Intel GPU-based Linux gamers are about to receive a performance uplift. According to the latest update to DXVK, a Vulkan-based translation layer for DirectX 8/9/10/11, Intel's "Battlemage" dGPUs and "Lunar Lake" Arc iGPUs now support memory defragmentation by default, helping Intel GPUs reduce VRAM usage and increase performance, which provides a smoother gaming experience when using the DXVK translation layer. In version 2.7, DXVK forces discrete GPUs to adhere to driver-set memory budgets whenever possible and dynamically frees unused resources to system memory when applications place a significant VRAM load. DXVK v2.7 release notes state that this will be particularly beneficial to Unity Engine and VRAM-constrained titles. This optimization will also enable higher texture quality in game settings, maintaining the same performance as before, and further building upon the memory optimization.

As a critical component of Valve's Proton software stack, this DXVK update also provides Intel-based handhelds, such as the MSI Claw, with increased performance when running SteamOS. Given that we measured a sizable crowd, comprising 3.4% of all handheld users, playing with MSI Claw, many handheld gamers considering running SteamOS on their device will find this DXVK update quite welcome.Here is a complete DXVK v2.7 changelog.

NVIDIA has released its DLSS 4 plugin for Unreal Engine 5.6, allowing developers to incorporate the company's most advanced upscaling and AI-driven frame generation tools into their projects. With this update, performance improvements introduced in UE 5.6, such as optimized memory handling and refined rendering pipelines, receive the benefits of Transformer-based upscaling, which significantly lowers VRAM usage while generating extra frames to improve motion smoothness. Developers using Unreal Engine 5.6 can now integrate multi-frame generation, ray reconstruction, deep learning anti-aliasing, and super resolution in one convenient package that requires fewer resources than earlier DLSS versions.

The DLSS 4 plugin relies on a Transformer neural network that evaluates the relationships among all pixels in a frame rather than using localized convolutional filters. This technique reduces video memory requirements by approximately 15-20% compared with the prior DLSS 3 SDK, freeing up capacity for higher-resolution textures and more detailed environments. Multi-frame generation predicts intermediate frames based on recent frame history, effectively boosting perceived frame rates without additional GPU draw calls. Ray reconstruction enhances reflections and global illumination by learning from high‑quality offline renders, delivering realistic lighting with minimal performance loss. Early feedback indicates a 30-50% uplift in ray-traced effects quality in GPU-bound scenes, as well as noticeably sharper visuals under dynamic lighting. The plugin supports Unreal Engine versions 5.2 through 5.6; however, only projects running on version 5.6 can access the full suite of improvements.

Monster Hunter Wilds players on PC have recently had a bit of a rough time of things when it comes to performance, with the game's recent Steam reviews seeing a number of players complaining about random stutters, FPS drops, and a general lack of optimization. Even gamers playing on relatively high-end recent GPUs, like the AMD Radeon RX 7900 XT report that their experience has been tainted by poor performance, even if the actual game content is good. With the latest Monster Hunter Wilds patch, though, Capcom is attempting to address those performance issues—and the resulting slew of negative reviews—as announced today in a post on X.

While there are new monsters, weapons, cosmetics, equipment, and other in-game content, the majority of the focus of the 1.020.00.00 update (full notes here) is on those performance updates. Capcom has changed the way shader compilation works, now making the CPU-intensive task take place the first time you run the game after an update as well updated a slew of upscaler and frame generation changes, primarily adding DLSS 4 and FSR 4 support for GPUs newer than the GeForce RTX 2000 series and AMD Radeon RX 9000 series. The new fix also allows players to mix upscaling and frame generation methods, which should allow players to better tune the game's visuals and performance. Additional fixes to the game include reduced VRAM usage from texture streaming and a more accurate calculation of estimated VRAM consumption. Steam users also now get a notification upon launching Monster Hunter Wilds if they are running an unsupported operating system or out-of-date GPU drivers, or if they are running the game in compatibility mode.

When we heard the news that NVIDIA will be halting shipments to Chinese customers for its GeForce RTX 5090D GPU in Q2, there were some speculations that the company is preparing to make a comeback. Leaving out the Chinese market is a tough decision from a business standpoint, but NVIDIA complies with export regulations imposed by the US administration under Donald Trump. However, the company now plans to keep its presence in China with the release of the export-abiding GeForce "RTX 5090DD" GPU. Carrying a GB202-240 die, it will be a cut-down version with reportedly only 14,080 CUDA cores present, paired with 24 GB of GDDR7 memory.

