NVIDIA will tap into 24 Gbit (3 GB) GDDR7 memory chips and increase memory sizes across the board for its RTX 50-series SUPER line of GPUs slated for later this year, if leaked specs of the upcoming GeForce RTX 5070 SUPER are anything to go by. Kopite7kimi, a reliable source with NVIDIA leaks, says that the RTX 5070 SUPER will feature 18 GB of memory—that's six 24 Gbit memory chips across a 192-bit wide GDDR7 memory bus. The card has a SKU designation of "PG147-SKU65" and ASIC code of "GB205-400-A1."

Besides 18 GB of memory, the RTX 5070 SUPER reportedly maxes out the "GB205" silicon it's based on, enabling all 50 SM present. The current RTX 5070 has 48 out of 50 SM enabled for 6,144 CUDA cores, whereas the RTX 5070 SUPER, with its 50 SM, should have 6,400 CUDA cores. The marginal increase in shader count will be bolstered by the 50% increase in memory size, and possible increases in GPU boost speeds. The memory bandwidth, however, remains unchanged. It ticks at the same 28 Gbps as on the RTX 5070, which yields 672 GB/s. The TGP will be increased to 275 W, from 250 W.

Certain Lenovo Legion pre-built gaming rigs are equipped with intriguing custom graphics card designs; reportedly not available to purchase as (separate) retail packages. As mentioned back in March, the system integrator's Legion 9000K gaming systems were configurable with slick metallic GeForce RTX 5090D and GeForce RTX 5080 options. At the time, Lenovo China hinted about a forthcoming GeForce RTX 5070 offering; also making use of a CNC-machined metal shroud and backplate. Since then, VideoCardz has kept a watchful eye on new product developments—an early April "official" leak suggested an eventual arrival of lower-end GeForce RTX 50-series "Blackwell" Legion models.

As of late last week, Lenovo China's social media accounts have unveiled cheaper Legion 7000K and GeekPro pre-built systems. These new-gen gaming PCs house the brand's fresh GeForce RTX 5070 12 GB (as promised) and GeForce RTX 5060 Ti 8 GB custom cards. VideoCardz reckons that company engineers have adhered to NVIDIA's reference specifications, so potential customers will be considering non-overclocked hardware. The Legion 7000K—starting at 11,999 RMB (~$1658 USD)—traditionally-proportioned enclosure can accommodate a longer card design; as demonstrated by promo shots. The tastefully ARGB-lit Legion GeForce RTX 5070 card's oblong format (2.5-slot) utilizes only two fans. In contrast, Lenovo's compact GeekPro—starter price: 7199 RMB (~$995 USD)—case seems to be formed around a stubbier graphics card volume; still large enough to be dual-fan. Its ridged backplate aesthetic brings previous-gen XFX to mind.

Manli introduced its new generation Polar Fox graphics design earlier this year; accompanied by an "exclusive two-dimensional" product mascot. Prior to 2025, the manufacturer did not market these mid-range offerings with cute character illustrations and graphics. Several Chinese graphics card specialist companies have adopted this aesthetic; seemingly taking major influence from Yeston. According to the Manli's PR material, the latest Polar Fox graphics cards are: "perfect for users who love anime...we've crafted a unique IP identity that makes Polar Fox truly special. The most special feature is wave-shaped heat sink, it enhances airflow to minimize noise, delivering a quieter and more comfortable experience."

The brand launched their icon-themed GeForce RTX 5070 12 GB factory overclocked custom design early on in March, starting off with sections finished with an unusually deep blue tone. As observed by VideoCardz, this aesthetic—covering a significant portion of the card's shroud and a central 9 cm cooling fan—is not a subtle prospect; especially when showcased via a vertical orientation. The brand has prepped a less flashy variant—swapping out the launch model's dark blue tones, in favor of a light-hearted shade of aqua teal. Manli's signature mascot backplate illustration remains the same across both SKUs, although the newer option features additional fractal graphics and "Polar Fox" text. Going back to last month; the manufacturer unveiled this lighter teal setup during an introduction of (mostly dual-fan) new GeForce RTX 5060 Ti models.

At some point this week, the population of ASUS GeForce cards has grown again. In a very quiet manner, the brand has added over twenty new SKUs to its Team Green portfolio. The Taiwanese company seems to be catching up with a nearby AIB rival—since February, MSI has stealthily expanded its graphics card library. It will likely take a while—for PC hardware news outlets—to comb through everything, but momomo_us and Wccftech have put spotlights on two compelling "White Edition" spin-offs of already launched ASUS GeForce RTX 5070 Ti 16 GB and RTX 5070 12 GB custom models. The manufacturer's TUF Gaming division is best known for running with an industrial/military aesthetic on mid-tier positioned products—a standard palette combines gunmetal gray tones, silver highlights and minimal ARGB lighting zones.

