an imminent arrival. A fresh insider leak suggests that AMD has something else Navi 48 GPU-related in the pipeline; possibly scheduled for launch before rumored Radeon RX 9060 XT cards. Earlier today, IT Home picked up on chatter regarding a mysterious Radeon RX 9070 GRE model. Apparently Zhongzheng Computer (note: machine translated name) issued an intriguing tidbit on its WeChat official account—the March 9 bulletin stated: "friends who don't have enough budget for Radeon RX 9070 XT can wait for RX 9070 GRE, which will have a better price-performance ratio. Radeon RX 9060 XT will have to wait for a while."

Based on this news, VideoCardz believes that Chinese market stock of custom Radeon RX 9070 16 GB (non-XT) cards was not topped up last week. Local sources have observed regional market conditions with almost zero availability—conjecture points to Team Red's local office "deliberately" paving the way for "gap-filling" Radeon RX 9070 GRE 12 GB options. As reported by TechPowerUp on multiple occasions, AMD's "GRE" (aka Golden Rabbit Edition) nomenclature debuted with their introduction of a Radeon 7900 GRE 16 GB model back in 2023—the Year of the Rabbit. This (now) very out-of-date naming scheme was revised earlier this year—with a modernized abbreviation of "Great Radeon Edition." Benchlife.info weighed in on rumors regarding a new-generation GRE package: "(it) uses the same Navi 48 die, that is, the RDNA 4 GPU architecture, as the Radeon RX 9070 XT and Radeon RX 9070 currently on sale, but the memory will be reduced to 12 GB and the memory interface will be 192-bit. Our sources have informed us that the Radeon RX 9070 GRE 12 GB is currently being planned by AIB partners and is ready to enter mass production." VideoCardz has kindly assembled a relevant comparison chart—see below. Naturally, these theorized specifications place the incoming GRE somewhere in between the already released Radeon RX 9070 16 GB cards, and a rumored Radeon RX 9060 XT class.

Microsoft is introducing live patch updates for Windows 11 Enterprise, version 24H2, that allow critical security fixes to be applied without interrupting users. These updates, known as hotpatches, are available for x64 devices running on AMD or Intel CPUs. Hotpatch updates are designed to install quickly and take effect immediately. Unlike standard monthly security updates that require a system restart, hotpatch updates provide instant protection against vulnerabilities while allowing users to continue working. This new process can reduce the number of restarts from twelve per year to just four. The update schedule follows a quarterly cycle. In January, April, July, and October, devices install a complete security update with new features and fixes that do require a restart. In the two months that follow each of these baseline updates, devices receive hotpatch updates that only include security fixes and do not need a reboot. This approach ensures that essential protections are applied quickly without impacting daily work.

To use hotpatch updates, organizations need a Microsoft subscription that includes Windows 11 Enterprise (or Windows 365 Enterprise) and devices running build 26100.2033 or later. These devices must also be managed using Microsoft Intune, where IT administrators can set up a hotpatch-enabled quality update policy. The Intune admin center automatically detects eligible devices and manages the update process. Hotpatch updates are currently available on Intel and AMD-powered devices. For Arm64 devices, hotpatch updates are still in public preview and require an extra configuration step: disabling CHPE support via a registry key or the upcoming DisableCHPE CSP. This update system represents a more efficient way to secure Windows client devices. By minimizing the need for restarts and delivering updates in a predictable, quarterly cycle, Microsoft aims to help organizations protect their systems with minimal disruption. We expect these live patches to trickle down to more Windows 11 versions, like Home and Pro editions.

Surface defects on the GPU die of a PowerColor Radeon RX 9070 XT Hellhound have been linked to excessive thermal hotspots, according to analysis following Igor's Lab's investigation. Microscopic examination revealed 1,934 pits across the silicon die surface despite proper thermal interface material application. The defects affect over one percent of the chip's surface area, with the largest measuring 12.59 µm in depth and 212.36 µm in diameter—significantly exceeding industry tolerances. These imperfections impede heat transfer, causing localized temperatures to reach 113°C, exceeding the 110°C threshold for RDNA GPUs and triggering thermal throttling. The defects appear to stem from flaws in the backgrinding process that thins silicon wafers for packaging. Improper process control creates surface irregularities that compromise thermal conductivity and structural integrity.

