The "standard" ASUS ROG Astral RTX 5090 OC Edition is already a really an "astronomically expensive" prospect (if you can find available stock); launch MSRP was $2800, but retailers pushed that figure beyond the $3000 mark. An even fancier golden variant exists—as a reminder; news outlets picked up on the existence of a "Dhahab" model early last month. This luxuriously decorated collector piece was likely introduced as a regional exclusive, for MENA (Middle-East and North Africa). This week, ASUS managed to sneak one gold encrusted sample out to San Francisco, California.

Ernest Cheng—Director of Marketing at the firm's North American branch—shared a photo (via LinkedIn) of the very unique ROG Astral RTX 5090 Dhahab OC Edition model; freshly scribbled on by Jensen Huang during GTC 2025. The ASUS exec commented on this blessing: "graphics card is one of a kind when it's been anointed. But it says a lot more when it's a Golden ROG RTX 5090 Astral." Press outlets reckon that this extremely special item will be auctioned off for charity; the presence of Team Green CEO's autograph and slogan ("RTX ON!") will boost its value severalfold. A Jensen Huang-signed ROG MATRIX RTX 4090 PLATINUM card attracted a top bid of $16,000; Der8auer (aka Roman Hartung) was officially congratulated as the winner back in late 2023.

Update 18:02 UTC: ASUS has confirmed that it will be supporting a local charity: "we are extremely honored to have this special edition ROG Astral RTX 5090 Dhahab OC graphics card, signed by NVIDIA CEO Jensen Huang. This card will be officially auctioned to support relief efforts for the California wildfires in Los Angeles."

Yesterday evening, the NVIDIA AI Developer social media account sent out a red alert regarding a time limited sale of flagship-tier Blackwell gaming hardware: "GeForce RTX 5090s are available at the Gear Store in the park right now at GTC 2025. 90 units are available for the next 30 minutes, with more coming tomorrow. Come say hi!" PC hardware news outlets have picked up on Team Green's latest stock bulletin, with Tom's Hardware disclosing some extra details. Under normal circumstances, NVIDIA's Gear Store Mobile Truck would be selling fairly standard merchandise—e.g. T-shirts, sweaters, hats, etc. According to the latest reports, the company's mobile pop-up store is taking orders for add-in-boards (AIB) GeForce RTX 5090 and RTX 5080 models. An information placard advertises old school/launch day guide prices of $1999 and $999 (respectively).

Tom's Hardware noted several caveats: "the graphics cards must be purchased from NVIDIA's van from 7 AM to 12 PM on Thursday or Friday and then picked up at South Hall main entrance the same day. The graphics boards are available to conference pass ($1145 for one day, or $2295 for five days) and exhibit pass holders only; with a limit of one card per person." According to folks on the ground, Team Green and its board partners have stockpiled a thousand of each highly-desirable GPU model at the San Jose Convention Center. The first waves of time-limited batches were made available yesterday (March 19). Demand for flagship and top-end GeForce RTX 50 series cards has far exceeded supply, starting back in late January. Following an absorption of plentiful feedback, NVIDIA revived its "Verified Priority Access" scheme a couple of weeks ago. This anti-scalping initiative was advertised as offering: "a limited number of verified GeForce gamers and creators in the United States the opportunity to purchase one GeForce RTX 5090 or RTX 5080 Founders Edition graphics card from the NVIDIA Marketplace."

Graphics card price watchers have highlighted refreshing downward motion in Europe, apparently affecting three out of the four GeForce RTX 50 series graphics cards. VideoCardz received a couple of email tip-offs from its pan-European audience, prompting the publication of a short investigative piece. NVIDIA's slight adjustment of official pricing for GeForce RTX 5090, RTX 5080, and RTX 5070 models is the result of a strengthened Euro. The US dollar's value has dropped by roughly 3.9 %; according to recent detective work, focusing on German trends. Team Green's "generous" reductions have arrived roughly two weeks after a stabilization of the USD-EUR exchange rate.

Curiously, the GeForce RTX 5070 Ti is an outlier here—NVIDIA did not reduce its German guide price (€879 + VAT) for this upper-mid-range offer. A Founders Edition does not exist at this GPU level, so Team Green has tasked its board partners with the creation of so-called "MSRP conformant" alternatives. One of VideoCardz's tipsters has observed various GeForce RTX 50 series models simply "rotting on shelves," due to potential buyers balking at unreasonable retailer-implemented price hikes. NVIDIA's minor changes (4.3 to 4.6 %) are unlikely to make a noticeable impact across the Euro zone.