Chinese gamers will notice a significant difference, but so will anyone trying to acquire these GPUs for non-gaming purposes. As the US administration has banned shipments of high-end chips to China to stop them from training military-grade AI, gaming GPUs have been a target of AI firms. Now, a modified GeForce RTX 5090 DD, which not only cuts with raw compute but also firmware, should be sufficient for gaming. Carrying a PG145 SKU 40 board, the new RTX 5090DD GPU will also feature a modified 384-bit bus for its 24 GB of memory, down from the original RTX 5090D's 512-bit bus.

Update 09:45 UTC: According to one of the most accurate NVIDIA leakers, kopite7kimi, the core count for this GPU is supposed to be 21,760 CUDA cores, with 575 W TDP. The leaker did note that there is some surprise, so NVIDIA has definitely changed something big under the hood to comply with export regulations.

Gearbox and 2K today opened Borderlands 4 pre-orders, with the rather unexpected $69.99 base game price resulting in a simultaneous sigh of relief following the gaming industry's recent move towards $80 AAA games. Following the pre-order launch, Borderlands 4's Steam page revealed the game's minimum system requirements as a bit of a mixed bag. According to the Steam page, the new looter shooter will only run on an AMD Radeon RX 5700 XT or NVIDIA GeForce RTX 2070, but this seems to largely be related to VRAM, since the minimum requirements also mentions that 8 GB is the minimum VRAM amount. An SSD and an eight-core CPU—listed as the AMD Ryzen 2700X or Intel Core i7-9700—will also be mandatory for Borderlands 4. Even the recommended specifications for Borderlands 4 are rather moderate, with the NVIDIA GeForce RTX 3080 and Radeon RX 6800 XT making the cut.

In addition to the reasonable PC specs, news also broke about an ECHO-4 Collector's Edition, a curious edition to the line-up of special edition bundles for Borderlands 4 that seems to be exclusive to GameStop for now. The Echo-4 Collector's Edition bundle is a host of physical Borderlands 4 memorabilia, including a cloth map of Kairos, an eight-inch Echo-4 figure with LED lights, a Vault symbol necklace, a card with five SHiFT codes for Borderlands 4 golden keys, and four double-sided lithography art prints—all for $149.99. The one glaring omission from the Echo-4 bundle is an actual game key for Borderlands 4. That is clearly labelled as "sold separately" on the GameStop page.

Prior to early May, the existence of AMD's Radeon RX 9060 XT 8 GB GDDR6 model seemed to be in question. Discerning graphics card buyers tend to opt for larger pools of VRAM. In the modern era, 12 GB or 16 GB options are preferred in budget-to-mid-range segments. Additionally, media outlets are growing weary with a continued delivery of new-gen 8 GB offerings. Yesterday, Digital Foundry's Richard Leadbetter expressed a similar sentiment: "unhappily, AMD did not sample the 8 GB version of the (Radeon RX 9060 XT) card—perhaps because it knows the kind of reception it would receive from a tech press that are mostly united about the need for 8 GB cards to be shown the door." The site's Technology Editor was not alone in voicing frustrations about certain conditions that apply to international members of the media.

Yesterday, Linus Tech Tips (LTT) pulled back a major curtain. Their leaking of an alleged "official review guideline" demonstrates some level of exerting too much control. LTT shared a screenshot of AMD's "commitment to the press" (see below). On the subject of evaluators gaining access to the cheaper (baseline MSRP: $299) Navi 44 XT GPU-based card, a liaison stated: "as for the 8 GB models, AMD has enabled global reviews of both 16 GB and 8 GB models of the Radeon RX 9060 XT based on regional market demand. So in short, yes, there are some other global publications that are receiving 8 GB models for testing." VideoCardz's tracking of day one critiques reveals an almost complete seeding of 16 GB variants (starting at $349). Vitally, TechPowerUp's W1zzard will be reaching into his own pocket(s). As outlined in his reviews of various better equipped options: "since it wasn't possible to get an 8 GB model for review, everyone focused on 16 GB models. I'll buy one of course as soon as possible, to get you the crucial info how it performs." Back in April, (on his own dime) TPU's resident GPU critic acquired a Gainward GeForce RTX 5060 Ti 8 GB retail specimen. At the time, NVIDIA caught plenty of flak—only GeForce RTX 5060 Ti 16 GB models were distributed to reviewers and influencers.