The discovery raises questions about the effectiveness of automated optical inspection systems in detecting subtle surface abnormalities. TSMC, the maker of these chips, uses optical recognition tools to see if any defects occurred. AMD stated: "We are aware of the reported issue and believe this to be an isolated incident. We are working with our partners and internal teams to understand the issue." The company maintains this is an isolated case, though the findings suggest potential gaps in quality control protocols. Users experiencing thermal issues with affected RX 9070 XT units should initiate the RMA process for replacement while AMD and PowerColor investigate whether these defects exist beyond the identified sample. We hope no further issues are present, and so far, PowerColor hasn't received any customer complaints. We are on the lookout for further situation development. We reviewed PowerColor's Radeon RX 9070 Hellhound, and found no issues on our unit, so this indeed remains an isolated case so far.

A small number of AMD Ryzen 7 9800X3D CPUs have had trouble starting up on some ASRock AM5 motherboards. The issue was first noticed in February 2025, and both AMD and ASRock immediately looked into the problem. The investigation found that the older BIOS versions on specific ASRock boards had memory compatibility problems. When a Ryzen 7 9800X3D was used, the computer could not complete its normal boot process, known as POST. ASRock quickly released a beta BIOS update to fix the memory issue, and after testing, a final version of the update was made available as of now. Users who update their BIOS should now see the problem disappear. In one case, a report suggested that a Ryzen 7 9800X3D had been damaged. However, further checks showed no burn marks on the motherboard. The issue was simply due to some debris, and after careful cleaning, the motherboard started normally and passed further tests.

AMD pointed out that many reasons can cause a computer to not start up properly, and a POST failure does not always mean the CPU is broken. They advise users to update their BIOS to the newest version available for their motherboard model. If the problem continues even after updating, users should reach out to customer support for help. In a few cases, if the problem cannot be fixed on the spot, the CPU might need to be replaced. This quick fix shows that AMD and ASRock work well together to solve problems. Even though only a few systems were affected, the issue was taken seriously and fixed fast. With the latest BIOS update, users can now expect their systems to boot without trouble, ensuring that the Ryzen 7 9800X3D offers good performance for everyday tasks and gaming.

Today, MLCommons announced new results for its industry-standard MLPerf Inference v5.0 benchmark suite, which delivers machine learning (ML) system performance benchmarking in an architecture-neutral, representative, and reproducible manner. The results highlight that the AI community is focusing much of its attention and efforts on generative AI scenarios, and that the combination of recent hardware and software advances optimized for generative AI have led to dramatic performance improvements over the past year.

The MLPerf Inference benchmark suite, which encompasses both datacenter and edge systems, is designed to measure how quickly systems can run AI and ML models across a variety of workloads. The open-source and peer-reviewed benchmark suite creates a level playing field for competition that drives innovation, performance, and energy efficiency for the entire industry. It also provides critical technical information for customers who are procuring and tuning AI systems. This round of MLPerf Inference results also includes tests for four new benchmarks: Llama 3.1 405B, Llama 2 70B Interactive for low-latency applications, RGAT, and Automotive PointPainting for 3D object detection.

When Microsoft first introduced the Copilot+ program alongside its renewed push for Windows-on-Arm laptops, the AI-powered assistant features were mostly limited to Snapdragon X-powered devices. In addition to the inclusion of these features on Intel and AMD systems, Microsoft is also announcing Voice Access, a new accessibility feature that will first launch on Qualcomm Snapdragon systems and make their way to Intel- and AMD-powered systems. These new updates come by way of the March 27 Preview update titled KB505365. However, there is still no mention of an AMD and Intel launch for the much maligned Recall feature that Microsoft was testing late last year and recalled due to privacy concerns.