NVIDIA today launched the RTX PRO 6000 Blackwell series of professional graphics cards. These cards are based on the latest GeForce "Blackwell" graphics architecture, and the three chips the company already launched on it. Leading the pack, is the RTX PRO 6000, a card that completely maxes out the massive "GB202" silicon, featuring more shaders than even the GeForce RTX 5090, albeit at lower clock speeds. The idea behind this product is to give pro-vis users more shader power, driving a large amount of GDDR7 ECC memory. Specifically, the card comes with 24,064 CUDA cores across all 192 SM physically present on the silicon, besides 768 Tensor cores, 192 RT cores, 768 TMUs, and 192 ROPs. The card gets a humungous 96 GB of ECC GDDR7 memory across the chip's 512-bit wide memory interface, probably using 48 Gbit density memory chips. The card has a TGP of 600 W, making out the 12V2x6 power input. It comes with a board design resembling the RTX 5090.

Next up, is the RTX PRO 6000 Blackwell Max-Q. This card has essentially the same core-configuration as the RTX PRO 6000, but with a reduced TGP, and a simpler 2-slot board design that uses a lateral-blower. This card is meant for machines with multiple such cards installed, though something that isn't quite a rendering server. Lastly, there's the RTX PRO 6000 Server Edition. This card, again, has identical core-config to the others in the lineup, but with a board design optimized for rackmount servers and large rendering farms. The cooler relies on the rack's airflow for cooling.

"Second wave" ASUS price hikes were documented online over the past weekend; affecting air-cooled premium ROG Astral and mid-tier TUF Gaming GeForce RTX 5090 models. Looking at the company's North American webshop, visitors noticed a freshly adjusted price for the ROG Astral GeForce RTX 5090 32 GB OC Edition—going from a previous level of $3079.99 up to $3359.99. Curiously, the asking price of a liquid-cooled sibling was not adjusted—remaining at a "first wave" point of $3409.99. The "cheapest" model—TUF Gaming GeForce RTX 5090 (non-OC)—experienced a $460 (representing 20%) price hike, bringing total cost of ownership up to $2759.99. As a reminder, NVIDIA's baseline MSRP guideline was $1999—as announced at CES 2025—but ROG Astral and TUF Gaming designs demand a premium or two for fancier feature sets. VideoCardz has fervently explored worrying market trends in the recent past; several of NVIDIA's big board partner players have jacked up asking prices for GeForce RTX 50-series graphics cards. Availability of stock is still a major sore point for potential buyers, who were not able to secure launch day wares. Despite a driving up of costs, the ASUS US webstore has absolutely zero stock of GeForce RTX 5090 SKUs—at the time of writing.

In addition, VideoCardz and other PC hardware media outlets noted price hikes affecting the manufacturer's stable of recently launched AMD Radeon RX 9070 Series TUF Gaming and PRIME models. In the absence of AMD-built (MBA) reference card designs, board partners were tasked with the providing of baseline "MSRP" conformant custom cards. The ASUS PRIME Radeon RX 9070 XT OC and RX 9070 OC Editions were readied as $599 and $549 options (respectively). Weekend sleuthing work put the spotlight on newly adjusted price points of $719.99 and $659.99 (respectively)—representing further cases of plain 20% elevations over baseline. AMD's debut batch of RDNA 4 cards was met with unprecedented demand earlier on in March, but secondary/tertiary stock shipments face unclear market conditions—Team Red GPU enthusiasts have (similarly) voiced their collective displeasure about elevated prices at retail. Mid-way through last week, the PC hardware community heard about ASUS leadership considering a new pricing strategy. The company is reportedly accelerating its manufacturing exodus from China.

As reported last month, GALAX had distributed prototypes of its upcoming flagship "Hall of Fame" (HOF) card—based on NVIDIA's Chinese market exclusive GeForce RTX 5090D GPU—to prominent figures within the PC hardware overclocking community. Earlier examples sported single 12V-2×6 power connectors, although GALAX's exposed white PCB design showed extra space for an additional unit. Evaluators conducted experiments involving liquid nitrogen-based cooling methods. The most vocal of online critics questioned the overclocking capability of initial GeForce RTX 5090D HOF samples, due to limitations presented by a lone avenue of power delivery. A definitive answer has arrived in the form of the manufacturer's elite team-devised GeForce RTX 5090D HOF Extreme Overclock (XOC) Lab Limited Edition candidate; a newer variant that makes use of dual 12V-2×6 power connectors. Several overclocking experts have entered into a GALAX-hosted competition—Micka:)Shu, a Chinese participant, posted photos of their test rig setup (see below).