GIGABYTE TECHNOLOGY Co. Ltd, a leading manufacturer of premium gaming hardware, today launched the AORUS GeForce RTX 5090 STEALTH ICE 32G graphics card, designed exclusively for the GIGABYTE Project STEALTH series. This new graphics card inherits the minimalist philosophy of Project STEALTH by adopting a hidden power connector design, allowing seamless cable routing and a cleaner build layout. Tailored for enthusiasts who prioritize aesthetic refinement and efficient assembly, the card offers both cutting-edge performance and a streamlined look that elevates any gaming setup.

Hide the Cables, Power in Stealth
Project STEALTH redefines the PC building experience by relocating all connectors to the rear of the components. This thoughtful design streamlines cable management, enabling quicker installation and a cleaner, more organized build.

If you recall NVIDIA's GTX TITAN Z GPU with two GPU dies and an impressive 6 GB of memory, it was a big deal at the time and quite an interesting engineering solution. Today, thanks to sources close to VideoCardz, we learn that Intel's Arc "Battlemage" B580 could get the next dual GPU, dual VRAM treatment. According to the source, an unnamed Intel add-in board partner is preparing to launch a GPU with two B580 dies and 48 GB of VRAM. This means that the 24 GB VRAM version of the Arc B580 we saw in EEC filings is basically confirmed, and that some models will be out for display very soon. While the BMG-G21 chip offers up to 20 Xe cores and 2,560 shader units, a dual BMG-G21 configuration will yield 40 Xe cores and 5,120 shader units working in tandem with 48 GB of VRAM.

Clearly, this model is mostly tailored to AI enthusiasts. No game (thankfully) requires 48 GB of VRAM for now, but it is still nonetheless an interesting solution to see. It has been a while since we got something like a TITAN Z, even in high-end form factors. There is a rumored Intel Arc PRO A60/B60 card carrying a B580 die with 24 GB of VRAM, tailored for local AI workloads and professional visualization. This GPU could be the base for the newly rumored dual B580, 48 GB SKU from an unnamed Intel AIB partner. We expect to hear more details at Computex, which is now just a week away!

A few weeks ago, we reported on AMD Radeon RX 9060 XT, which is scheduled to come right after this year's Computex show. Some early leaks have pointed to the existence of two Radeon RX 9060 XT variants: one with 16 GB of GDDR6 VRAM and another with an 8 GB GDDR6 capacity. Recent rumors have begun speculating that the 8 GB card is not coming at all, which BenchLife now debunks. According to the publication: "As for the AMD Radeon RX 9060 XT 8 GB version, we have reliable sources telling us that there is currently no plan to stop supply or cancel it. As for the news from the market, it is just a rumor. The main reason is as mentioned earlier, it is entirely due to the reaction to the GeForce RTX 5060 Ti."

The 8 GB version of NVIDIA's latest GeForce RTX 5060 Ti wasn't well received. It wasn't even supplied to reviewers, and out own review was delayed as we waited to buy a card off the shelf. TechPowerUp's reviewer W1zzard confirmed that "If you want ray tracing, then RTX 5060 Ti 16 GB is your best option, and of course anything upcoming from AMD in that price bracket—we've been hearing rumors about an RX 9060 Series after Computex, an RX 9070 GRE is also likely, but both are unknowns in terms of performance and pricing." Perhaps if AMD can price this 8 GB card aggressively, it will receive praise from consumers. If not, it will be a turning point for the mid-range PC gamers, who now demand more VRAM for their cards so they are not left behind with future title releases, especially as they become more demanding.

Officially, AMD is set to reveal the next wave of RDNA 4—in desktop form—at some point within the second quarter of this year. Unlike its main rival—NVIDIA—company representatives have not publicly discussed a comprehensive branch-off into dedicated laptop graphics solutions. Team Red's previous-gen "Navi Mobile" Radeon RX 7000M lineup received little fanfare, and early 2025 "official" talk indicated that company engineers were focused on delivering RDNA 4 to desktop. Additionally, late March leaks suggested an increased future reliance on integrated RDNA 3.5-based graphics solutions. Seemingly out of nowhere, All The Watts!! has produced a list of six tentative Radeon RX 9000M family variants. As interpreted by VideoCardz, the leaker has cryptically included supposed compute unit (CU) counts and VRAM configurations. A rumored Radeon RX 9080M flagship SKU could be based on AMD's Navi 48 GPU design; prepped with 64 RDNA 4 compute units (or 4096 stream processors) and 16 GB of video memory.

