During the Intel AI Summit in Seoul, South Korea, Intel teased its upcoming product portfolio, featuring next-generation memory technologies. Being in Seoul, memory makers like SK Hynix are Intel's main partners for these products. Teased at the summit is Intel's upcoming AI accelerator, called "Jaguar Shores," which utilizes the next-generation HBM4 memory, offering 2.0 TB/s of bandwidth per module across 2,048 IO pins. SK Hynix plans to support this accelerator with its memory, ensuring that Intel's big data center-grade AI accelerator is equipped with the fastest memory on the market. Since the "Falcon Shores" accelerator is only intended for testing with external customers, we don't have an exact baseline to compare to, and Jaguar Shores specifications are scarce.

Next up, Intel confirmed that its upcoming seventh-generation "Diamond Rapids" Xeon processors will use the second generation of MRDIMMs (Multiplexer Rank Dual Inline Memory Modules), an upgrade from the first-generation MRDIMMs used in the Xeon 6 family. The upgrade to MRDIMMs Gen 2 will allow Intel to push transfer rates to 12,800 MT/s, up from 8,800 MT/s in Xeon 6 with MRDIMMs Gen 1. Alongside this 45% speed bump in raw transfer rates, the memory channels are jumping to 16, up from 12 in the current generation, yielding an additional bandwidth boost. Given that MRDIMMs operate by connecting more memory ranks using a multiplexer, and that these modules buffer data and commands, the increased data transfer rate comes without any additional signal degradation. As Intel is expected to pack more cores, this will be an essential piece in the toolbox to feed them and keep those cores busy on the Oak Stream platform, based on the LGA9324 socket.

EMTEK, a South Korean manufacturer, has quietly added two new SKUs to its current-gen MIRACLE WHITE D7 graphics card series. As of late last week, their custom GeForce RTX 5060 Ti 16 GB and 8 GB models have turned up for sale at domestic e-tail outlets. The higher capacity option starts off at 725,900 KRW (~$535 USD), while its slightly cheaper sibling demands at least 639,660 KRW (~$472 USD). As covered on TechPowerUp in the recent past, EMTEK products are not well known properties in the West—due to regional retail exclusivity.

Unsurprisingly, the latest MIRACLE WHITE cards carry the same overall visual theme as slightly older offerings. Unlike the slightly larger GeForce RTX 5070 D7 12 GB SKU, EMTEK has readied a slimmer 2-slot thick cooling solution (versus 2.5) with no integrated lighting zones. Promotional renders show a triple-fan array and relatively substantial heatsink tempering a (not white) stubby custom PCB. Gigabyte pulled off something similar on a GB206 "Blackwell" GPU-based EAGLE model, albeit with only one fan positioned over crucial hardware. As per usual (for this series), the GeForce RTX 5060 Ti MIRACLE WHITE D7 models utilize NVIDIA's reference specifications.

Often, when semiconductor companies talk about chip manufacturing, they are referring to the specific process of applying masks and etching silicon wafers that takes place in advanced foundries. While these are critical steps in chip production, they don't result in a functional processor in and of themselves. Fully processed wafers are often transferred from the foundry manufacturer to an OSAT (Outsourced Semiconductor Assembly and Test) firm. There, newly built processors are mounted to underlying substrates, connected to other processors, and checked for errors.

ASE Technology Holdings is the largest OSAT in the world, with facilities in Taiwan, China, Japan, Korea, Malaysia, and Singapore. It assembles and tests electronics for semiconductor manufacturers across the planet and has collaborated with AMD to develop 2.5D interposers and other advanced packaging since 2007. But it's one thing to build parts for a customer and something altogether different to adopt that customer's hardware into your own lines of business. ASE's collective requirements run the gamut from lightweight mobile systems to data center systems with dozens to hundreds of available CPU cores.

Yesterday, ASRock's South Korean office issued a statement regarding problematic hardware pairings; affecting current-gen motherboards and AMD Ryzen 9000 "Granite Ridge" desktop processors. Curiously, the Taiwanese manufacturer's local branch slipped in an intriguing tidbit into its message, addressed to members of the Quasar Zone forum: "ASRock has released a new BIOS version 3.25. This version is based on AMD AGESA 1.2.0.3d, and has been improved to enhance compatibility with next-generation CPUs and optimize PBO (Precision Boost Overdrive) settings to enable more stable system operation." Due to media outlets focusing on an important BIOS update, a major hint was largely ignored. Fortunately, VideoCardz's keen eye was not distracted by current affairs.