Micka's early access sample managed to achieve top placement GPU on UL Benchmarks' 3DMark Speed Way Hall of Fame, with a final score of 17169 points. A screenshotted GPU-Z session shows the card's core frequency reaching 3277 MHz. Around late January, ASUS China's general manager (Tony Yu) documented his overclocking of a ROG Astral RTX 5090 D GAMING OC specimen up to 3.4 GHz; under liquid nitrogen cooled conditions. GALAX has similarly outfitted its flagship model with selectively binned components and an "over-engineered" design. The company's "bog-standard" HOF model is no slouch, despite the limitation imposed by a single power connector. The GALAX OC Facebook account sent out some appreciation to another noted competitor (and collaborator): "thanks to Overclocked Gaming Systems—OGS Rauf for help with the overclock of GeForce RTX 5090D HOF, and all of (our) GALAX products." The OGS member set world records with said "normal" HOF card—achieving scores of 59,072 points in 3DMark's Fire Strike Extreme project, and 25,040 points in Unigine Superposition (8K-optimized).

NVIDIA has prepared a "Blackwell" Titan equivalent—RTX PRO 6000 Blackwell GPU for professional visualization and local AI solutions. Based on the GB202 silicon, the RTX PRO 6000 is the closest solution to the full power of NVIDIA's prosumer-oriented Blackwell SKUs. With 24,064 CUDA cores on board, this configuration is just 512 CUDA cores shy from the complete GB202 24,576-core configuration. This is likely due to the yield defects, meaning that this is perhaps the highest CUDA core count model we will see based on GB202. Additionally, at 24,064 core RTX PRO 6000 Blackwell is carrying 2,340 more cores than the top-end consumer GeForce RTX 5090 with 21,760 CUDA cores. Based on the 600 W TGP, the new RTX PRO 6000 Blackwell should have lower clocks, given its bigger core count to maintain the TGP for its massive memory configuration. Remember, GDDR7 memory modules consume power, too.

With 96 GB of GDDR7 memory and ECC memory correction on board, it will primarily target professionals in 3D rendering, simulations, and local AI development. To cool the beastly configuration, NVIDIA opted for a double-flow-through cooler used on the RTX 5090, with an open-air design. Operating on a PCIe 5.0 x16 interface with four DisplayPort 2.1 connectors, the design remains double-slot in thickness. VideoCardz obtained pictures of it, which you can see below. Interestingly, the color accents of the new cooler are darker. We can expect to see more from NVIDIA and its new RTX PRO 6000 Blackwell GPU from the upcoming GTC 2025 on March 17.

Leadtek has expanded its graphics card lineup into GeForce RTX 50-series territory, with a quiet introduction of three new generation WinFast models. We have not heard much—gaming-wise—from the Taiwanese manufacturer in recent years. Going back to Computex 2023, TechPowerUp staffers inspected several GeForce RTX 40-series offerings. Leadtek is involved in the production of NVIDIA professional graphics solutions—for data science and HPC workload purposes. Curiously, Team Green press material tends to only lightly mention its manufacturing partner's involvement. The newest WinFast entries have rolled out with "Blackwell" silicon and refreshed enclosures. As reported by VideoCardz, Leadtek has focused mainly on Asian PC gaming markets over the past decade—so the brand-new WinFast RTX 5090 HURRICANE 32G, RTX 5080 HURRICANE 16G, and RTX 5070 Ti HURRICANE 16G models are not expected to arrive on Western shores.

The 2025 WinFast HURRICANE package utilizes a triple-fan array paired with a robust heatsink design, and a "revamped" metal backplate. A total of ten heat pipes are advertised as providing strong heat dissipation—Leadtek's product description goes into more detail: "three 10 mm and seven 8 mm heat pipes, along with large aluminium fins and a vapor chamber base, quickly and efficiently dissipate heat." The three freshly introduced models all conform to NVIDIA reference specifications, so the HURRICANE cooling solution seems to be a bit over-engineered for purpose. Price details are not available at the time of writing, and an official press release is not present on Leadtek's website. The WinFast RTX 5090 HURRICANE card seems to sport a chunkier shroud, when lined up against its less powerful siblings. Leadtek has decorated its flagship with "Esports" ARGB lighting strips, while the RTX 5080 and RTX 5070 Ti variants make do with silver-ish accents instead.

Taiwanese press outlets have put a spotlight on an inexplicably devised "GIGABYTE RTX 5090 Super Bundle." According to regional reports, a local company—Deyuan Computer—has listed this unusual multi-piece package on Taiwan's MOMO e-commerce platform. GeForce RTX 5090 graphics cards are a hot commodity nowadays—unfortunately, limited supplies and scalping activities have created a volatile environment for regular punters. Retail outlets have implemented various "nifty" sales methods in attempt to create a fairer buying environment—pre- and post-launch. Hong Kong's HKEPC Media has acknowledged that locals are very used to a "bundling culture" that requires the purchase of superfluous extra items.