According to the latest Windows Experience Blog post, users of AMD- and Intel-powered PCs will now be able to access features, like Live Captions, Cocreator, Restyle Image, and Image Creator more broadly across the line-up of Copilot+ PCs with Intel Core Ultra 200V and AMD Ryzen AI 300 CPUs. Live Captions is officially pitched as an accessibility feature, meanwhile Restyle Image and Image Creator are AI-powered image editing and generation features, and Cocreator lies somewhere in between as a text-to-image tool that is meant to augment drawing in Paint. Cocreator will be rolling out as of the announcement, and Restyle Image and Image Creator will be available in the Photos app on Intel and AMD systems. As for Voice Access, Microsoft claims that it will allow users to be more flexible with their language when using speech to navigate their PCs, as opposed to "learning complex steps, commands and syntax that voice access previously required" for voice navigation on PC. Voice Access will initially be limited to Snapdragon X PCs, but it will roll out to AMD and Intel Copilot+ PCs later this year.

Over the past weekend, GMKtec's Weibo channel announced that pre-orders for its recently unveiled EVO-X2 mini PC model will start on April 7 (through JD.com), for customers located in China. Almost two weeks ago, the manufacturer boasted about its brand-new offering being the "world's first AI mini PC" equipped with AMD's Ryzen AI "Strix Halo" Max+ 395 APU. Extra international attention was gained, due to Lisa Su's autographing of a showcased unit during proceedings at the 2025 AI PC Innovation Summit (held on March 18, in Beijing). Pricing and availability were not mentioned during this press event, but GMKtec's Saturday (March 29) bulletin has revealed a (roughly) $2067 USD price point for the EVO-X2 launch model.

The manufacturer's blog entry stated that the: "EVO-X2 AI supercomputing host is coming, 128 GB + 2 TB priced at 14999 yuan, pricing reconstructs the desktop computing power boundary! Equipped with AMD Ryzen AI Max+ 395 flagship processor, 16-core 32-thread architecture with 5.1 GHz acceleration frequency, combined with 128 GB LPDDR5X memory and 2 TB high-speed storage, it can realize local deployment of 70 billion parameter large models, and AI performance exceeds NVIDIA's GeForce RTX 5090D graphics card." This potent compact AI-crunching solution is tempered by GMKtec's "innovative" Arctic Ocean cooling system. They advertise this design as using: "dual-turbofans and VC heat sinks to achieve silent heat dissipation at a peak power consumption of 140 W. The body adopts a recycled aluminium suspension design, equipped with HDMI, DP and USB4 interfaces, and supports Wi-Fi 6 + 2.5G network access." The brand has not yet announced an international release, but their EVO-X2 mini PC could face serious competition. Late last month, Framework debuted its Desktop product range—consisting of configurable 4.5L Mini-ITX systems—with a top-end Ryzen AI Max+ 395 (128 GB) model starting at $1999.

Successors to AMD's current-generation lineup of Ryzen 8000G desktop APUs are reportedly in the pipeline—according to the latest HXL/9550pro predictive declaration, finalized units could arrive at retail later this year. They propose that an "AMD AM5 New APU" family could arrive alongside an unannounced MSI Unify-X enthusiast-grade motherboard design, within the final quarter of 2025. Press interpretations of this inside track information point to possible upcoming "Ryzen 9000G" processors, utilizing Team Red's Zen 5 and RDNA 3.5 technologies. This potent combination already exists, albeit in mobile form—namely within Team Red's stable of Ryzen AI "Strix Halo, Strix Point," and "Krackan Point" APUs.

Industry experts opine that AMD will most likely deploy high-end "Strix Point" silicon to desktop, or more fancifully: "Gorgon Point." The latter codename turned up via leaks last week. Around early 2024, we witnessed Team Red's transfer of "Phoenix"—from original mobile formats—to their AM5 desktop platform. TechPowerUp's W1zzard evaluated the Ryzen 5 8500G "Phoenix 2" APU last summer; this plucky budget-friendly model sports Zen 4 and Zen 4c cores. Theoretically a flagship "Ryzen 9000G" SKU could emerge with twelve processor cores (4x "Zen 5" + 8x "Zen 5c"), a Radeon 890M iGPU, and an XDNA 2 NPU.