According to their evening report (dated June 4), their local contacts have confirmed that there are no misunderstandings when Westerners machine translate ASRock Korea's original statement. Naturally, guess work—about the company's teasing of next-gen processors—will land on two upcoming Team Red product families. VideoCardz quickly eliminated a successive "Zen 6" CPU series, aka "Ryzen 10000." Despite recent leaks, industry watchdogs anticipate (at best) a mid-2026 debut. Given the uptick of semi-official Ryzen 9000G "Gorgon Point" leaks throughout Q2'25, VideoCardz reckons that AMD and mainboard partners are preparing for the arrival of this Zen 5 + RDNA 3.5-based APU series. Industry insiders have proposed a fourth quarter launch window.

At Computex 2025, ASRock deliberated over the problems that have plagued certain AMD 800-series motherboard models. Fatal combinations—usually affecting popular Ryzen 7 9800X3D gaming CPUs—have been highlighted by sections of the PC hardware community, and (subsequently) press outlets. Months ago, the Taiwanese mainboard manufacturer's Japanese branch took a combative stance (in response to widespread criticism). Following continued failures of Team Red "Granite Ridge" desktop processors on "mid-to-high-end" ASRock mainboards—even in non-3D V-Cache guises—company representatives have interacted with key media conduits, including Tech YES City's Bryan Bilowol. Coinciding with the release of fresh BIOS fixes, long-form video coverage of Q&A sessions have turned up on YouTube.

Company reps have alluded to a warranty replacement scheme, but full details were not disclosed during May-time Computex proceedings. Many critics are still waiting for an official company statement; hopefully addressed to a global audience. In the meantime, one regional office has weighed in with some much-needed outreach: "ASRock Korea takes the recent AMD 800 series motherboard issues experienced by some users very seriously and will follow up with you to protect your trust. ASRock is continuously monitoring the performance and stability of the platform and will provide a firmware update as soon as possible if any issues are identified. In addition, if the system you are using is damaged due to this issue, we will take responsible measures for both the CPU and the motherboard distributed through the official domestic importer, and please submit the case through ASRock Korea's official distributor for diagnosis and processing. ASRock Korea will continue to listen to the voices of our customers and do our best to remain a trusted brand." This message was posted on Quasar Zone; a high-traffic PC hardware forum. A good number of manufacturers have responded to cases raised by members of this South Korean community.

The Nintendo Switch 2 hybrid console is due to launch globally next Wednesday (June 4). The highly anticipated next-gen handheld is powered by a custom NVIDIA processor. To the surprise of many industry watchdogs, both parties have semi-recently disclosed a couple of technical details regarding their fruitful hardware collaboration. Historically, Nintendo has guarded many aspects of its past generation hardware. Throughout the 2020s, data miners and leakers have unearthed plenty of pre-release information—leading to theories about the Switch 2 chipset's origins. During the Switch 1 era, TSMC was the chosen manufacturing partner. NVIDIA's off-the-shelf Tegra X1 mobile SoC powered the first wave of Nintendo Switch (2017) devices, in 20 nm form. A 2019 revision resulted in Switch Lite and (refreshed) Switch models being equipped with a more efficient 16 nm solution, also present within 2023's premium OLED variant.

Since then, Switch 2's alleged NVIDIA Tegra T239 SoC was linked to a Samsung 8 nm node process. Earlier this month, extremely brave Chinese leakers produced "full die shot" evidence of South Korean foundry origins. Bloomberg insider news articles have implied that Samsung Semi's mature 8 nm FinFET node is better suited—rather than an equivalent TSMC product—for the Switch 2's custom NVIDIA chipset. Unnamed sources have mentioned critical factors; namely stable production and process compatibility. Industry moles reckon that Samsung leadership is actively and aggressively pushing for a longer Switch 2 chipset production deal. Renewed terms could include a future die shrink; pre-launch analysis indicates a sizeable 207 mm² footprint. Beyond foundry biz negotiations, additional murmurs suggest company executives dangling an OLED panel supply agreement. Industry experts have viewed Samsung's key entry—into the gaming console chip market—as a seismic development. A DigiTimes article dives into a so-called "tripartite cooperation"—involving Nintendo, NVIDIA, and the South Korean semiconductor giant. The Samsung Foundry has floundered and struggled in recent times, but is keen to catch up with its arch rival. Fresh rumors have AMD and Sony considering Samsung's chip making channels; possibly with futuristic PlayStation hardware in mind.