The freshly introduced TUF Gaming GeForce RTX 5070 Ti 16 GB GDDR7 White OC Edition will be deployed with the exact same specifications and "military-grade" components present on its traditionally-shaded sibling. In a similar manner, the new PRIME GeForce RTX 5070 White OC Edition 12 GB GDDR7 SKU leverages a repeat of original model specifications. The ASUS global website does not provide pricing or availability details, at the time of writing. Under normal circumstances, PRIME graphics cards were intended to be "baseline MSRP" conformant. It is not clear whether this white-hued + overclocked variant will hover close to NVIDIA's recommended GeForce RTX 5070 starting level: $549.00. Unfortunately, both of these brand-new almost "all-white" setups do not extend to contained PCB designs—we are probably looking at standard model boards. Gaming graphics card enthusiasts have heaped praise on PowerColor; thanks to their inclusion of suitably pale PCBs within "Spectral White" products.

Mere months after the launch of GeForce RTX 5080 16 GB and RTX 5070 12 GB graphics cards, industry rumors are already swirling about a potential "SUPER" semi-successor. Sections of NVIDIA's previous-gen "Ada Lovelace" GeForce RTX 40-series family received mid-generation refresh treatment around early 2024. It is not clear whether it will take a year+ for the speculated arrival of upgraded "Blackwell" desktop gaming solutions, but members of the Chiphell forum have been openly discussing alleged "in-progress" GeForce RTX 5080 SUPER 24 GB and RTX 5070 SUPER 18 GB card designs. Past weekend theorizations were highlighted by ITHome and VideoCardz—in particular, one Chiphell participant posited the two follow-ups will not be deployed with noticeably "faster" performance.

Larger pools of onboard GDDR7 VRAM could pave the way for improvements in AI productivity, although greater capacities could lead to beneficial conditions in gaming scenarios—e.g. extra provisions for large textures and complex assets. The speculated GeForce RTX 5080 SUPER card could borrow aspects from Team Green's GeForce RTX 5090 Mobile 24 GB model; namely the latter's usage of 3 GB GDDR7 memory modules. Press interpretations—of fresh Chiphell leaks—point to company engineers enabling existing PCB designs with 8×3 GB (24 GB) and 6×3 GB (18 GB) parts; thus preventing a major overhaul of board layouts. Global PC gaming hardware communities have often expressed a dislike of NVIDIA's repeated deployment of 8 GB and 12 GB capacity products. A theoretical GeForce RTX 5070 SUPER 18 GB option could receive a warmer welcome. Very early speculation proposes an introduction—of SUPER "Blackwell" models—just before or during CES 2026.

MSI has quietly introduced a refreshed GAMING DUKE product line; starting with GeForce RTX 5070 12 GB graphics card models—in standard and overclocked forms, with an "almost perfect" two-slot thick profile. The manufacturer has a recent track record of adding new designs—to its GeForce RTX 50-series "Blackwell" portfolio—with little press fanfare. We heard about SHADOW 3X options (stealthily) turning back up in February, followed by dual-fan siblings—almost a month later. VideoCardz's past weekend discovery of GeForce 5070 12G GAMING DUKE 3X OC and non-OC (global website) product pages sends MSI's custom GB205 GPU-based offering count up to a grand total of 21 SKUs. Their news article posits that new generation DUKE cards will slide somewhere in-between existing SHADOW and VENTUS ranges—therefore we expect these new models to be relatively inexpensive; perhaps "MSRP conformant."

Going back to last April, MSI unveiled an "Ada Lovelace" generation GeForce RTX 4060 GAMING DUKE 8 GB SKU—as a Chinese market exclusive. The brand's latest marketing blurb mentions "speed and strength" being the cornerstones of this year's all-black refreshed design. Elaborating further: "GAMING DUKE channels the power of the dragon, ensuring smooth and stable performance in every battle. Its shadowy, armored design, reinforced with a metal backplate, offers solid protection, while glowing accents hint at the energy within." Graphics card industry watchdogs anticipate the release of more GAMING DUKE products—smaller GeForce RTX 5060 Ti and RTX 5060 designs could be next.