Around RDNA 4's launch period, ASRock's web presences listed an intriguing Radeon RX 9070 XT Steel Legend Dark 16 GB model—seemingly not mentioned in official introductory press material. As covered on TechPowerUp earlier this month, this unusual darkened spin-off of the "traditional white" Steel Legend design received little fanfare—the obvious reason being that gray/black stock was not readied for AMD's March 6 rollout to market. Over the past weekend, VideoCardz put an investigative spotlight on a Hermitage AkiHabara/GDM Japan press release. It turns out that ASRock's Radeon RX 9070 XT Steel Legend Dark SKU will become available early next month, starting with the Japanese "domestic market."

The manufacturer's global website was not updated with a similar PR item, so an international/widespread launch is likely planned for a later date. The "expected market price" upon release is 133,800 yen (including tax), which converts to roughly US$891. Industry watchdogs believe that the Dark variant will roll out "globally" with the exact same price tag as the already launched pale sibling. The Taiwanese brand did not introduce a Radeon RX 9070 (non-XT) Steel Legend Dark overclocked model, to exist alongside their white-hued RX 9070 Steel Legend OC offering. The wallet-friendlier triple-fan ASRock RX 9070 Challenger SKU—in black, with a splash of ARGB—is their only other Navi 48 XT GPU-based option.

Ayar Labs, the leader in optical interconnect solutions for large-scale AI workloads, today announced the industry's first Universal Chiplet Interconnect Express (UCIe) optical interconnect chiplet to maximize AI infrastructure performance and efficiency while reducing latency and power consumption. By incorporating a UCIe electrical interface, this solution is designed to eliminate data bottlenecks and integrate easily into customer chip designs.

Capable of achieving 8 Tbps bandwidth, the TeraPHY optical I/O chiplet is powered by Ayar Labs' 16-wavelength SuperNova light source. The integration of a UCIe interface means this solution not only delivers high performance and efficiency but also enables interoperability among chiplets from different vendors. This compatibility with the UCIe standard creates a more accessible, cost-effective ecosystem, which streamlines the adoption of advanced optical technologies necessary for scaling AI workloads and overcoming the limitations of traditional copper interconnects.

Today, AMD announced 5th Gen AMD EPYC processors power the Oracle Cloud Infrastructure (OCI) Compute E6 Standard shapes. 5th Gen AMD EPYC processors, the world's best server CPUs for enterprise, AI and cloud, enable OCI Compute E6 shapes to deliver up to a 2X increase in cost to performance, compared to the previous E5 instance generation based on testing by OCI.

The new OCI Compute E6 shapes build on the success of the previous E5 generation to deliver leadership performance and cost efficiency for general-purpose and compute-intensive workloads. These OCI shapes add to the selection of more than a thousand compute instances powered by AMD EPYC processors across all major cloud service providers.

MAINGEAR, the leader in premium-quality, high-performance gaming PCs, today unveiled its new 2025 pre-configured MG-1 gaming systems, equipped with NVIDIA's cutting-edge GeForce RTX 50-Series GPUs and powered by the latest AMD Ryzen 9000 Series CPUs and Intel Core Ultra Series 2 processors. This exciting lineup offers gamers a significant leap in power, performance, and value, reinforcing MAINGEAR's commitment to redefining the gaming experience across all levels of play.

"2025 is bound to be the best year yet for PC gamers, and the introduction of our newest MG-1 systems continues to drive MAINGEAR's mission to deliver the ultimate gaming experience," said Wallace Santos, CEO of MAINGEAR. "With NVIDIA's RTX 50-Series GPUs, and the latest AMD Ryzen and Intel Core Ultra processors, we've got more options than ever before, ensuring the perfect fit for gamers of all skill levels and budgets."

AMD today announced the completion of its acquisition of ZT Systems, a leading provider of AI and general-purpose compute infrastructure for the world's largest hyperscale providers. The acquisition will enable a new class of end-to-end AI solutions based on the combination of AMD CPU, GPU and networking silicon, open-source AMD ROCm software and rack-scale systems capabilities. It will also accelerate the design and deployment of AMD-powered AI infrastructure at scale optimized for the cloud.