By comparison, these offerings look quite reasonable when lined up against a Taiwanese business dealing in staggeringly silly combination platters. To claim ownership of a single GIGABYTE RTX 5090 GAMING OC graphics card, potential buyers are strong-armed into adding forty(!) GIGABYTE H610M-K-DDR4 motherboards to virtual shopping carts. A grand total expenditure of NT$191,500 (~$5832 USD) unlocks "GIGABYTE RTX 5090 Super Bundle" ownership. By itself, the high-end/award-winning GIGABYTE graphics card would cost NT$91,900 (~$2796 USD). An individual entry-level H610M-K-DDR4 Micro ATX model comes in at NT$2490 (~$76 USD). Willing customers will have to deal with an inventory of slightly outdated motherboards—a full build's worth of components would make more sense. Around late January, Taiwan's AutoBuy advertised an anti-scalping deal for GeForce RTX 5080 cards—requiring shoppers to load up on parts for a near complete Intel Core Ultra 200S CPU-based system.

A formerly dormant member of China's Chiphell discussion board has resurfaced with a major GALAX leak; they claimed ownership of a pre-release GeForce RTX 5090D HOF OC LAB edition graphics card. The Hong Kong-based AIB did not update its "Hall of Fame" series with flagship NVIDIA "Blackwell" GPU silicon in time for last month's launch. Their highest-end offering arrived in the form of a GeForce RTX 5080 HOF OC LAB Plus-X SKU; reportedly an over-specced juggernaut. According to the latest reports, GALAX's top GeForce RTX 5090D candidate is in the pipeline—early details indicate an impressive feature set. The manufacturer is well known for its selective methodology; GPU industry watchdogs believe that the incoming flagship utilizes only the very best pre-binned GB202 GPU dies and GDDR7 memory modules.

Despite a fitting of elite-tier parts, the lucky owner reported issues with their allegedly sanction box-ticking card: "this time the overclocking was more troublesome. The bracket and backplate of the graphics card gun had to be remade. Finally, it was done. I won't say more. I'm going to fill it with liquid nitrogen." VideoCardz believes that GALAX has sent preview samples to local overclocking specialists, minus finalized air-cooled solution attachments. As disclosed by the Chiphell member
(laoshu1919), early evaluators appear to be experimenting with liquid cooling. A photo of GALAX's GeForce RTX 5090D HOF OC LAB PCB design was shared on the forum; showing a grand total of 38 high-quality power phases. The GPU die is surrounded by 26 of these units, while 10 are positioned closer to the card's VRAM. Another 2 are located nearby to a single 16-pin power connector, with enough extra room to accommodate two more. Judging from laoshu1919's photo, this prototype seems capable of hosting another 16-pin power connector. The sample's I/O bracket likely sports GALAX's signature "1-Click OC" button.

The new ASUS ROG Astral graphics card design debuted last month, with the rollout of NVIDIA GeForce RTX 5090 and 5080 "Blackwell" GPUs. The flagship model—in overclocked form—is not a wallet-friendly prospect; as highlighted by W1zzard's in-depth evaluation. The "astronomically-priced" premium-tier quad-fan model is a hot property; in more ways than one—late last week, an unfortunate ownership experience was shared online. NVIDIA subreddit member—Impossible-Weight485—uploaded photo evidence, accompanied by a short story: "I was playing PC games this afternoon, and when I was done with the games, my PC suddenly shut down while I was browsing websites. When I restarted the PC, the GPU caught on fire, and smoke started coming out. When I took out the GPU, I saw burn marks on both the GPU and the motherboard." Post-absorption, initial community and press feedback posited that the problem originated with a Multi-Layer Ceramic Capacitor (MLCC), located not far from the card's PCI-E interface.

High-profile figures soon swooped in, with different theories and offers. A Team Green subreddit moderator weighed in: "not adding this one to our GeForce RTX 50 Series 12VHPWR Megathread. This looks to be a blown power phase, and not melting power connector. The original poster provided additional photos of the cable, in addition to the GPU connector photo in the post. Both looks pristine...Yes, I watched Buildzoid's video (see below), hence updating this comment...Thanks to Buildzoid for the education!" The owner uploaded another interior shot, seemingly showing burn damage on their ASUS ROG CROSSHAIR X870E HERO motherboard. The severity of this incident attracted the attention of Gamers Nexus—Lelldorianx (aka Stephen Burke) reached out to the damaged card's owner: "messaging you. We'd buy the board and GPU from you if you want to just take the cash and buy something else (or) skip the RMA process." Burke and his colleagues are actively investigating various GeForce RTX 50-series "pratfalls"—earlier this month, reports indicated that the team was already engaged in the sourcing of problematic units.

It would be safe to state that the NVIDIA GeForce RTX 5090 and RTX 5080 launch was anything but ideal. Gamers had to deal with whacky NVIDIA marketing material with absurd performance claims, followed by disappointing generational improvement for the RTX 5080, only to be left dealing with abysmal supply leading to obscene shortages and scalper-induced price inflation. However, it does seem like things are about to take a positive turn - NVIDIA is rumored to have ramped up production for its GB202 GPU, which the RTX 5090 is based on, according to a reliable source.