Zalman has debuted its ZM-DF120 fan series at Computex 2025; with an emphasis on reducing cable clutter and easy installation. This is enabled by a magnetic mounting system and per-unit pogo pin interfaces. Similar designs have emerged from rival companies in the recent-ish past; eventually leading to ugly legal battles. The South Korean manufacturer is (bravely) stepping into this arena with its heavily ARGB-lit ZM-DF120 120 mm PWM fan options; showcased in black and white finishes. Side infinity mirror panels are hyped as providing dazzling visual effects. According to specification placards, fan speed range is 600-2000RPM (±10%). Maximum noise level is 34.5 dB (±10%), and max. air flow is 70.05 CFM (±10%). Zalman did not list any static pressure data. As expected, their ZM-DF120 fans leverage a hydro bearing platform.

Zalman has showcased their ZET series at Computex 2025; a brand-new family of air CPU tower coolers. This lineup seems to serve as a natural successor to past shell-less CNPS offerings. Circular cooling solutions are an unusual sight in the 2020s, but the South Korean manufacturer is intent on breaking beyond current (boxy) trends. Company representatives and product placards mentioned a main shape inspiration: aircraft engine designs.

The flagship ZET5 model was presented in black and white forms; both augmented with tasteful ARGB lighting rings. The brand's premium option has a 200 W TDP rating. Two modular fans are attached via a magnetic system, and connectivity is granted by a pogo pin system. The ZET5 features copper heatpipes, and a surprisingly non-radial aluminium fin array. A tube of ZM-STC10 thermal compound will be bundled in with these top-end products. Tentative pricing was not available at the time of writing.

On June 11, the PC version of Stellar Blade will be released. The PC version of Stellar Blade will support several new features and introduce exciting additional content. This additional content will also be available on the PS5 console version, This additional content will also be available on the PS5 console version. PS5 users will receive the free update content through a patch download. It will deliver the best experience to everyone - both users who couldn't join Eve's mission due to platform limitations, and fans who have already witnessed the end of Eve's protocol. With this in mind, we are pleased to share detailed information about the Stellar Blade PC version.

NVIDIA DLSS 4 & AMD FSR 3 upscaling and unlocked frame rate
Blade, blood, and beauty. Eve is a ruthless executioner and an elegant liberator. The PC version of Stellar Blade supports NVIDIA DLSS 4 resolution scaling and frame generation, NVIDIA DLAA image enhancement, and NVIDIA Reflex latency reduction features. AMD FSR 3 is also supported. Additionally, it supports unlocking the frame rate, allowing you to maximize your PC's potential. With these enhancements, Eve's action will transcend its limits, becoming even more merciless and all the more mesmerizing.

Around mid-February, South Korean sources alleged that NVIDIA was in the process of developing an innovative new memory form factor. The System on Chip Advanced Memory Module (SOCAMM) design is reportedly a collaborative effort. Team Green's usual set of memory partners—SK Hynix, Samsung, and Micron—were mentioned in early 2025 news articles. Just over a month later, official press material revealed a key forthcoming deployment—Micron stated: "(our) SOCAMM (product), a modular LPDDR5X memory solution, was developed in collaboration with NVIDIA to support the NVIDIA GB300 Grace Blackwell Ultra Superchip. In a (rumored) blow to all involved parties, ZDNet Korea posits that Team Green has postponed the commercialization of their "next-generation low-power DRAM module" IP. According to industry moles, the SOCAMM standard will not debut with this generation of enterprise-focused "Grace-Blackwell" chips. Instead, fresher theories indicate a postponement into next-gen territories—possibly rescheduled to arrive alongside the firm's "Rubin" GPU architecture.

NVIDIA has reportedly sent out notices to major memory partners—(alleged) May 14 updates were received by Samsung Electronics and SK Hynix (in South Korea) and Micron (USA). As a result, SOCAMM supply timelines are (apparently) adjusted. A newer "Cordelia" board design—acting as a substrate for GB300 chips, and compatible with SOCAMM—was in the picture. The latest whispers suggest a return to an existing "Bianca" board configuration, that supports current-gen LPDDR memory modules. ZDNet believes that company engineers have run into several obstacles: "Blackwell chips have been continuously experiencing difficulties in securing design and packaging yields. In fact, the 'Cordelia' board is known to have reliability issues, such as data loss, and SOCAMM has reliability issues, such as heat dissipation characteristics." NVIDIA briefly previewed its futuristic "Rubin Ultra" AI GPU design during GTC 2025—on-stage, a "second half of 2027" release window was teased.