ASUS China showcased a compelling new custom graphics card design during their presentation of a Hatsune Miku-themed product collection. Global audiences were treated to TUF Gaming x Hatsune Miku crossover peripherals/accessories, but Chinese gamers will get access to a wider gamut of options—under the TX Gaming banner; a regional spin-off of TUF Gaming—including a cutely decorated GeForce RTX 5060 Ti 8 GB model. ASUS seems to be reusing this card's basic shroud/backplate design—minus elaborate character illustrations and graphics—on a brand-new TX Gaming GeForce RTX 5070 12 GB SKU. A "Born to be Different" message adorns a largely plain white front section, and familiar "GeForce RTX" texts are placed on a side panel and on the 2.5-slot thick card's metallic silver backplate. A large X-formation cut-out grants a glimpse at an enclosed fin stack.

In the past, ASUS TX Gaming products have been accompanied by sci-fi "idol" characters—on related retail box art and in promotional material. Evidently, ASUS has extended this "marketing technique" into the NVIDIA "Blackwell" GPU generation. As expected, the TX Gaming GeForce RTX 5070 models—in standard and factory overclocked forms—will launch as Chinese market exclusives. VideoCardz did not find any global listings of these offerings, nor launch details—e.g. a release date and MSRPs. They noted that a TUF Gaming equivalent model features five display connectors; as opposed to the TX Gaming's deployment of four (1x HDMI 2.1b, 3x DisplayPort 2.1b).

Earlier today, ZOTAC unveiled a special edition spin-off of its relatively new GeForce RTX 5070 X-Gaming 12 GB design. The manufacturer's Weibo channel outlined the very exclusive nature of this MMORPG-themed offering (via machine translation): "watch the finals and get a graphics card! ZOTAC and the first 'Jian Wang 3' competitive group hero joint customized graphics card, fine customization, supports NVIDIA DLSS 4 and Reflex 2 technology, excellent performance. On April 12-13, Nanjing Ledongli Jiangning International E-sports Center, watch the heroes aiming for the top!" According to an ITHome report, the Chinese brand has redecorated its overclocked GeForce RTX 5070 X-Gaming model with Jian Wang 3 characters. ZOTAC's promotional shots showcase a customized backplate, and the same illustration of Jian Wang 3's main cast adorning the card's retail box.

Late last month, Kuroutoshikou Japan introduced a less impressive Blade and Soul NEO-themed rebadge of their standard Radeon RX 7600 SKU—essentially, PowerColor's Fighter RX 7600 design. ITHome noted that ZOTAC has a short-ish history of readying very limited edition designs for past Jian Wang 3 Masters Tournaments. 2024's prize was based on the manufacturer's GeForce RTX 4070 Ti SUPER TRINITY OC White SKU. 2023 finalists were gifted with customized GeForce RTX 4070 Ti X-Gaming OC cards. Both of the previous gen tie-in designs sported Jian Wang 3 graphics or symbols on their shrouds, while the latest entry simply reuses its base SKU's "youthful graffiti" aesthetic.

EMTEK has released a new custom GeForce RTX 5070 graphics card in South Korea; fresh retail/e-tail listings have popped up online via the Danawa price comparison engine. Similar circumstances were observed around mid-February for the launch of the brand's GeForce RTX 5080 MIRACLE WHITE D7 16 GB SKU. EMTEK's GB205 "Blackwell" GPU-based offering sports a slightly smaller shroud design; its larger siblings are 2.5-slotters. As noted by VideoCardz, the GeForce RTX 5070 MIRACLE WHITE D7 12 GB model's 329 mm-long triple-fan cooling solution tempers a less potent key component.

EMTEK's brand-new card conforms to NVIDIA's reference specifications, so a relatively slim heatsink seems appropriate for this deployment. A dual BIOS switcher grants access to "Cooling" and "Silent 0-db" modes. Another nearby physical switch can enable/disable the MIRACLE WHITE D7's integrated "Auto ARGB" system. EMTEK's pricier pale-toned offerings—in GeForce RTX 5080 and RTX 5070 Ti guises—feature intriguing USB-C connected Windows 11-controlled lighting schemes. The cheapest price for a RTX 5070 MIRACLE WHITE D7 card is 1,030,000 won (~$700 USD) according to Danawa SK aggregation. EMTEK products are only available in South Korea, therefore attract very little Western press coverage. Interestingly, the company also acts as a regional distributor of various PALIT GeForce and Sapphire Radeon graphics cards.