Spotted by VideoCardz, MEGAsizeGPU has claimed that the supply for the GeForce RTX 5090 GPU will soon be "stupidly high", which is absolute music to our ears. In a reply thread, the source further claimed that at least one AIB partner already has "tons of cards", which sure does paint a promising picture for the future. As such, the source expects that the supply will reach customers in about a month, which is to be expected since production has been cranked only recently. Apparently, demand for the GB200 GPU has been lower than usual, forcing NVIDIA to switch to producing GeForce GPUs instead. Of course, the margins for the gaming GPUs are lower, but the production capacity has to go somewhere.

ASUS has released version one of a BIOS update tool for its ultra-premium air-cooled ROG Astral graphics cards, OC and standard flavors. Yesterday's update advertises an improved "Quiet Mode," that implements a "more silent fan curve." TechPowerUp reviewed the quad-fan configured ASUS ROG Astral GeForce RTX 5090 and RTX 5080 OC Edition cards late last month. Resident GPU evaluator, W1zzard, awarded the astronomically-priced flagship model with an "Editor's Choice" badge, but simultaneously pinned on a "But Expensive" honor. The lesser (GB203-based) Astral did not receive any accolades. Both models have courted criticism for louder than expected operation; W1zzard did not enjoy listening to the Astral GeForce RTX 5090 OC Edition's collective fan concerto. "Out of the box" settings were not great, and the quiet BIOS mode did not "help much" in reducing the highest-end Astral's cacophony.

TPU's GPU guru elaborated further: "the second BIOS runs a more relaxed fan curve, but it's not much quieter and achieves 36.5 dBA with 70°C. Sure, good temperature, but isn't the point of a 'quiet' BIOS that isn't quiet, even if temperatures are higher? MSI's Suprim lineup does much better noise-wise, with temperatures that aren't that much higher." The ASUS ROG Astral RTX 5080 OC Edition seems to be a less noisy prospect, given that its cooling solution is not dealing with the largest "Blackwell" GPU die. Its sound signature was less offensive, but W1zzard reckoned that there was room for improvement. On this subject he stated: "with the default (performance) BIOS, temperatures are extremely low, but noise levels are a little bit on the high side with 36 dBA. I would have preferred a more balanced setting. Good thing that ASUS includes a secondary quiet BIOS with their card. Now the card runs whisper quiet, emitting only 26 dBA, which is highly impressive for a card in this performance segment. It is not the quietest card though, the MSI Suprim SOC is a tiny bit quieter, and it is so out of the box, without requiring a manual BIOS switch change." ASUS has seemingly absorbed initial feedback from review outlets (plus early adopters)—yesterday's update arrived just under two weeks from launch time. Watch out for possible upcoming reassessments.

German overclocking specialist Der8auer has showcased a custom water block for NVIDIA's flagship RTX 5090 Founders Edition, achieving a remarkable 30-degree Celsius reduction in core GPU temperatures. The prototype copper cooling solution demonstrated peak operating temperatures of just 43.8°C, compared to the stock cooler's 73.8°C under identical testing conditions. The design features an all-copper construction with a substantial 14 mm-thick cold plate, engineered specifically for optimal thermal transfer from the GB202 die. While maintaining the dual-slot form factor of the original Founders Edition cooler, the water block incorporates multiple cooling channels beneath a transparent top panel that reveals the coolant flow.

Memory temperatures also significantly improved, dropping by 38 degrees Celsius below stock measurements. Despite its impressive thermal performance, the prototype faces several engineering challenges. The team was constrained by the graphics card's display output ribbon cable, which prevented a more compact design closer to the actual PCB dimensions. Additionally, the Founders Edition card's unique split PCIe finger design presented structural concerns that require additional reinforcement before any potential commercial release. Der8auer's prototype prioritizes function over aesthetics, omitting common features like RGB lighting in favor of raw cooling performance. The water block's industrial design reflects its proof-of-concept nature, though the transparent top panel offers practical utility for monitoring coolant flow. While no retail release timeline has been announced, Der8auer indicated that addressing the PCIe slot structural support issue remains a critical milestone before any commercial version could be considered.

A die-shot of NVIDIA's GB202, the silicon powering the RTX 5090, has surfaced online, providing detailed insights into the "Blackwell" architecture's physical layout. The annotated images, shared by hardware analyst Kurnal and provided by ASUS China general manager Tony Yu, compare the GB202 to its AD102 predecessor and outline key architectural components. The die's central region houses 128 MB of L2 cache (96 MB enabled on RTX 5090), surrounded by memory interfaces. Eight 64-bit memory controllers support the 512-bit GDDR7 interface, with physical interfaces positioned along the top, left, and right edges of the die. Twelve graphics processing clusters (GPCs) surround the central cache. Each GPC contains eight texture processing clusters (TPCs), with each GPC housing 16 streaming multiprocessors (SMs). The complete die configuration enables 24,576 CUDA cores, arranged as 128 cores per SM across 192 SMs. With RTX 5090 offering "only" 21,760 CUDA cores, this means that the full GB202 die is reserved for workstation GPUs.