On May 9, the South Korean Radio Agency (RRA) logged four unannounced ASRock Radeon RX 9060 XT SKUs. Since early Spring, harukaze5719 has kept a watchful eye on intriguing pre-launch registrations. For example, they discovered Gigabyte's Gaming OC 16 GB and 8 GB models over a month ago. AMD's board partners seem to be prepping custom options; well in advance of next week's teased unveiling (at Computex 2025). ASRock seems to be cutting things quite fine, with their registering of (allegedly) Navi 44 XT GPU-based Steel Legend and Challenger models. Rumors of a canceled AMD Radeon RX 9060 XT 8 GB (GDDR6) variant emerged a while back, but certain insiders have insisted that this cheaper option will arrive alongside a 16 GB sibling.

Recent RRA filings indicate an upcoming two-pronged approach, involving a good number of AIBs. Buoyed by harukaze5719's fresh findings, VideoCardz carried out additional detective work. They soon unearthed "in stock" ASRock Steel Legend and Challenger factory overclocked Radeon RX 9060 XT cards. Apparently, a Vietnamese store is already courting local client interest—prices are not visible on the unnamed shop's webstore, but staffers are reportedly "encouraging customers to inquire" about costs of ownership. Gaming GPU enthusiasts will recall the amusingly extra early arrival of custom Radeon RX 9070 XT and RX 9070 retail stock in Europe; in anticipation of a supposedly scuppered January launch window. Unlike its main rival, AMD seems to be running on schedule with its next wave of RDNA 4 gaming products.

South Korean semiconductor insiders and analysts believe that Samsung's Foundry business is catching up with a main rival. Earlier this month, TSMC leadership openly discussed an unprecedented demand for 2 nm wafer products. Industry moles believe that the Taiwan's top chipmaker is still ahead of contenders in nearby nations. As a result of an alleged leading and comfortable position, TSMC is reportedly upgrading its state-of-the-art facilities with brand-new equipment—indicating a push into 1.4 nm fields. According to a fresh Chosun Biz news article, Samsung engineers are in the process of narrowing the gap between their 2 nm Gate-All-Around (GAA)—also known as SF2—manufacturing node process and TSMC's equivalent technology.

Last month, leaks suggested SF2 trial yields passing the ~40% mark—in comparison, a ~60% figure was uttered by TSMC insiders. Chosun Biz's sources claim that the South Korean foundry team is close to getting their "2 nm process performance evaluation" into a crucial final stage. Yesterday's report posits that NVIDIA and Qualcomm are in the equation; these VIP clients are purportedly considering SF2 as a "second channel" option. Chosun Biz reckons that Team Green is sizing up Samsung Foundry flagship tech for next-gen commercial and enterprise GPUs. Meanwhile, the San Diego-based smartphone processor specialist could be eyeing up SF2 (for a future AP). The latest inside track info points to 2 nm GAA trial production runs breaking beyond aforementioned (approximate) 40% yield rates. TSMC 2 nm wafer charges are reportedly greater than expected, so big industry players are allegedly investigating "cheaper" non-Taiwanese production avenues.

Last week, GIGABYTE issued an official response to an initial case of "thermal conductive gel slippage," involving an ultra-expensive AORUS GeForce RTX 5080 MASTER ICE, a vertical-mounted graphics card setup, and very non-intensive MMO gaming sessions. The Taiwanese manufacturer believes that this problem is isolated within a first wave of products: "every graphics card is inspected and verified against our quality standards before leaving the factory. The thermal conductive gel is an insulating, deformable, putty-like compound. It is engineered to remain in place when applied properly, and can endure at least 150 °C before any melting or liquification could happen. In some early production batches for the GIGABYTE GeForce RTX 50 Series, a slightly higher volume of gel was applied to ensure sufficient thermal coverage. The overapplication may cause the excessive gel to appear more prominent, extended, and could potentially be separated from the designated area. While the appearance of extra gel might be concerning, this cosmetic variance does not affect the card's performance, reliability, or lifespan. We had already inspected the issue, and adjusted the gel to the optimal amount in (subsequent) production runs."

Despite sending out a public assurance to a worried audience—"(we) take your concerns seriously and want to provide clear information"—GIGABYTE will not be recalling problematic products. VideoCardz reckons that the company is "downplaying" current conditions. Based on further evidence—shared by several members of the TechPowerUp forum (commenting on news coverage)—unfortunately, the first reported case (emerging from South Korea) was not an isolated incident. Given the contents of GIGABYTE's public bulletin, they seem to be aware that this special thermal material (reserved for fancier SKUs) is troubling owners of early batch "GeForce RTX 50 Series and Radeon RX 9000 Series graphics cards." TPU forumite, remekra, shared two images and the following bit of feedback (plus a warning): "I have mine mounted in Lian Li SUP01 case, so GPU is basically standing that's why it drips into the direction of ports. So far it does not overheat on memory modules. I will hold off sending it to GIGABYTE customer service, as I don't have good memories of them; so until it overheats or stops working I will use it. But if you have a vertical case or stand then be aware."