To Western audiences, ZOTAC's X-Gaming graphics card product line is not a well known quantity—a search of TechPowerUp's news archive produces one result (from 2021). As reported by VideoCardz, X-Gaming models are only available to buy in China. Roughly two months ago, the Hong Kong-headquartered manufacturer teased four new designs based on NVIDIA's emerging GeForce RTX 50-series "Blackwell" GPUs. As mentioned during our very recent coverage of ZOTAC's GeForce RTX 5080 Apocalypse OC model, the brand introduced familiar SOLID and AMP Extreme INFINITY families. A 2025 refresh of the company's "youthful" X-Gaming series was advertised as a subversion of: "the coldness of traditional graphics card industrial design through high-saturation contrasting color splicing, modular geometric segmentation and other techniques."

Since its beginnings, X-Gaming's core design concept is described as "breaking the dimensional wall between hardware and fashion." ZOTAC's local branch has just released brand-new X-Gaming GeForce RTX 5070 Ti and RTX 5070 SKUs, in factory-overclocked and standard/reference configurations. VideoCardz believes that the GeForce RTX 5080 X-Gaming OC model launched—in China—earlier in the year. So far, all of the launched variants share the same visual theme—ZOTAC cites "inspiration" from a famous source: "the GeForce RTX 50 X-Gaming series graphics cards have a white base and large areas of contrasting colors. The front pattern design is inspired by the well-known artist Mr. Doodle (aka Sam Cox), whose iconic intensive linearism style is accurately restored on the X-Gaming: the front of the graphics card is covered with hundreds of lines, and the seemingly disordered graffiti combination creates a vibrant visual effect. Through a special printing process, the front pattern presents a unique visual effect under different light angles, further increasing the texture and layering of the card's appearance." Regional ZOTAC enthusiasts expected the GeForce RTX 5070 Ti option to borrow its "larger" sibling's robust cooling solution, but the GB203 GPU-based model makes do with a thinner shroud profile—also present on the GeForce RTX 5070 (non-Ti) X-Gaming SKU.

So far, AX Gaming has mainly released triple-fan cooled custom GeForce RTX 50-series graphics cards under their "Rebel" product line. For example, TechPowerUp's GPU database lists a GeForce RTX 5070 X3W SKU—promo images showcase fairly straightforward white shroud and backplate designs; somewhat belying the family's namesake. Earlier today, VideoCardz put a spotlight on an extravagantly-decorated sibling—the X2W. TechPowerUp's news archive seems to lack in AX Gaming content, but a forumite correctly informed others about this sub-brand existing under the umbrella of Inno3D. The Singaporean AIB launched its AX Gaming off-shoot a few years ago; with its main market focus being China.

The freshly revealed AX Gaming Rebel GeForce RTX 5070 X2W 12 GB model sports a very busy shroud design; thus effortlessly distinguishing itself from nearby plain triple-fan products. Mayan civilization-inspired graffiti graphics are spreads across the card's front face. This funky aesthetic is also present on the model's retail packaging. Interestingly, the two cooling fans are adorned with a hand graphic; clutching a modern gamepad. Disappointingly, the manufacturer has not applied this loud patterning onto the X2W's backplate—we are looking at a mostly white landscape adorned with some sort of Rebel series emblem and "GeForce RTX" text. AX Gaming's specification sheet indicates that their X2W SKU sticks to NVIDIA's reference specifications. The manufacturer's "Punk 4.0" cooling system consists of two fans, paired with a heatsink that features five heatpipes. VideoCardz reckons that the housed PCB also exists within Inno3D's RTX 5070 TWIN X2 cards.

A week ago, MSI introduced its lineup of custom GeForce RTX 5070 graphics card models—headlined by their premium VANGUARD option, with a product stack going down to a budget friendly INSPIRE 3X offering. Since then, the Taiwanese manufacturer has added another entry to its NVIDIA "Blackwell" GB205 GPU-based family. As observed by VideoCardz, MSI is overpopulating its GeForce RTX 5070 stack with two more models. The SHADOW 2X OC and SHADOW 2X (non-OC) have—very recently—popped up on the company's website, with absolutely zero PR fanfare. The triple-fan SHADOW 3X designs received a similar treatment last month—brand-new stealth-black GeForce RTX 5080 and RTX 5070 Ti custom designs seemingly appeared overnight.

As befits the nomenclature, MSI's SHADOW 2X design sports a TORX 5.0 dual-fan cooling solution. As covered by TechPowerUp's news section in the past, the SHADOW series shares similarities with VENTUS—both families serve as "baseline MSRP" conformant products, due to a minimalistic aesthetic and barebones feature set (i.e. focused on the essentials). MSI marketing blurb describes SHADOW as: "a performance-focused design that delivers the gaming experience players want, making it the ideal choice when upgrading or building a gaming rig." The company has courted controversy in recent times; VideoCardz and other media outlets have leveled plenty of criticism throughout February and March—the lack of "MSRP" models and reported price hiking at launch became major sticking points. Industry watchdogs believe that the freshly unveiled SHADOW 2X models will be in short supply for the foreseeable future.