AMD expects the transaction to be accretive on a non-GAAP basis by the end of 2025. The world-class design teams will join the AMD Data Center Solutions business unit led by AMD Executive Vice President Forrest Norrod. AMD is actively engaged with multiple potential strategic partners to acquire ZT Systems' industry-leading U.S.-based data center infrastructure manufacturing business in 2025.

ASUS has readied its first all-in-one water cooler with a built-in curved AMOLED screen called ROG RYUO IV SLC 360 ARGB. The upcoming ASUS ROG RYUO AIO cooler integrates a 6.67-inch 60 Hz AMOLED display for system monitoring, significantly larger than any commercial AIO cooler with a built-in display on the pump/water block combo. The cooler operates at 39.6 dB while its 120 mm fans deliver 71.4 CFM airflow at speeds between 500-2650 RPM through PWM/DC control. A copper water block connected to a 360 mm aluminium radiator forms the thermal transfer system, joined by 200 mm FEP sleeved tubing. Socket compatibility includes current-generation Intel LGA-1851/1700 and AMD AM4/AM5 platforms, with no support for legacy sockets.

The expanded display provides real-time data visualization during intensive computational workloads without requiring additional monitoring software. ASUS plans to release the cooler before the "618" Mid-Year Shopping Festival. While the initial product features a 360 mm radiator, smaller variants may follow for space-constrained builds. No pricing details have been announced. A similar solution to this is TRYX Panorama AIO, which features a 6.5-inch L-shaped 3D screen as well, but ASUS has outdone it with a 6.7-inch screen variant now. We are curious how this cooler will look in real life, and some photos are already available below.

Here are some of the first pictures of the ASUS ROG Crosshair X870E Extreme, the company's new flagship Socket AM5 motherboard. Until now, the company's AMD 800-series chipset motherboard lineup topped off with the ROG Crosshair X870E Hero for connectivity and the ROG Crosshair X870E Apex for overclocking chops, and the company is looking to push things up a notch, especially given that AMD's Ryzen 9000X3D series is established as the fastest processor series for gaming PC builds. This board is designed to compete with the likes of the MSI MEG X870E Godlike and the GIGABYTE X870E AORUS Xtreme. It is firmly into the E-ATX territory in terms of dimensions. The board draws power from a combination of 24-pin ATX, two 8-pin EPS, and an 8-pin PCIe power. It uses the company's most powerful CPU VRM solution for the AMD platform, above even the 22-phase solution of the Hero.

The ASUS ROG Crosshair X870E Extreme comes with a large number of 4-pin PWM fan and water pump headers. Storage connectivity includes one M.2-22110 Gen 5 on the board with a chunky heat pipe-based cooler; a second Gen 5 and two Gen 4 slots on the board, and a DIMM.2 module that provides additional slots. The expansion slots provided are a PCI-Express 5.0 x16 and a second PCI-Express 4.0 x4 (physical x16). There are a boat-load of USB ports, including a couple of 40 Gbps USB4, 20 Gbps USB 3.2 Gen 2x2, and several 10 Gbps and 5 Gbps USB 3.2 ports. Networking interfaces include Wi-Fi 7, a 10 GbE, and a 5 GbE. The board features the company's most premium onboard audio solution that probably combines a Realtek ALC4082 with an ESS ES9219 DAC for the front channels, an audiophile-grade OPAMP, and premium capacitors. The board offers many of the exclusive overclocker-friendly features found in the company's ROG Maximus Extreme motherboards. A star-attraction is its 4-inch true-color display over the VRM heatsinks, which can be programmed to show anything. There is no word on availability or pricing.

AMD Radeon RX 9070 XT and RX 9070 "MBA" models have turned up for sale through Chinese black market channels; with local PC hardware enthusiasts quickly snapping up these "reference design" curios—for "scientific" analysis, and in some cases: bragging rights. Officially, Team Red has relied on its board partners to produce an all-custom first wave of RDNA 4 gaming graphics cards. Recent discoveries of authentic-looking "Made by AMD" specimens suggested a very limited distribution of first-party units to trusted partners. According to a fresh ITHome news article, Team Red's Radeon RX 9070 XT reference PCB design was fully revealed via a comprehensive teardown. Kyogre shared heavily watermarked photos with the online publication; showcasing their disassembly of a "leaked" example.