Yesterday, Chosun Daily published a news report that alleges a key "Galaxy Z Flip 7-related" decision made by Samsung leadership. According to smartphone industry moles, the oft-leaked/rumored "Exynos 2500" chipset destined to debut in the company's next-gen (horizontal) foldable smartphone design. As stated in the South Korean insider article: "this is the first time that Samsung's own mobile application processor (AP) is being installed in a foldable phone." Prior to 2025, Galaxy Flip Z product lines made use of Qualcomm Snapdragon chipsets. Throughout early 2025, leaks have linked the "troubled" Exynos 2500 mobile processor to futuristic Galaxy Z Flip 7, Fold 7, and affordable "FE" Enterprise Edition models. The emergence of a superior 2 nm "Exynos 2600" flagship chip—apparently tailored for Galaxy S26 devices (2026)—has allegedly relegated the lesser SoC into lower leagues.

Semiconductor industry watchdogs reckon that the "Exynos 2500" will be manufactured via Samsung Foundry's own 3 nm GAA node process (aka SF3). Notebookcheck commented on this odd choice: "industry estimates from earlier suggested (SF3) yields were around 40%, making it less than ideal for mass production. But it seems Samsung decided to proceed anyway to save on costs, and likely give a new lease of life to its struggling foundry business." Jukanlosreve followed up with additional inside knowledge, just after the publication of Chosun Daily's news piece. The keen tracker of foundry-related revelations let slip with this observation: "the Exynos 2500 being used in the Flip 7 is said to be a lower-clocked chip due to its low yield. So it might as well be called the E2500E." Samsung is expected to unveil its Galaxy Z Flip 7 smartphone family during an "Unpacked" July event. Experts believe that Foundry employees will accumulate useful experience from the mass production of 3 nm parts; thus leading to improved output of finalized 2 nm (SF2) production lines.

The vast majority of High Bandwidth Memory (HBM) new stories—so far, in 2025—have involved or alluded to new-generation SK hynix and Micron products. As mentioned in recently published Samsung Electronics Q1 financial papers, company engineers are still working on "upcoming enhanced HBM3E products." Late last month, a neighbor/main rival publicly showcased their groundbreaking HBM4 memory solution—indicating a market leading development position. Samsung has officially roadmapped a futuristic "sixth-generation" HBM4 technology, but their immediate focus seems to be a targeted sales expansion of incoming "enhanced HBM3E 12H" products. Previously, the firm's Memory Business has lost HBM3 ground—within AI GPU/accelerator market segments—to key competitors.

Industry insiders believe that company leadership will attempt to regain lost market shares in a post-2025 world. As reported by South Korean news outlets, Kim Jae-joon (VP of Samsung's memory department) stated—during a recent earnings call, with analysts—that his team is: "already collaborating with multiple customers on custom versions based on both HBM4 and the enhanced HBM4E." The initiation of commercial shipments is anticipated at some point in 2026, hinging on mass production starting by the second half of this year. The boss notified listeners about development "running on schedule." A Hankyung article alleges that Samsung HBM4 evaluation samples have been sent out to "NVIDIA, Broadcom, and Google." Wccftech posits a positive early outlook: "Samsung will use its own 4 nm process from the foundry division and utilize the 10 nm 6th-generation 1c DRAM, which is known as one of the highest-end in the market. On paper, (their) HBM4 solution will be on par with competing models (from SK hynix), but we will have to wait and see."

An unlucky owner of a GIGABYTE AORUS GeForce RTX 5080 MASTER ICE 16 GB graphics card has reported a baffling instance of thermal gel leakage. A forum post—titled: "5080 oh my god thermal problem"—on the Quasar Zone BBS alerted the wider world to this bizarre fault. The South Korean MMORPG enthusiast described circumstances up until the point of critical liquefaction: "it's been exactly a month since I bought it. I use it for (Blizzard's) World of Warcraft. Two hours of use per day. I set up the card with a riser kit. Thermal (material) is crawling out?!" Early 2025 press coverage has largely focused on other types of unwanted high temperature events involving GeForce RTX 50-series cards, but the seeping out of "server-grade thermal conductive gel" compound is something new. As reported by several PC hardware news outlets, GIGABYTE has utilized fancy thermal conductive gel within flagship SKUs—instead of traditional/conventional thermal pads. This gel was placed over the card's VRAM and MOSFET sections; following fairly light usage (as described above) some of this material started to head down—getting ever closer to the unit's PCIe interface.