Earlier today, GALAX unveiled a brand-new custom GeForce RTX 5070 graphics card—according to ITHome, this offering marks the debut of the manufacturer's FIRE product series. This nomenclature has raised eyebrows across Western press outlets—given recent "high temperature" events involving NVIDIA "Blackwell" GPUs; GALAX's latest introduction seems to be a bit clumsy (timing-wise). The freshly introduced triple-fan model sports a "Cyber Black Flame" design; promotional imagery showcases an almost all-black aesthetic. The shroud features carbon fiber texturing and pattern-embossed sections. Its backplate displays prominent white GALAX branding and GeForce RTX text. The FIRE's cooling system is referred to as "wandering star."

According to its specification sheet, the GALAX GeForce RTX 5070 FIRE 12 GB model sticks closely to Team Green reference points. Currently, TechPowerUp's GPU database lists four GB205 GPU-based GALAX SKUs—in 1-Click or EX Gamer forms—all factory overclocked affairs. ITHome's report does not mention any price point for the debut FIRE card, but expectations point to a possible baseline MSRP conformant release in China. GALAX has prepped the GeForce RTX 5070 FIRE with a very barebones dual-slot design—akin to MSI's relatively new SHADOW 3X series. The GeForce RTX 5070 SHADOW 3X model also adheres to NVIDIA reference specs, including a TDP rating of 250 W.

ASUS has taken the unusual step of formally announcing day one availability of its brand-new offerings—in TUF Gaming and PRIME forms—to the NVIDIA subreddit community. Today (March 5) marks the retail rollout of AIB-built GeForce RTX 5070 graphics cards—unfortunately, NVIDIA Founders Edition enthusiasts will have to wait a little longer for a now delayed launch. VideoCardz has called out ASUS on a number of occasions over the past month and a half, regarding alleged GeForce RTX 50-series "price manipulation" practices. In their latest bit of company coverage, the publication concedes that ASUS representatives are doing a better job in communicating launch day circumstances to its target audience. In contrast, other board partners have (reportedly) not reached out to the hardware gaming community.

The ASUS_MKTLeeM account has produced a useful "Day 1 Pricing and Availability in US and CA" buying guide for TUF Gaming GeForce RTX 5070 OC Edition ($739.99 USD) and ASUS Prime GeForce RTX 5070 models (standard: $549.99, OC: $699.99). The company rep disclosed the outlook for this opening salvo: "we expect that most, if not all of these will be available at 6am PST for the US. Of course, things happen sometimes; and a particular e-tailer or retailer may have a website issue or may be pending for stock to reach them. Feel free to ask later in the day and we can track this down...For Canada, I do not have confirmation on which stores will have Day 1 stock at this time, unfortunately. If I receive confirmation before the end of the day I will update it." They noted that Best Buy Canada may experience delayed shipments (1-3 weeks) of TUF Gaming RTX 5070 OC Edition stock. Unfortunately, higher-end options are not releasing alongside "cheaper" brethren parts—MKTLeeM stated: "the ROG Strix cards will not be available Day 1. There is currently no ETA for when these cards will be available, but we will update the post and/or the availability post in the comments when we have an update." Press outlets have highlighted the apparent limited range of ASUS launch day models. Interestingly, the TUF Gaming RTX 5070 (non-OC) option was not included in MKTLeeM's subreddit thread.

NVIDIA today released to market the GeForce RTX 5070 performance segment graphics card for maxed out gaming at 1440p with ray tracing. Based on the "Blackwell" graphics architecture, the RTX 5070 gets a few of the latest features, including Neural Rendering, DLSS 4 with Multi Frame Generation, and an updated ray tracing hardware that's ready for Mega Geometry. The RTX 5070 debuts the new 5 nm "GB205" silicon, a mid-sized GPU featuring a well-rounded 50 SM or 6,400 CUDA cores. The RTX 5070 almost maxes this out, with 48 SM or 6,144 CUDA cores on tap.