Team Red's proprietary triple-fan cooling solution and backplate were removed; leading to the exposure of gory innards. ITHome provided a quick rundown visual clues—the analyzed unit features: "dual PCIe 8-Pin power supply interfaces, and a black shell with a tough and simple shape. It is about three slots thick and is equipped with a rear open and ventilated cooling backplane with a core back frame...Looking deeper into the internal structure of the graphics card, we can find that this graphics card is equipped with five heat pipes, the corresponding position of the GPU core uses a copper base, and the surrounding memory positions are also in contact with the metal frame through thermal pads. In addition, it is equipped with four video output interfaces."

Shipping manifests have served as fairly reliable sources of pre-launch information—Everest (aka Olrak29) has discovered many juicy details in recent times. Their latest sleuthing session—combing through NBD documents—has indicated AMD's (alleged) prepping of a larger socket design for next-generation mobile processors. A leaked document alludes to the existence of various "MEDUSA01" jig and block "FP10" socket validation parts. Current-generation Ryzen AI "Strix Point" 300 series APUs utilize the FP8 socket format. Based on the "MEDUSA01" shipping manifest, it seems that a successor will arrive with a larger footprint—measurements of 25 mm x 42.5 mm are repeated throughout the leaked description list. Industry watchdogs surmise that "Medusa Point's" BGA FP10 socket will be approximately 6% larger than its predecessor.

Mid-way through last month, insider theorizations pointed to "Medusa Point" being a chiplet-based design. A "single 12-core Zen 6 CCD" was linked to a TSMC 3 nm-class node, with "N4P" reportedly selected for a separate mobile client I/O die. Readily available 4 nm Ryzen AI "Strix Point" processors are monolithic in nature. Initial inside track info mentioned RDNA 4 technology in the same equation as "Medusa Point," but recent Team Red's recent-ish targeting of "GFX1153" places RDNA 3.5 as the de facto choice.

ASUS silently releases its first AMD B850 motherboard featuring the GC-HPWR connector, allowing it to deliver up to 600 W to a compatible graphics card without needing external cables. The TUF Gaming B850 BTF Wi-Fi W motherboard supports Ryzen 7000, 8000, and 9000 CPUs, including the latest 9950X3D, and offers all the typical features expected from a mid-tier AMD motherboard. On a quick look over main specifications, we can see it uses a 14+2+1 power stage design and an eight-layer PCB, supports DDR5 memory up to 8000 MT/s, PCIe 5.0 x16 with updated Q-Release system, one Gen 5 M.2 slot, ALC1220P audio, and BIOS flashback.

ASUS first created its Back to the Future (BTF) lineup with the intention of moving all motherboard connectors to the back of the PCB. While this makes installation and cable routing a bit more challenging, it results in a much tidier-looking motherboard and a cleaner overall system look without the usual mess of wires. BTF-compatible GPUs help eliminate power cables by replacing the standard 8-pin or 12VHPWR connectors with a metal plug at the end of the PCIe slot. ASUS is now bringing this connector to its more affordable B850 boards, which could help make this rare GPU power solution more common. As a BTF board, it hides all cable connectors on the front—even for GPU power as we said, you'll find the GC-HPWR connector at the end of the PCIe x16 slot, just below the RAM slots. However, you don't have to use this feature—the motherboard works fine with any modern graphics card, whether it relies on the new power option or standard connectors. That said, to achieve a completely cable-free setup, you'll need to pair it with one of ASUS's BTF graphics cards. This currently limits you to the RTX 40-series, as ASUS has yet to release any 50-series BTF cards. The fact that ASUS is adding this new type of connector to midrange motherboards could suggest that it plans to expand the BTF lineup in the future. ASUS didn't reveal pricing or any release date information for the TUF Gaming B850 BTF Wi-Fi W motherboard.