The leaked Radeon RX 9070M XT's basic details match certain data points present on the rumored Radeon RX 9070 GRE 12 GB desktop card's specification sheet. VideoCardz reckons that the RX 9070M XT will also utilize Team Red's Navi 48 GPU, but with a reduced compute unit count of 48 (or 3072 SPs). The four remaining RDNA 4 laptop SKUs—RX 9070M, RX 9070S, RX 9060M and RX 9060S—are said to be based on company's Navi 44 GPU IP, and configured with 8 GB pools of VRAM. Two days ago, AMD hinted about new generation Radeon Mobile gaming experiences getting introduced at next month's Computex 2025 trade show.

Just before the official launch of Radeon RX 9070 XT graphics cards, members of the Chiphell forum expressed concerns about unnamed evaluation samples exhibiting worrying memory module temperatures. Days later, Western review outlets published similar findings across several board partner specimens. Typically, pre-launch and retail units have utilized SK hynix GDDR6 VRAM modules—TechPowerUp's W1zzard uncovered numerous examples of the manufacturer's "H56G42AS8DX-014" model during teardown sessions. Curiously, Chinese PC hardware enthusiasts have happened upon Radeon RX 9070 XT cards that utilize Samsung GDDR6 memory modules. It is not clear whether Yeston has outfitted its flagship Sakura Atlantis OC SKU with Samsung components from the very beginning, but one owner documented a GPU-Z diagnostics session—late last month—showing "GDDR6 (Samsung)" onboard. Earlier this month, insiders posited that NVIDIA was in the process of changing its main GDDR7 memory vendor—from Samsung to SK hynix—for the GeForce RTX 50-series.

According to a recent Guru3D news piece, additional cases were reported. Online conjecture points to "higher end" custom models being updated with "cooler" modules. Hilbert Hagedoorn—Guru3D's head honcho—has gathered compelling information via community feedback channels: "the transition to Samsung memory has yielded noticeable thermal advantages. Early reviews and comparisons indicate that the new memory modules help lower temperatures significantly. Custom RX 9070 XT models running demanding benchmarks like Furmark have demonstrated memory temperatures of 75°C or lower without necessitating an increase in voltage or a reduction in clock speeds. Notable manufacturers, including Sapphire, XFX, and GIGABYTE, have already integrated Samsung's GDDR6 into their custom variants. However, as of now, AMD has not authorized its partners to explicitly differentiate between models with SK hynix and Samsung memory, likely to avoid confusion among consumers and maintain a consistent product lineup." Additionally, TechPowerUp forum members have found Samsung GDDR6 memory on Radeon RX 9070 XT and RX 9070 custom models. Thanks for the tip, Fluffmeister.

Not long after Intel's launch of the Arc Xe2 "Battlemage" B580 12 GB graphics card design, insiders started generating noise about potential spin-offs bound for release in 2025. In theory, the speculated "B580 24 GB" variant could arrive as a workstation-oriented discrete graphics solution—possibly lined up as a next-gen entry within Team Blue's Arc Pro family. Three mysterious BMG (aka "Battlemage") PCI identifiers turned up at the end of January; sending online PC hardware debates into overdrive; one faction believed that Team Blue was readying fabled productivity-focused B-series cards—complete with enlarged pools of GDDR6 VRAM. Apparently, Sparkle's Chinese branch has provided comment on newer rumors—from March, according to VideoCardz. The Taiwanese manufacturer is a key Intel board partner in the field of Arc GPU-based graphics card products—across gaming and professional desktop lines. Unfortunately, the company's head office (in Taiwan) has dismissed "official" claims about a May/June launch of an unnamed 24 GB model. Sparkle's Chinese social media account engaged with members of the PC hardware community, and outlined an "original plan" to release something new within the second quarter of 2025—apparently the incoming card is "still being arranged."

Earlier today, NVIDIA's public relations department published their "GeForce RTX 5060 Desktop Family" introductory article. Curiously, the company's brand-new GeForce RTX 5060 Ti graphics cards do not headline Team Green's "dedicated" PR release—instead, the general "game changing" GeForce RTX 5060 series is advertised with a starting price point of $299. Clarification arrives several paragraphs deep into the blurb—as explained with some "PR" magic: "starting April 16th, we're bringing DLSS 4 and Blackwell's suite of innovations to every gamer with the launch of the GeForce RTX 5060 desktop family, beginning with the release of the GeForce RTX 5060 Ti. And in May, the GeForce RTX 5060 arrives, with prices starting at $299." Very specific cost of ownership digits leaked last week; indicating a refreshing reduction over earlier (disappointing) predictions.