Other key specs of the RTX 5070 include 192 Tensor cores, 48 RT cores, 192 TMUs, and 80 ROPs. The SKU also gets all 48 MB of L2 cache present on the silicon. When compared to the previous generation RTX 4070, these counts see a numerical increase. The RTX 4070 has 64 ROPs and 36 MB of cache. While the memory size is unchanged at 12 GB, the memory bandwidth sees an increase by 33% thanks to the new 28 Gbps GDDR7 memory being used. The RTX 5070 has a starting price of $550, but real-world pricing could easily reach $700, if not more.

Inet AB—one of the largest Swedish e-tail stores—has alerted its customer base with foreboding information. NVIDIA and its board partners are expected to launch GeForce RTX 5070 graphics cards tomorrow (March 5)—reviews have started to trickle out; including TechPowerUp's freshly-published evaluation of the Founders Edition. According to the Scandinavian shop's blog, they anticipate major supply problems: "unfortunately it looks like we won't be able to release the cards then. This is simply because we don't seem to be getting any cards in stock. Just like with other models in the RTX 50 series, we will release new cards one by one as they land in stock, and only make a fuss in cases where we might receive large deliveries." A week and a half ago, supply chain insiders alluded to significant problems that have reportedly affected GeForce RTX 5070 and RTX 5060 production pipelines.

Like many other global PC hardware retail outlets involved in the sale of "Blackwell" GPUs, Inet has offered helpful guidelines in anticipation of limited shipments: "we hope to start releasing cards and (pre-built) computers with GeForce RTX 5070 in the coming weeks. If you want to buy RTX 5070, we recommend that you keep an eye on inet.se and use 'Notify me' on each product page, then you will receive an email when we release them for order. Keep in mind, however, that many people use this function, so the cards may run out quickly after the email goes out." VideoCardz and Australia's Hardware Unboxed have both predicted a grand battle between GeForce RTX 5070 and Radeon RX 9070 XT cards. Team Green tends to stick to its intended release schedules—according to industry watchdogs—even in the face of mounting problems. The tides could be turning again; recent events have allegedly pushed the global launch of GeForce RTX 5060 cards into April.

NVIDIA is preparing a late-March 2025 launch of the GeForce RTX 5060 Ti following its early-March RTX 5070 launch. The card will reportedly come with 16 GB and 8 GB memory variants, much like its predecessor, the RTX 4060 Ti, with the 16 GB variant launching late-March, and the 8 GB variant in April. A Wccftech report sheds light on a few interesting specs of the RTX 5060 Ti. Apparently, the card comes with a total graphics power (TGP) of 180 W, which should make it possible for partners to sell cards with single 8-pin PCIe power connectors. NVIDIA is looking to target the $400 to $500 price segment with the RTX 5060 Ti series, but a lot will depend on how AMD addresses this space with its lean new "Navi 48" silicon that could probably be cut down to create SKUs that square off against the RTX 5060 Ti.

The GeForce RTX 5060 Ti, much like the RTX 4060 Ti, will come with a 128-bit wide memory bus, however, NVIDIA will give the GPU a massive upgrade in bandwidth with 28 Gbps GDDR7, giving it 448 GB/s of memory bandwidth, which is a 55% increase over the 288 GB/s that the RTX 4060 Ti has. The GB205 (or GB206?) GPU that the RTX 5060 Ti is based on could implement PCI-Express 5.0, which means that should NVIDIA opt for an 8-lane PCIe host interface to save board costs, the GPU could still enjoy bandwidth comparable to PCI-Express 4.0 x16 on systems with Gen 5 PCIe.

NVIDIA's upcoming mid-range GeForce RTX 5070 12 GB model is almost ready for launch, according to recent reports. Industry moles reckon that GB205 GPU-based specimens are already in the clutches of press and influencer outlets; review embargoes are due to be lifted on March 4, for $549 MSRP conformant SKUs (as disclosed by a VideoCardz source). Last week, we heard whispers about Team Green's (allegedly) troubled production cycle for incoming GeForce RTX 5070 and RTX 5060 models.

Insiders insist that these issues have caused a delay; many believed that NVIDIA had (prior) plans for a February GeForce RTX 5070 launch. A revised schedule was leaked to VideoCardz; the publication posits that GeForce RTX 5070 cards will launch at retail on March 5, with non-MSRP ($549+) reviews projected to go live on the same day. Based on various leaks, NVIDIA and AMD will likely clash with their respective new offerings. Right now, reviewers could be dealing with sizable piles of competing Team Green and Team Red hardware. Graphics card enthusiasts will be looking forward to incoming comparisons—GeForce RTX 5070 and its Ti sibling versus Radeon RX 9070 XT and RX 9070 (non-XT).