The SM design includes four slices sharing 128 KB of L1 cache and four texture mapping units (TMUs). Individual SM slices contain dedicated register files, L0 instruction caches, warp schedulers, load-store units, and special function units. Central to the die's layout is a vertical strip containing the media processing components—NVENC and NVDEC units—running from top to bottom. The RTX 5090 implementation enables three of four available NVENC encoders and two of four NVDEC decoders. The die includes twelve raster engine/3D FF blocks for geometry processing. At the bottom edge sits the PCIe 5.0 x16 interface and display controller components. Despite its substantial size, the GB202 remains smaller than NVIDIA's previous GH100 and GV100 dies, which exceeded 814 mm². Each SM integrates specialized hardware, including new 5th-generation Tensor cores and 4th-generation RT cores, contributing to the die's total of 192 RT cores, 768 Tensor cores, and 768 texture units.

Did you catch our launch review of the NVIDIA GeForce RTX 5090 Founders Edition graphics card? If you did, you'd have noticed the hoops we had to jump through to disassemble the card without breaking anything. As we carefully peeled the card's aesthetic finery and worked out way down to the GPU, we found that NVIDIA is using a liquid-metal thermal interface material (TIM) between the "GB202" GPU and the unified vapor chamber plate of the Double Flow Through cooling solution. After our photography the problem arose of putting the card back together, and we wondered if using a mainstream thermal paste would be a suitable replacement since we didn't have the patience to toy with liquid metal given our review volume. So we pulled out our trusty tube of Arctic MX6, gave it a suitable application, skipped the gasket, and put the card back together.

Here's what we found—the GPU temperatures rise by about 2°C on average over the stock liquid metal TIM. In the graph below, you will see both TIM applications compete with each other over a 370-second stress from a benchmark run. In the chart the start temperatures are slightly different, this does not mean that idle temperatures are higher with liquid metal. At the end of test the maximum temperature reached with the stock liquid metal TIM is around 77.6°C. The Arctic MX6, on the other hand settles to 79.4°C. This +1.8°C temperature increase really isn't significant at all—room temperature changes between summer and winter will cause bigger swings. We also tested performance, and it was spot on, the same as pre-disassembly—not a hint of thermal throttling. Both values are safely below the 90°C thermal threshold for the RTX 5090—that's right, NVIDIA raised the thermal throttle point, it's not 83°C anymore like on the RTX 40-series Ada. Taking the RTX 5090 apart and putting it back together was a challenging experience, but we're glad we didn't have to do a liquid metal application to ensure trouble-free operation. This is good news for all DIYers—don't bother with replacing the liquid metal—a thin thermal paste application works fine, too.

Recent shipping manifests suggest that NVIDIA works on a graphics card with 96 GB of GDDR7 memory. Documents reveal a product utilizing a 512-bit memory bus and a clamshell (memory on both PCB sides) design that combines two 3 GB modules per memory controller. This setup effectively doubles the memory capacity of existing workstation-oriented cards. The product is believed to use the GB202 chip, the only Blackwell desktop GPU with a 512-bit interface. The documents refer to a board labeled PG153, a designation not seen in any of NVIDIA's existing consumer GPUs. This finding points toward a professional or workstation model rather than a gaming product. There is a possibility that it could be part of the RTX 6000 Blackwell or RTX 8000 Blackwell series.

NVIDIA's current top workstation card, the RTX 6000 "Ada," features 48 GB of memory. A move to 96 GB would be a substantial jump, enabling more complex workloads for content creation, data analysis, and AI. This GPU could carry a significantly higher power target than current workstation models. However, professional GPUs often maintain lower clock speeds to keep power consumption within limits that accommodate more stable operation in professional environments. There is no confirmed information regarding the card's official name or final specifications, such as core count or actual clock frequencies. NVIDIA's workstation GPUs have historically provided a higher core count than their gaming counterparts. If the rumored 96 GB GPU follows this pattern, it may surpass even the potential GeForce RTX 5090, which comes with 32 GB of GDDR7. NVIDIA is expected to hold its annual GPU Technology Conference in March. This event is viewed as a likely venue for official announcements. Until then, these details remain unverified.

Is NVIDIA reviving the RTX TITAN brand of halo-segment graphics cards with "Blackwell"? A curious-looking GB202-200-A1 spy-shot making rounds on ChilHell hints at the possibility. The upcoming GeForce RTX 5090 is the company's flagship product from the RTX 50-series "Blackwell" generation, although it does not max out the 4 nm "GB202" silicon on which it is based. The RTX 5090 enables 170 out of the 192 SM (streaming multiprocessors) physically present on the "GB202." This leaves NVIDIA with a lot of room to carve out either a halo-segment SKU such as the RTX TITAN Blackwell, or a Pro-Vis (professional visualization) product that also targets the AI research community.