Assisted by the AORUS RTX 5080 MASTER ICE's vertical orientation, the (apparently) highly deformable, but non-fluid thermal gel was susceptible to the effects of gravity. JC Hyun System Co., Ltd.—GIGABYTE's official domestic importer (for South Korea)—weighed in with a separate bulletin: "we are aware of the thermal gel issue with the GIGABYTE GeForce RTX 50 series, which was first posted on Quasar Zone—(we) are currently discussing the thermal gel issue with GIGABYTE HQ and future customer service regulations. In addition, we sincerely apologize for the confusion caused to many customers who love and use GIGABYTE products due to inaccurate guidance provided to customers who received the products due to unclear customer service regulations regarding the issue that occurred this time. Lastly, when the manufacturer's customer service policy regarding this thermal gel issue is finalized, we will also forward the service policy to CS Innovation so that it can be processed smoothly in accordance with the service policy. We will also provide information through a separate post so that more customers can be aware of the information." As mentioned by Notebookcheck, GIGABYTE uses this special thermal gel solution on other highly expensive custom: "RTX 50-series cards like the GeForce RTX 5090 XTREME WATERFORCE 32G, RTX 5090 MASTER ICE, RTX 5070 Ti MASTER, and others."

Earlier today, TechPowerUp covered the alleged performance prowess of Huawei's CloudMatrix 384 system super node. According to SemiAnalysis opinion, the system's Ascend 910C AI accelerators are a generation behind—in terms of chip performance—when compared to NVIDIA's GB200 "Blackwell" AI GPU design. SMIC seemed to be in the picture, as Huawei's main fabrication partner—possibly with an in-progress 5 nm node process. Instead, SemiAnalysis has surmised that the Ascend 910C is based on plenty of non-native technologies. Huawei's (current and prior) "aggressive skirting of export controls" has likely enabled the new-gen AI chip's better than expected performance stats. SemiAnalysis documented the early sample's origins: "while the Ascend chip can be fabricated at SMIC, we note that this is a global chip that has HBM from Korea (Samsung), primary wafer production from TSMC (Taiwan), and is fabricated by 10s of billions of wafer fabrication equipment from the US, Netherlands, and Japan...One common misconception is that Huawei's 910C is made in China. It is entirely designed there, but China still relies heavily on foreign production."

Despite China's premiere foundry business making pleasing in-roads with a theorized "7 nm N+2" manufacturing test line, Huawei has seemingly grown impatient with native immature production options. Today's SemiAnalysis article presents a decent dose of inside knowledge: "while SMIC does have 7 nm, the vast majority of Ascend 910B and 910C are made with TSMC's 7 nm. In fact, the US Government, TechInsights, and others have acquired Ascend 910B and 910C and every single one used TSMC dies. Huawei was able to circumvent the sanctions on them against TSMC by purchasing ~$500 million of 7 nm wafers through another company, Sophgo...It is rumored Huawei continues to receive wafers from TSMC via another 3rd party firm, but we cannot verify this rumor." Another (fabless) Chinese chip design firm—Xiaomi—appears to still have direct/unrestricted access to TSMC manufacturing lines, albeit not for enterprise-grade AI products.

According to the latest South Korean semiconductor industry whispers, Samsung's 2 nm GAA node process (aka SF2) development team has hit another pleasing experimental production milestone. An Asia Economy SK news article has sourced insights from inside track players—one unnamed mole posited that: "the 2 nm yield currently under development at Samsung Foundry is much better than previously known...and more positive than the (reportedly abandoned) 3 nm process." A combination of relatively new leadership and a rumored welcoming of first wave High-NA EUV equipment has likely bolstered next-gen efforts, after late 2024's alleged failure of 3 nm prototypes. Leaks from earlier in 2025 indicated SF2 test yields wavering around 20-30%; far from ideal—back then, insider reports suggested that TSMC was well on the way to achieving 60% rates with a competing 2 nm product line. Asia Economy has picked up on mutterings about Samsung's current progress—latest outputs: "have exceeded 40% in the wafer testing stage at a post-processing company."

Industry watchdogs reckon that the South Korean's foundry business is making good progress; perhaps on track to commence speculated mass production by the third quarter of this year—just in time to get finalized flagship "Exynos 2600" mobile chips in the manufacturing pipeline. The Taiwanese rumor mill indicated a major milestone "completion" of TSMC's 2 nm trial phase at some point last month—insiders mentioned excellent yield rates: in the region of 70-80%. Cross-facility mass production could start later this year, but experts propose that the market leader will be implementing price hikes. These "elevated charges" could send loyal TSMC customers in the direction of an alternate source of 2 nm wafers: Samsung. Fresh semicon biz gossip has the likes of Apple, AMD and NVIDIA in the picture.