NVIDIA's GeForce RTX 5070 Ti today goes on sale, and we have reviewed a few of them. However, the RTX 5070 non-Ti variant has reportedly faced some production issues. According to CTEE reports, volume production was pushed back by one month, with manufacturing expected to reach full capacity by mid-March. Given that RTX 5070 is officially coming on March 5, we are left to wonder if enough capacity will be available for the launch day or if it will follow the same footsteps of scarcity that current RTX 5090 and RTX 5080 have experienced.

The unannounced RTX 5060's mass production has been pushed to mid-April, with both models requiring additional debugging due to unexpected issues. NVIDIA's engineers are ironing out all bugs to ensure stable GPU and drivers arrive on time. With the RTX 5070 using the GB206 GPU, the RTX 5060 is expected to implement a GB206 variant, with the in-development RTX 5060 Ti featuring a slightly larger GB205 GPU. For memory, RTX 5070 is expected to utilize 12 GB of GDDR7, and RTX 5060 should come with an 8 GB GDDR7 configuration.

Thanks to kopite7kimi, we are able to finalize the leaked specifications of NVIDIA's upcoming GeForce RTX 5070 and RTX 5070 Ti graphics cards.
Starting off with RTX 5070 Ti, it will feature 8,960 CUDA cores and come equipped with 16 GB GDDR7 memory on a 256-bit memory bus, offering 896 GB/s bandwidth. The card is reportedly designed with a total board power (TBP) of 300 W. The Ti variant appears to use the PG147-SKU60 board design with a GB203-300-A1 GPU. The standard RTX 5070 is positioned as a more power-efficient option, with specifications pointing to 6,144 CUDA cores and 12 GB of GDDR7 memory on a 192-bit bus, with 627 GB/s memory bandwidth. This model is expected to operate at a slightly lower 250 W TBP.

Interestingly, the non-Ti RTX 5070 card will be available in two board variants, PG146 and PG147, both utilizing the GB205-300-A1 GPU. While we don't know what the pricing structure looks like, we see that NVIDIA has chosen to make more considerable differentiating factors between its SKUs. The Ti variant not only gets an extra four GB of GDDR7 memory, but it also gets a whopping 45% increase in CUDA core count, going from 6,144 to 8,960 cores. While we wait for the CES to see the initial wave of GeForce RTX 50 series cards, the GeForce RTX 5070 and RTX 5070 Ti are expected to arrive later, possibly after RTX 5080 and RTX 5090 GPUs.

Possible specifications of the various NVIDIA GeForce "Blackwell" gaming GPUs were leaked to the web by Kopite7kimi, a reliable source with NVIDIA leaks. These are specs of the maxed out silicon, NVIDIA will carve out several GeForce RTX 50-series SKUs based on these chips, which could end up with lower shader counts than those shown here. We've known from older reports that there will be five chips in all, the GB202 being the largest, followed by the GB203, the GB205, the GB206, and the GB207. There is a notable absence of a successor to the AD104, GA104, and TU104, because NVIDIA is trying a slightly different way to approach the performance segment with this generation.

The GB202 is the halo segment chip that will drive the possible RTX 5090 (RTX 4090 successor). This chip is endowed with 192 streaming multiprocessors (SM), or 96 texture processing clusters (TPCs). These 96 TPCs are spread across 12 graphics processing clusters (GPCs), which each have 8 of them. Assuming that "Blackwell" has the same 256 CUDA cores per TPC that the past several generations of NVIDIA gaming GPUs have had, we end up with a total CUDA core count of 24,576. Another interesting aspect about this mega-chip is memory. The GPU implements the next-generation GDDR7 memory, and uses a mammoth 512-bit memory bus. Assuming the 28 Gbps memory speed that was being rumored for NVIDIA's "Blackwell" generation, this chip has 1,792 GB/s of memory bandwidth on tap!

NVIDIA is confirmed to implement the GDDR7 memory standard with the top three GPU ASICs powering the next-generation "Blackwell" GeForce RTX 50-series, Tweaktown reports, citing XpeaGPU. By this, we mean the top three physical silicon types from which NVIDIA will carve out the majority of its SKUs. This would include the GB202, the GB203, and GB205; which will power successors to everything from the current RTX 4070 to the RTX 4090. NVIDIA is expected to build these chips on the TSMC 4N foundry node.