Team Green's publicity team has confirmed starting MSRPs of $429 for 16 GB models, and $379 for 8 GB variants. These official numbers are buried three-quarters of the way into NVIDIA's PR document. Older evidence pointed to a possible repeat of lower end GeForce RTX 40-series guide prices—as it turns out, the GeForce RTX 5060 card's $299 "cost of entry" aligns with its predecessor's launch figure. Based on leaked pre-built PC listings, industry watchdogs deduced the $299 MSRP earlier on in the month. Additionally, Team Green's PR material teased the upcoming launch of related mobile hardware: "the same features, innovations and advantages of the GeForce RTX 5060 desktop family are coming to laptops this May, when GeForce RTX 5060 laptops arrive on shop shelves, starting at $1099." Real life scenario prices are expected to remain "sky-high," even for rumored cheaper devices—equipped with GeForce RTX 5050 Mobile GPUs.

AMD's Radeon RX 9070 XT and RX 9070 "MBA" graphics cards are no longer "best kept secrets"—as demonstrated recently by the "leaking out" of supposed reference models through black market/back alley channels in China. Late last month, a dual-fan non-XT specimen was snapped up by Chiphell forum member—alleged benchmark results were soon shared within that community. A few days later, a "Made-by-AMD" Radeon RX 9070 XT sample was dissected and compared to Sapphire's PULSE Radeon RX 9070 XT 16 GB SKU. An additional MBA XT example emerged last week, courtesy of another in-depth Chiphell thread. A "bored" enthusiast happened upon a 5499 RMB (~$748 USD) when idly browsing through Xianyu listings (Taobao's Ebay equivalent platform). Their buying experience was described as follows: "(I) found a 'public version' Radeon RX 9070 XT in Tianjin. It was said to be manufactured by (an) OEM, so I bought it without hesitation...I made an appointment to meet today and got it successfully. I don't have to wait for a graphics card anymore."

For unknown reasons, AMD decided to launch its first wave of RDNA 4 gaming graphics cards sans first-party designs. Leaked specimens have attracted much attention in China; with owners bragging about their respective ownerships of reasonably priced rarities. The latest back channel customer expressed satisfaction when inspecting Team Red's all-black flagship Navi 48 GPU-based solution: "after I got it, I have to say that it is not very heavy and is quite light. But the appearance is really what I like." After initial tests, they discovered that VRAM temperatures were not up to snuff—as alluded to (pre-launch) by other Chiphell figures. Their personal DIY improvements were described: "(I) took it apart to measure the thickness of the thermal grease pad. Everyone said the temperature of the video memory was high, so I decided to change to something better....(with) original silicone grease FurMark 2K resolution for 20 minutes: maximum core temperature was 62 degrees, maximum hot spot temperature was 84 degrees, maximum memory temperature was 88 degrees, maximum power was consumption 346 W."

As we have witnessed in the past few weeks, leaked retailer listings are undoubtedly an interesting way to gain more information regarding upcoming hardware. A set of recent listings by the retailer PC21, as well as Acer Hong Kong, have revealed some details regarding the upcoming NVIDIA GeForce RTX 5060 Laptop and RTX 5050 Laptop GPUs. First of all, it appears that we now have concrete information regarding the launch date for the RTX 5060 laptop GPU. According to Acer, the RTX 5060 Laptop-equipped producs will start shipping sometime in May. Moreover, as the leaked listing has revealed, RTX 5060 and RTX 5050 Laptop GPUs will arrive with 8 GB of VRAM, both based on the new GDDR7 spec.

This is quite an interesting development, considering that the RTX 5050 Desktop GPU is expected to ship with GDDR6 memory, so it does appear that the Laptop variant will have superior memory speeds. Considering that the RTX 5070 laptop GPU will also boast 8 GB of GDDR7 memory, it is clear that the primary differences between the RTX 5050, RTX 5060, and RTX 5070 Laptop variants will be limited to core counts. Those who demand more than just 8 GB of VRAM will have to shell out extra for the RTX 5070 Ti Laptop, which is expected to ship with 12 GB of GDDR7 memory on a 192 bit bus. As for pricing, the leaked MSI VenturePro 16 with a Core 7 240H CPU, 16 GB DDR5 memory, 512 GB SSD and an RTX 5050 Laptop GPU is seemingly priced at €1580, or roughly $1720. Clearly, even the lowest-end of Blackwell GPUs are all set to boast price tags lofty enough to make even those with deep pockets weep.