Samsung's foundry business seems to be busying itself with the rumored refinement of a 2 nm GAA (SF2) manufacturing node process—for possible mass production by the end of 2025, but company leadership will very likely be considering longer term goals. Mid-way through last month, industry moles posited that the megacorporation's semiconductor branch was questioning the future of a further out 1.4 nm (SF1.4) production line. Officially published roadmaps have this advanced technology rolling out by 2027. Despite present day "turmoil," insiders believe that a new team has been established—tasked with the creation of a so-called "dream semiconductor process." According to a fresh Sedaily news article, this fledgling department has started development of a 1 nm foundry process.

Anonymous sources claim that Samsung executives are keeping a watchful eye on a main competitor—as stated in the latest South Korean report: "there is a realistic gap with Taiwan's TSMC in technologies that are close to mass production, such as the 2 nm process, the company plans to speed up the development of the 1 nm process, a future technology, to create an opportunity for a turnaround." A portion of the alleged "1 nm development chip team" reportedly consists of veteran researchers from prior-gen projects. Semiconductor industry watchdogs theorize that a canceled SF1.4 line could be replaced by an even more advanced process. Sedaily outlined necessary hardware upgrades: "the 1.0 nanometer process requires a new technology concept that breaks the mold of existing designs as well as the introduction of next-generation equipment such as high-NA EUV exposure equipment. The company is targeting mass production after 2029." Samsung's current Advanced Technology Roadmap does not extend beyond 2027—inside sources claim that the decision to roll with 1.0 nm was made at some point last month.

So far, NVIDIA's GeForce RTX 50xx graphics card models have shipped with Samsung GDDR7 memory modules onboard. According to a fresh MEGAsizeGPU (aka @Zed__Wang) claim, a change in vendor has already occurred. The tenured tracker of Team Green inside track information believes that the company has: "started to use SK hynix GDDR7 for the GeForce RTX 50 series graphics cards. Started with GeForce RTX 5070 first." Officially, NVIDIA's latest board designs can support GDDR7 modules produced by the "big three:" Samsung, SK hynix and Micron (see BIOS info below). Team Green's comfortable market leading position probably grants plenty of negotiation power to pick and choose the best component deals. Day one evaluators performed teardowns on GeForce RTX 50 series review samples; TechPowerUp's W1zzard found Samsung "K4VAF325ZC-SC32" GDDR7 units—rated for 32 Gbps—onboard various GeForce RTX 5080 16 GB models. As outlined by VideoCardz, the rest of NVIDIA's "Blackwell" gaming product stack sticks with 28 Gbps-rated Samsung GDDR7 modules, extending to its Mobile portfolio.

While semiconductors are exempt (for now at least) from Trump's tariffs, other components going into our PCs are not. According to Tom's Hardware, which spoke to multiple system integrators, tariffs are about to hike PC costs by at least 20%, with system integrators hurt the most. The tariff package imposes a 54% rate on Chinese goods, 34% on top of earlier tariffs, and significant duties on Taiwan, South Korea, and Vietnam products. These countries supply essential PC components such as SSDs, RAM, cases, and graphics cards. Wallace Santos, CEO of Maingear, highlighted the immediate effects on production: "Tariffs have a direct impact on our cost structure… which we have to pass down to our customers." He further explained that some suppliers have halted production in China, leading to scarcity and escalating costs. Santos estimates that prices for his PCs will rise "20 to 25% as a result of the tariffs."

Other company leaders express concern over the limited alternatives available. Kelt Reeves, CEO of Falcon Northwest, stated, "Sadly the overwhelming majority of PC component manufacturing is not done in the US and never has been. There's no US alternative supplier for most PC parts." Reeves added that even US-based system integrators are "facing skyrocketing costs" due to the tariffs, which are set to worsen an already challenging market situation caused by ongoing GPU shortages. Jon Bach, CEO of Puget Systems, shared his perspective in a recent blog post, noting that his company might absorb some costs to minimize consumer price increases. However, even before the latest tariff updates, Bach predicted a price rise of "20 to 45 percent by June." Critics of the tariffs warn of broader economic issues. Gary Shapiro, CEO of the Consumer Technology Association, condemned the policy as "massive tax hikes on Americans that will drive inflation, kill jobs on Main Street, and may cause a recession for the US economy." With these tariffs taking effect, the PC industry faces a period of adjustment marked by increased costs and significant supply chain challenges.