NVIDIA's top Pro-Vis product tended to have more SMs enabled than the top GeForce RTX product, while having lower clock speeds, so its target users have access to more FP64-capable CUDA cores, more Tensor cores, etc. However, over the past two generations, NVIDIA discontinued the practice of giving its RTX GPUs large numbers of FP64 cores that are disabled on GeForce RTX products, to make die-space for the Tensor cores. This hence makes it more likely that a maxed-out "GB202" is the RTX TITAN Blackwell, and not a Pro-Vis product.

ZOTAC's website has been updated with technical specifications for its GeForce RTX 5090 GPU-based custom models—the Hong Kong-based hardware company is perhaps the first NVIDIA board partner to publicly disclose these details. According to VideoCardz, this discovery was made by a loyal reader—product pages for SOLID, SOLID OC and AMP Extreme INFINITY seemed to have appeared online within the past couple of days. A resultant report suggests that Team Green has only recently communicated (potentially) finalized specs with its AIBs.

The aforementioned discoverer of ZOTAC's GeForce RTX 5090 spec sheets noticed an unusual memory clock figure on the SOLID OC model's listing—30 Gbps instead of 28 Gbps. This was an error—ZOTAC has since amended that particular data point (see VideoCardz's screenshots below)—their presumably more expensive AMP Extreme INFINITY card's memory clock spec is set at 28 Gbps. Interestingly, ZOTAC's upcoming flagship is the first example of an NVIDIA GPU configured with a power consumption rating of 600 W. It is not immediately apparent whether this TDP figure is an out-of-the-box default—VideoCardz reckons that the GeForce RTX 5090 AMP Extreme INFINITY will arrive with dual-BIOS functionality. A high-performance mode could be user selected. Will rival flagship GeForce RTX 5090 custom cards roll out with similar TDPs? TechPowerUp anticipates the emergence of pre-launch technical details—from other brands/manufacturers—over the next two weeks.

At the 2025 International CES, we went hands on with the new NVIDIA GeForce RTX 5090 Founders Edition "Blackwell" graphics card. This thing is huge—longer and taller than the RTX 4090 FE, and yet just 2 slots thick. This is because NVIDIA's designers have figured out that the heat dissipation area of the heatsink lost to thinning the card can be recovered by stretching it in other directions. The card retains the essential aesthetic of Founders Edition cards from the past two generations going back to the RTX 30-series, but changes the concept of the dual-axial flow-through.

While past generations used an intake fan on one side, blowing onto the PCB, and another fan at the tail end of the backplate pull air through the heatsink and out the back, the RTX 5090 FE has two large fans, both of which blow cold air through the heatsink, and out the back of the card. The PCB is located in the center of the card, and relies on a set of breakout PCBs for host interface and display outputs.

NVIDIA's unannounced GeForce RTX 5090 graphics card has leaked, confirming key specifications of the next-generation GPU. Thanks to exclusive information from VideoCardz, we can see the packaging of Inno3D's RTX 5090 iChill X3 model, which confirms that the graphics card will feature 32 GB of GDDR7 memory. The leaked materials show that Inno3D's variant will use a 3.5-slot cooling system, suggesting significant cooling requirements for the flagship card. According to earlier leaks, the RTX 5090 will be based on the GB202 GPU and include 21,760 CUDA cores. The card's memory system is a significant upgrade, with its 32 GB of GDDR7 memory running on a 512-bit memory bus at 28 Gbps, capable of delivering nearly 1.8 TB/s of bandwidth. This represents twice the memory capacity of the upcoming RTX 5080, which is expected to ship with 16 GB capacity but 30 Gbps GDDR7 modules.

Power consumption has increased significantly, with the RTX 5090's TDP rated at 575 W and TGP of 600 W, marking a 125-watt increase over the previous RTX 4090 in raw TDP. NVIDIA is scheduled to hold its CES keynote today at 06:30 pm PT time, where the company is expected to announce several new graphics cards officially. The lineup should include the RTX 5090, RTX 5080, RTX 5070 Ti, RTX 5070, and an RTX 5090D model specifically for the Chinese market. Early indications are that the RTX 5080 will be the first card to reach consumers, with a planned release date of January 21st. Release dates for other models, including the flagship RTX 5090, have not yet been confirmed. The RTX 5090 is currently the only card in the RTX 50 series planned to use the GB202 GPU. Pricing information and additional specifications are expected to be revealed during the upcoming announcement.