The refreshed Mac Studio is here, and it appears that Mark Gurman's reports were accurate once again. The system was updated with the M4 Max and the M3 Ultra SoCs - and once again, that is not a typo. For whatever reason, Apple refused to fit the Mac Studio with an M4-flavored Ultra SoC, instead settling for an undeniably confusing product lineup. The M4 Max, with up to 16 CPU cores and 40 GPU cores, will undoubtedly have the upper hand in single-core performance by as much as 30%, whereas the M3 Ultra will have superior multithreaded and GPU performance, courtesy of its 32 CPU cores and 80 GPU cores. Moreover, the price gap between the base M3 Ultra and M4 Max SKUs will remain the same, despite the former being based on an older generation.

However, the M3 Ultra will allow the system to be configured with up to a whopping 512 GB of unified memory, with memory bandwidth of 819 GB/s. While that number is not particularly mind-bending for a workstation-class system, the fact that the M3 Ultra's 80-core GPU will have access to over half a terabyte of fast-enough memory is a game changer for select few ultra-high-end workloads. Of course, this amount of VRAM is not intended for the average Joe, but the Ultra SoCs were always meant to be a halo product. The M3 Ultra variant can also be equipped with up to 16 TB of storage - at Apple's ridiculous pricing, of course. Needless to say, Apple's performance claims are as vague as always, and interested customers will have to wait for independent reviews and benchmarks to make sense of Apple's confusing SoC strategy with the new Mac Studio.

Morefine is a relatively reputed brand of mini PCs that we have covered in the past. The company has now introduced a powerful new eGPU dubbed the G1, which can also function as a dock, courtesy of its healthy selection of ports. The eGPU can be equipped with the RTX 4060 Laptop, RTX 4080 Laptop, or even the RTX 4090 Laptop GPUs. The product is already available for purchase, with prices starting from $659 for the RTX 4060 L variant, $1,099 for the RTX 4080 L variant, and the highest-end RTX 4090 Laptop variant is unfortunately out of stock as of this writing, which otherwise carries a price tag of $1,699.

For those who need a refresher, the RTX 4060 Laptop packs 3,072 CUDA cores paired with 8 GB of GDDR6 VRAM, allowing for performance that is roughly in the same league as the Radeon RX 7600M GPU found in many other eGPUs. The RTX 4080 Laptop and RTX 4090 Laptop GPUs pack 7,424 and 9,728 CUDA cores coupled with 12 and 16 GB of GDDR6 VRAM respectively. Morefine has revealed the 3D Mark scores for each of the systems, which is certainly a good thing for prospective buyers. The 14 x 10 x 5.4 cm chassis is impressively compact, and as mentioned previously, the ports on offer are equally impressive. An OCuLink port is provided, along with Thunderbolt 3 (85W PD), which can swapped with a dual USB-C module for folks with systems that cannot take advantage of OCuLink. Other ports include triple USB-A 3.2, dual HDMI 2.1, and DisplayPort 1.4 ports.

AMD's internal benchmarks of its upcoming RDNA 4-based RX 9070 series graphics cards have been leaked, thanks to VideoCardz. The flagship RX 9070 XT delivers up to 42% better performance than the Radeon RX 7900 GRE at 4K resolution across a test suite of over 30 games, with the standard RX 9070 showing a 21% improvement in the same scenario. The performance data, encompassing raster and ray-traced titles at ultra settings, positions the RX 9070 series as a direct competitor to NVIDIA's RTX 4080 and RTX 5070 Ti. Notably, AMD's testing methodology focused on native rendering and ray tracing capabilities rather than upscaling technologies like FSR. The RX 9070 XT demonstrated large gains at 4K resolution, achieving a 51% performance uplift compared to the two-generations older RX 6900 XT. Meanwhile, the base RX 9070 model showed a 38% improvement over the RX 6800 XT at 4K with maximum settings enabled.

While AMD confirms its new cards are designed to compete with NVIDIA's RTX 50 series, specific comparative benchmarks against the RTX 5070 Ti were absent from the presentation. AMD acknowledges it has yet to acquire the competitor's hardware for testing. The company is expected to provide a comprehensive performance overview, potentially including additional GPU comparisons, during its official announcement on February 28. Both RX 9070 series cards will feature 16 GB of VRAM, matching the memory configuration of the RX 7900 GRE used as a primary comparison point. By the official launch date, AMD will have time to push final driver tweaks for optimal performance. Nonetheless, more information will surface as we near the official release date.