Earlier today, industry moles in South Korea have heard whispers about Samsung Electronics planning a new pricing strategy for NAND and DRAM product lines. According to an MK news articles, local sources believe that company leadership will: "raise memory chip prices—by 3-5% from the current level—for major global customers. It is reported that some customers have already begun contract negotiations that reflect the increase conditions." Regional watchdogs posit that the megacorporation is reacting to very current geopolitical tensions (i.e. tariffs). Earlier this week, a main rival—Micron—informed customers about forthcoming memory price increases. Naturally, the North American memory chip giant is not "fully" affected by recent seismic shifts. A "significant growth demand" has caused jacked up charges—effective across DRAM, NAND flash, and HBM portfolios—projected throughout 2025 and 2026.

Returning to South Korean shores and Samsung, one unnamed semiconductor insider opined to MK: "oversupply continued throughout last year, but supply has recently decreased as major companies have begun to reduce production...In addition, artificial intelligence (AI) devices are appearing one after another in China, and demand for semiconductors is gradually increasing due to industrial automation." DRAMeXchange—an appropriately named market research organization—has kept track relevant trends. As disclosed by the MK news piece—as of last month, general-purpose DRAM DDR4 prices: "remained flat for the fourth month in a row." Looking at conditions for DDR5 (used in high-performance PCs and enterprise equipment), prices soared by 12%. DRAMeXchange observed NAND costs rising by 9.6%: "continuing an upward trend for the third consecutive month."

According to a fresh Money Today SK news article, Samsung is expected to launch a next-generation Galaxy Ring model later this year—this tiny wearable device is touted to operate with a "dream battery" design. The South Korean giant's Electro-Mechanics division is reportedly tasked with the challenging development of "all-solid-state" batteries for all manner of ultraportable products. Yesterday's report suggests that Samsung's upcoming Galaxy Ring sequel—apparently scheduled for launch within Q4'25—will be driven by the Electro-Mechanics team's pioneering effort. The production of all-solid-state battery units is an expensive endeavor, so industry watchdogs have predicted tough retail conditions for the forthcoming "Galaxy Ring 2" rollout—the original unit was not exactly a "hot property" in terms of sales figures.

Money Today's inside sources reckon that the Electro-Mechanics branch will—eventually—fit all-solid-state battery designs inside new-gen earphones (aka Galaxy Buds) by Q4 2026, and very futuristic smartwatches by the end of 2027. Given cost considerations, larger all-solid-state solutions—potentially for usage in smartphones—are not in the pipeline. Around early February of this year, the development of Samsung's (inevitable) "Galaxy S26" mobile series was linked to alleged 6000+ mAh silicon-carbon battery units. The South Korean's smartphone engineering team is reportedly trying to play catch up with more advanced solutions, as devised by competitors in China. The status of Samsung's proprietary silicon-carbon prototype is the subject of much online debate, but certain insiders believe that employees are still working hard on the perfection of an ideal "battery formula."

Over the past weekend Jukanlosreve declared via social media that Samsung's: "Exynos 2600 (mobile SoC) is definitely back, and it will be used in the Galaxy S26 series. But the chip volume is so limited that it'll likely be similar to the Exynos 990 situation. I'm not sure if SF2 is actually any good." Mid-way through March, the keen observer of semiconductor industry conditions posited that Samsung's Foundry business could abandon a 1.4 nm (SF1.4) process node. SF2 (aka 2 nm GAA) seems to be in a healthier place, according to insiders—thanks to rumored assistance from an external AI-specialist partner. The development of next-generation flagship Exynos smartphone processors are allegedly closely tied with Samsung Foundry's 2 nm GAA manufacturing process.

As alluded to by Jukanlosreve's recent prediction, the statuses of leaked 2 nm-based "Exynos 2600" and "Exynos 2500" chips were often questioned by industry watchdogs in the past. The latter is purportedly destined for rollout in forthcoming affordable "Galaxy Z Flip FE" models, albeit in mature 4 nm form. Vhsss_God—another source of inside track info—has weighed in on the topic of Samsung's next-gen chipset roadmap. Compared to Jukanlosreve's musings, their similarly-timed weekend projection seemed to be quite fanciful: "exclusive leak...Samsung doesn't want to use Exynos or Qualcomm Snapdragon chips anymore. S26 line is targeted to launch with the new Samsung developed chip (2 nm)—formerly referred to as Exynos 2600. (The company) will try its hardest to ditch Snapdragon on the entire Galaxy line next year." Perhaps there is too much of a negative stigma attached to Samsung's long-running chipset nomenclature, but the majority of foundry moles continue to label incoming designs as Exynos processors.

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.