There will be certain GPU ASIC types in the "Blackwell" generation that will stick to older memory standards such as GDDR6 or even the GDDR6X. These would be successors to the current AD106 and AD107 ASICs, powering SKUs such as the RTX 4060 Ti, and below. NVIDIA co-developed the GDDR6X standard with Micron Technology, which is the chip's exclusive supplier to NVIDIA. GDDR6X scales up to 23 Gbps and 16 Gbit, which means NVIDIA can avail plenty of performance for the lower-end of its product stack using GDDR6X; especially considering that its GDDR7 implementation will only run at 28 Gbps, despite chips being available in the market for 32 Gbps, or even 36 Gbps. Even if NVIDIA chooses the regular GDDR6 standard for its entry-mainstream chips, the tech scales up to 20 Gbps.

Following Monday's blockbuster announcements of the "Blackwell" architecture and NVIDIA's B100, B200, and GB200 AI GPUs, all eyes are now on its client graphics derivatives, or the GeForce RTX GPUs that implement "Blackwell" as a graphics architecture. Leading the effort will be the new GB202 ASIC, a successor to the AD102 powering the current RTX 4090. This will be NVIDIA's biggest GPU with raster graphics and ray tracing capabilities. The GB202 is rumored to be followed by the GB203 in the premium segment, the GB205 a notch lower, and the GB206 further down the stack. Kopite7kimi, a reliable source with NVIDIA leaks, says that the GB202 silicon will be built on the same TSMC 4N foundry node as the GB100.

TSMC 4N is a derivative of the company's mainline N4P node, the "N" in 4N stands for NVIDIA. This is a nodelet that TSMC designed with optimization for NVIDIA SoCs. TSMC still considers the 4N as a derivative of the 5 nm EUV node. There is very little public information on the power- and transistor density improvements of the TSMC 4N over TSMC N5. For reference, the N4P, which TSMC regards as a 5 nm derivative, offers a 6% transistor-density improvement, and a 22% power efficiency improvement. In related news, Kopite7kimi says that with "Blackwell," NVIDIA is focusing on enlarging the L1 caches of the streaming multiprocessors (SM), which suggests a design focus on increasing the performance at an SM-level.

Speculative NVIDIA GeForce RTX 50-series "GB20X" GPU memory interface details appeared online late last week—as disclosed by the kopite7kimi social media account. The inside information aficionado—at the time—posited that the "memory interface configuration of GB20x (Blackwell) is not much different from that of AD10x (Ada Lovelace)." It was inferred that Team Green's next flagship gaming GPU (GB202) could debut with a 384-bit memory bus—kopite7kimi had "fantasized" about a potentially monstrous 512-bit spec for the "GeForce RTX 5090." A new batch of follow-up tweets—from earlier today—rips apart last week's insights. The alleged Blackwell GPU gaming lineup includes the following SKUs: GB202, GB203, GB205, GB206, GB207.

Kopite7kimi's revised thoughts point to Team Green's flagship model possessing 192 streaming multiprocessors and a 512-bit memory bus. VideoCardz decided to interact with the reliable tipster—their queries were answered promptly: "According to kopite7kimi, there's a possibility that the second-in-line GPU, named GB203, could sport half of that core count. Now the new information is that GB203 might stick to 256-bit memory bus, which would make it half of GB202 in its entirety. What this also means is that there would be no GB20x GPU with 384-bit bus." Additional speculation has NVIDIA selecting a 192-bit bus for the GB205 SKU (AKA GeForce RTX 5070). The GeForce RTX 50-series is expected to arrive later this year—industry experts are already whispering about HPC-oriented Blackwell GPUs being unveiled at next week's GTC 2024 event. A formal gaming family announcement could arrive many months later.

The next-generation GeForce RTX 50-series graphics cards will be powered by the Blackwell graphics architecture, named after American mathematician David Blackwell. kopite7kimi, a reliable source with NVIDIA leaks revealed what the lineup of GPUs behind the series could look like. It reportedly will be led by the GB202, followed by the GB203, and then the GB205 and GB206, followed by the GB207 at the entry level. What's surprising here, is the lack of a "GB204" succeeding the AD104, GA104, TU104, and a long line of successful performance-segment GPUs by NVIDIA.

The GeForce Blackwell ASIC series begins with "GB" (GeForce Blackwell) followed by a 200-series number. The last time NVIDIA used a 200-series ASIC number for GeForce GPUs was with "Maxwell," as the GPUs ended up being built on a more advanced node, and with a few more advanced features, than what the architecture was originally conceived for. For "Blackwell," the GB202 logically succeeds the AD102, GA102, TU102, and a long line of "big chips" that have powered the company's flagship client graphics cards. The GB103 succeeds AD103, as a high SIMD count GPU with a narrower memory bus than the GB202, powering the #2 and #3 SKUs in the series. There is curiously the lack of a "GB104."