Recent rumors hinted at a 32 GB variant of the Radeon RX 9070 XT being in the works, which were quickly dismissed as false information by AMD's Frank Azor. However, reliable sources seem to point to the contrary, stating that a 32 GB variant of the RX 9070 XT, likely dubbed the RX 9070 XTX, is under active development indeed. The source, as pointed out by Wccftech, has a decent track record with AMD-related claims, which sure does add weight to the assertion. Unlike previous XTX-class cards from AMD, which boasted higher clock speeds and core counts, the 9070 XTX is almost certain to feature the same core count as the XT, since the latter already utilizes the full Navi 48 chip - unless, of course, there is an even higher-end chip under wraps.

The VRAM amount seems to indicate that the card will likely be positioned to appease AI enthusiasts. There is also the possibility that the rumored card will be launched under a different branding entirely, although that is not what the post at Chiphell states. Interestingly, Frank Azor did specifically mention that a 32 GB "RX 9070 XT" card is not on the horizon - he did not state that a higher-end XTX card isn't either, which sure does leave room for us to speculate. Benchlife has also chimed in on the matter, claiming that they are aware of AIB partners working on a 32 GB RDNA 4 card with the Navi 48 GPU, which in some ways, confirms the information that came out of Chiphell. The RDNA 4 cards are set to see the light of day soon enough, it seems the wait won't be much longer. However, if the 32 GB card is indeed in the pipeline, it's likely still further down the road.

AMD's Radeon RX 9000 series GPUs are expected to come with up to 16 GB of GDDR6 memory. However, AMD is reportedly expanding its RX 9070 lineup with a new 32 GB variant, according to sources on Chiphell. The card, speculatively called the RX 9070 XT 32 GB, is slated for release at the end of Q2 2025. The current GDDR6 memory modules used in GPUs carry a capacity of 2 GB per module only, meaning that a design with 32 GB of VRAM would require as many as 16 memory modules on a single card. No 2 GB+ GDDR6 memory modules are available, meaning that the design would require memory module installation on both the front and back of the PCB. Consumers GPUs are not known for this, but it is a possibility with workstation/prosumer grade GPUs employing this engineering tactic to boost capacity,

While we don't have information on the GPU architecture, discussions point to potential modifications of the existing Navi 48 silicon. This release is positioned as a gaming card rather than a workstation-class Radeon PRO 9000 series product. AMD appears to be targeting gamers interested in running AI workloads, which typically require massive VRAM amounts to run locally. Additionally, investing in a GPU with a big VRAM capacity is essentially "future-proofing" for gamers who plan to keep their cards for longer, as recent games have been spiking VRAM usage by a large margin. The combination of gaming and AI workloads may have made AMD reconsider some of its product offerings, potentially giving us the Radeon RX 9070 XT 32 GB SKU. We have to wait for the Q2 to start, and we can expect more details by then.

Update 20:55 UTC: AMD's Frank Azor on X debunked rumors of the 32 GB SKU coming to gamers. So, this will not happen. Instead, we could be looking at prosumer oriented AMD Radeon Pro GPU with 32 GB of memory instead.

According to early February reportage, Team Green's GeForce RTX 5060 and RTX 5060 Ti graphics cards are expected to launch onto market next month. Very basic technical information has leaked online; insiders reckon that both product tiers will be utilizing the NVIDIA "Blackwell" GB206 GPU. Rumors have swirled regarding intended VRAM configurations—loose online declarations point to variants being prepared with 8 or 16 GB of GDDR7 VRAM, on a 128-bit bus. Regulatory filings indicate two different configs with the eventual arrival of GeForce RTX 5060 Ti models, but certain industry watchdogs insist that the GeForce RTX 5060 SKU will be an 8 GB-only product.

A curious-looking ZOTAC Trinity OC White Edition GeForce RTX 5060 16 GB variant surfaced via a TikTok video—post-analysis, expert eyes declared that the upload contained doctored material. A BenchLife.info report pointed to a notable inconsistency on the offending item's retail packaging: "DLSS 3 should not appear on the GeForce RTX 50 series box, because the Blackwell GPU architecture focuses on DLSS 4." The publication presented evidence of ZOTAC RTX 4070 Ti SUPER Trinity OC White Edition box art being repurposed in the TikToker's video. Hardware soothsayer MEGAsizeGPU added their two cents: "this is fake. There is no plan for a GeForce RTX 5060 16 GB, and the box is photoshopped from the last-gen ZOTAC box." At the end of their report, BenchLife reckons that NVIDIA has not sent a "GeForce RTX 5060 color box template" to its board partners.