According to two of the most accurate leakers, kopite7kimi and hongxing2020, NVIDIA's GeForce RTX 5090 and RTX 5080 will feature 575 W and 360 W TDP, respectively. Previously, rumors have pointed out that these GPU SKUs carry 600 W and 400 W TGPs, which translates into total graphics power, meaning that an entire GPU with its RAM and everything else draws a certain amount of power. However, TDP (thermal design power) is a more specific value attributed to the GPU die or the specific SKU in question. According to the latest leaks, 575 Watts are dedicated to the GB202-300-A1 GPU die in the GeForce RTX 5090, while 25 Watts are for GDDR7 memory and other components on the PCB.

For the RTX 5080, the GB203-400-A1 chip is supposedly drawing 360 Watts of power alone, while 40 Watts are set aside for GDDR7 memory and other components in the PC. The lower-end RTX 5080 uses more power than the RTX 5090 because its GDDR7 memory modules reportedly run at 30 Gbps, while the RTX 5090 uses GDDR7 memory modules with 28 Gbps speeds. Indeed, the RTX 5090 uses more modules or higher capacity modules, but the first-generation GDDR7 memory could require more power to reach the 30 Gbps threshold. Hence, more power is set aside for that. In future GDDR7 iterations, more speed could be easily achieved without much more power.

Fresh details have surfaced about NVIDIA's upcoming flagship "Blackwell" graphics card, the GeForce RTX 5090, suggesting power delivery and board design changes compared to its predecessors. According to Benchlife, the new Blackwell-based GPU will feature a new 16+6+7 power stage design, departing from the RTX 4090's 20+3 phase configuration. The report confirms earlier speculation about the card's power requirements, indicating a TGP of 600 watts. This specification refers to the complete power allocation for the graphics subsystem, though the actual TDP of the GB202 chip might be lower. The RTX 5090 will ship with 32 GB of next-generation GDDR7 memory and utilize a 14-layer PCB, possibly due to the increased complexity of GDDR7 memory modules and power delivery. Usually, GPUs max out at 12 layers for high-end overclocking designs.

The upcoming GPU will fully embrace modern connectivity standards, featuring PCI Express 5.0 x16 interface compatibility and implementing a 12V-2×6 power connector design. We spotted an early PNY RTX 5090 model with 40 capacitors but an unclear power delivery setup. With additional power phases and more PCB layers, NVIDIA is pushing the power delivery and signal integrity boundaries for its next-generation flagship. While these specifications paint a picture of a powerful gaming and professional graphics solution, questions remain about the broader RTX 50 series lineup. The implementation of the 12V-2×6 connector across different models, particularly those below 200 W, remains unclear, so we have to wait for the CES-rumored launch.

Here is the first picture of what is very likely the GeForce RTX 5090 "Blackwell," the successor to the RTX 4090 "Ada." The picture, its massive GPU, and layout appear to confirm the weekend's bare PCB leak. The RTX 5090 is powered by the "GB202" silicon, the largest gaming GPU based on the "Blackwell" graphics architecture. The silicon in the picture has the ASIC code "GB202-300-A1." From this ASIC code, we can deduce that the RTX 5090 may not max out the silicon (i.e. enable all SM present on it), as maxed-out NVIDIA ASICs tend to have the variant designation "450."

The "GB202" ASIC is surrounded by sixteen GDDR7 memory chips, which reportedly make the 32 GB memory size of the RTX 5090. The chip count, coupled with the large GPU package size (high pin-count), confirm that the "GB202" features a 512-bit wide memory bus. Assuming a memory speed of 28 Gbps, this memory bus should yield a stellar memory bandwidth of 1,792 GB/s. The GPU and memory are surrounded by the card's 24-phase VRM solution. This draws power from a single 16-pin 12V-2x6 power connector. NVIDIA will likely max out the 600 W continuous power-delivery capability of the connector, and give the card a TGP of around 500-550 W, if not more.

NVIDIA's GeForce RTX 5090 graphics card printed circuit board has allegedly been shown in the flesh, showing the memory layout and some interesting engineering choices. The custom PCB variant (non-Founders Edition) houses more than 40 capacitors, which is perhaps not standard on the FE reference board, and 16 GDDR7 memory modules. The leaked PCB, which extends beyond standard dimensions and traditional display connector configurations, is reportedly based on NVIDIA's PG145 reference design. The memory modules are distributed systematically: five on the left, two below, five on the right, and four above the GPU die. The interface is PCIe 5.0 x16.

As NVIDIA has reportedly designated 32 GB GDDR7 memory capacity for these cards, this roughly translates into 16 x 2 GB GDDR7 memory modules. At the heart of the card lies what sources claim to be the GB202 GPU, measuring 24×31 mm within a 63×56 mm package. Power delivery uses a 16-pin 12V-6x2 power connector, as expected. The entire PCB features only a single power connector, so the 16-pin 12V-2x6, but with an updated PCIe 6.0 CEM specification, is the logical choice.