Kuroutoshikou—a Japanese PC hardware brand—has introduced a special Blade and Soul NEO collaboration Radeon RX 7600 8 GB model. The company's press release goes into great detail about their retail packaging being "decorated with the beautiful characters and gorgeous stages of Blade," but the card itself is not adorned with any fancy illustrations or markings. In fact, the bare black design seems to be borrowed almost directly from PowerColor's Fighter stable (minus two stickers). As pointed out by VideoCardz, Kuroutoshikou already offers a "normal edition" Radeon RX 7600 SKU—minus NCSoft-related material. An exclusive in-game bonus seems to be the main draw here—as disclosed in the company's marketing spiel: "if you purchase this product and apply on the special page, you will receive a campaign code for the limited costume 'Gentleness' that can be used in Blade and Soul NEO."

Kuroutoshikou will deliver their "RD-RX7600-8G/Blade & Soul NEO Collaboration" edition to market on April 4, which could place it in a similar release timeframe with natural successors—AMD and board partners are expected to launch Radeon RX 9060 Series graphics cards in the second quarter of 2025. Surprisingly, the lower end of RDNA 3 keeps generating news articles—PowerColor seems to be working on a "V2" Fighter Radeon RX 7600 SKU, while plenty of China market-exclusive Radeon RX 7650 GRE cards have appeared within the past month or two. A possible surplus of Navi 33 GPUs could explain the sudden "reemergence" of previous-gen hardware. Going back to Kuroutoshikou; their current graphics card portfolio seems to consist of slightly rebranded PowerColor designs for the AMD side of things, while NVIDIA GeForce offerings are rebadged GALAX IPs.

The AMD Ryzen 5 9600 launched recently as a slightly more affordable variant of the popular Ryzen 5 9600X. Despite launching over a month ago, the 9600 still appears rather difficult to track down in retail stores. However, a recent PassMark benchmark has provided some insights as to the performance of the non-X variant of AMD's six-core Zen 5 budget CPU. Unsurprisingly, the Ryzen 5 9600X and the Ryzen 5 9600 are neck-and-neck, with the 9600X scraping past its non-X counterpart by a mere 2.2% in the CPU benchmark.

According to the PassMark result, the Ryzen 5 9600 scored 29,369, compared to the Ryzen 5 9600X's 30,016, while single-core scores were 4581 for the 9600X and 4433 points for the 9600, representing a 3.2% disparity between the two CPUs. The result is not surprising, since the only real difference between the 9600 and the 9600X is 200 MHz boost clock. All other specifications, including TDP, core count, cache amount and base clock speed, are identical. Both CPUs are also unlocked for overclocking, and both feature AMD Precision Boost 2. While the Ryzen 5 9600 isn't available just yet, it will seemingly be a good option for those who want to stretch their budget to the absolute maximum, since recent reports indicate that it will be around $20 cheaper than the Ryzen 5 9600X, coming in at around the $250-260 mark.

At the tail end of 2024, AMD technical staffers added the "GFX1153" target to their first-party GPU supported chip list. Almost three months later, PC hardware news outlets and online enthusiasts have just picked up on this development. "GFX1150" family IPs were previously linked to Team Red's RDNA 3.5 architecture. This graphics technology debuted with the launch of Ryzen AI "Strix Halo," "Strix Point" and "Krackan Point" mobile processors. Recent leaks have suggested that Team Red is satisfied with the performance of RDNA 3.5-based Radeon iGPUs; warranting a rumored repeat rollout with next-gen "Medusa Point" APU designs.

Both "Medusa Point" and "Gorgon Point" mobile CPU families are expected to launch next year, with leaks pointing to the utilization of "Zen 6" and "Zen 5" processor cores (respectively) and RDNA 3.5 graphics architecture. RDNA 4 seems to be a strictly desktop-oriented generation. AMD could be reserving the "further out" UDNA tech for truly next-generation integrated graphics solutions. In the interim, Team Red's "GFX1153" IP will likely serve as "Medusa Point's" onboard GPU, according to the latest logical theories. Last year, the "GFX1152" target was associated with Ryzen AI 7 300-series "Krackan Point" APUs.