It's not every day that we witness a famous NSA whistleblower voice their disappointment over modern gaming hardware. Edward Snowden, who likely needs no introduction, did not bother to hold back his disapproval of NVIDIA's recently launched RTX 5090, RTX 5080, and RTX 5070 gaming GPUs. The reviews for the RTX 5090 have been mostly positive, although the same cannot be said for its affordable sibling, the RTX 5080. Snowden, voicing his thoughts on Twitter, claimed that NVIDIA is selling "F-tier value for S-tier prices".

Needless to say, there is no doubt that the RTX 5090's pricing is quite exorbitant, regardless of how anyone puts it. Snowden was particularly displeased with the amount of VRAM on offer, which is also hard to argue against. The RTX 5080 ships with "only" 16 GB of VRAM, whereas Snowden believes that it should have shipped with at least 24, or even 32 GB. He further adds that the RTX 5090, which ships with a whopping 32 GB of VRAM, should have been available with a 48 GB variant. As for the RTX 5070, the security consultant expressed desire for at least 16 GB of VRAM (instead of 12 GB).

As we reported yesterday, the Radeon RX 9070 XT appears to be all set to disrupt the mid-range gaming GPU segment, offering performance that looks truly enticing, at least if the leaked synthetic benchmarks are anything to go by. The highest-end RDNA 4 GPU is expected to handily outperform the RTX 4080 Super despite costing half as much, with comparison to its primary competitor, the RTX 5070, yet to be made.

Now, a fresh leak has seemingly hinted at how heavy the RDNA 4 GPU is going to be on its buyers' pockets. Also sourced from Chiphell, the Radeon RX 9070 XT is expected to command a price tag between $479 for AMD's reference card and roughly $549 for an AIB unit, varying based on which exact product one opts for. At that price, the Radeon RX 9070 XT easily undercuts the RTX 5070, which will start from $549, while offering 16 GB of VRAM, albeit of the older GDDR6 spec. There is hardly any doubt that the RTX GPU will come out ahead in ray tracing performance, as we already witnessed yesterday, although traditional rasterization performance will be more interesting to compare.

MSI has unveiled its groundbreaking NVIDIA GeForce RTX 50 Series graphics cards, featuring cutting-edge designs including Suprim Liquid, Suprim, Vanguard, Gaming Trio, Ventus, and Inspire. Engineered with enhanced thermal solutions, the cards are crafted to meet the high-performance demands of next-gen GPUs, delivering advanced cooling and peak performance.

Powered by NVIDIA Blackwell, GeForce RTX 50 Series GPUs bring game-changing capabilities to gamers and creators. Equipped with a massive level of AI horsepower, the RTX 50 Series enables new experiences and next-level graphics fidelity. Multiply performance with NVIDIA DLSS 4, generate images at unprecedented speed, and unleash creativity with NVIDIA Studio. Plus, access NVIDIA NIM microservices - state-of-the-art AI models that let enthusiasts and developers build AI assistants, agents, and workflows with peak performance on NIM-ready systems.

By now, it's a surprise to almost nobody that NVIDIA plans to launch its next-generation RTX 5000-series "Blackwell" gaming graphics cards at the upcoming CES 2025 event in Las Vegas in early January. Previously, leaks and rumors gave us a full run-down of expected VRAM amounts and other specifications and features for the new GPUs, but these have yet to be confirmed by NVIDIA—for obvious reasons. Now, though, it looks as though iBuyPower has jumped the gun and prematurely revealed the new specifications for its updated line-up of pre-built gaming PCs with RTX 5000-series GPUs ahead of NVIDIA's official announcement. The offending product pages have since been removed, but they both give us confirmation of the previously leaked VRAM amounts and of the expected release cadence for RTX 5000, which will reportedly see the RTX 5070 Ti and RTX 5080 launch before the RTX 5090 flagship.

On iBuyPower's now-pulled pages, the NVIDIA GeForce RTX 5080 16 GB and GeForce RTX 5070 Ti 16 GB can be seen as the GPUs powering two different upcoming Y40 pre-built gaming PCs from the system integrator. The VRAM specifications here coincide with what we have previously seen from other leaked sources. Unfortunately, while an archived version of the page for the pre-built containing the RTX 5080 appears to show the design for an ASUS TUF Gaming RTX 5080 with a triple-fan cooler, it looks like iBuyPower is using the same renders for both the 5080 and 5070Ti versions of the pre-built PCs. What's also interesting is that iBuyPower looks to be pairing the next-gen GPUs with 7000-series AMD X3D CPUs, as opposed to the newly released AMD Ryzen 9000 X3D chips that have started making their way out into the market.