Tom's Hardware just published an intensive data collection of online prices of the latest GPUs from AMD and NVIDIA, suggesting that real-world prices are nowhere near MSRP. With an investigation into the 30-day eBay average price based on listings, the data shows that a lot of GPU SKUs are retailing for well-above-average price premiums. The data tracker also looked for the best-priced listing of a specific SKU. For instance, NVIDIA's GeForce RTX 5070 is seen retailing at around $700—a 27% increase over its official MSRP of $550—while the RTX 5080 pushes these premiums even further by selling at over 50% above its suggested price. The flagship RTX 5090 tops the chart with a staggering $4,222 on secondary markets compared to its $2,000 MSRP, an increase of roughly 111%.

In contrast, earlier models like the RTX 4060 and RTX 4060 Ti show relatively minor inflation, with increases of only about 3% and 5%, respectively, according to recent eBay averages. Still, selling years-old GPUs at MSRP today proves that there is demand. On the AMD side, the Radeon RX 9070 series is no less dramatic. The base RX 9070, with a $550 MSRP, averages around $817 in real-world sales—a nearly 48% premium—while its high-end sibling, the RX 9070 XT, jumps to approximately $1,001 from a $600 MSRP, marking an increase of roughly 66%. Early figures even suggest that first-week sales for the RX 9070 series were ten times higher than those of previous AMD models, justifying the price surge. AMD is working on addressing this supply, which should improve in April, and NVIDIA is working with AIB partners, too, to deliver more Blackwell GPUs.

Late last week, a leak suggested a forthcoming Spectral White Limited Edition refresh of the existing Red Devil Radeon RX 9070 XT graphics card design. PowerColor's pale reinterpretation of flagship level hardware appeared online courtesy of an accidental Amazon listing; indicating $899/€929 pricing. The Taiwanese manufacturer's Spectral White product line—historically favoring Hellhound SKUs—is viewed favorably as offering a nearly "all-white" aesthetic; extending to proper coloration of PCB and I/O panel designs.

As expected, PowerColor has distributed Limited Edition packages to media and influencer outlets—as evidenced by Overclock3D's very recent showcasing of a sample unit. In response to last week's leak, OC3D's "Tiny" Tom Logan "previewed" the ghostly Red Devil RX 9070 XT variant—he noted that a "performance/review" NDA is still in effect; due to be lifted on April 2. Video coverage explored a slight "bending of the rules," but Logan did not power up his evaluation specimen. As evidenced by the unboxing process, PowerColor has bundled in several Spectral White-tinted accessories—indicating a commendable level of attention to detail. By contrast, the manufacturer's slightly inelegant and contradictory nomenclature is confirmed by outer box labelling—loudly exclaiming: "RED DEVIL Spectral White." Amusingly, Logan and his team have nicknamed this new sub-variant: "White Devil."

Around late January, out-of-date AMD marketing material teased the existence of a Radeon RX 9070 series reference card design. Almost a month later, PC hardware news outlets picked up on an official signal about Team Red's launch lineup consisting entirely of board partner-produced options. First-party enthusiasts were disappointed by the apparent total lack of "Made by AMD" (MBA) solutions, but some unusual specimens appeared online roughly two weeks post-RDNA 4's launch. Reports pointed to triple-fan Radeon RX 9070 XT and dual-fan RX 9070 MBA cards being exchanged for cash via Chinese black market channels. Photographed examples seemed to sport a somewhat muted black shroud design—not quite as exciting when compared to AMD's marketed/rendered brushed metal effect promo units.

Members of the Chiphell forum have spent months leaking many aspects of Team Red's foray into a new generation of graphics architecture—going back to the days of old nomenclature: Radeon RX 8800 XT. Yesterday, one participant revealed their fresh purchase of a Radeon RX 9070 non-XT MBA card. They sold their old GeForce RTX 4070 SUPER 12 GB graphics card, in favor of Navi 48 GPU-based OEM hardware. The post focused mainly on photo uploads and screenshots, but a brief description stated: "purchased at original price (TPU note: presumably 4499 RMB), room temperature is 16 degrees Celsius. Dual fans on the front. The back panel has an AMD logo, but it's a sticker." As theorized by VideoCardz, AMD likely produced a limited number of pre-release "public" MBA cards. The publication reckons that partner companies have received a smattering of samples for evaluation or software development purposes. The presence of an old school Radeon logo (pre-RDNA era) is a head scratcher, given the unit's supposed first-party origin.