Navigating AMD's various modern processor model naming schemes is tricky business, and another layer of complexity has been added this week; with newly revealed "Strix Point" and "Krackan Point" CPUs. Three previously unannounced SKUs have appeared on AMD's Chinese website: Ryzen AI 9 H 365, Ryzen AI 7 H 350, and Ryzen AI 5 H 340. At first glance, these models codes seem to be familiar—after a double take, we see a small difference in nomenclature. The addition of a middle-placed/detached "H" has press outlets and hardware enthusiasts scratching their collective heads. After discovering Team Red's Ryzen AI 9 H 365 APU listing, Lonely City Hardware posted a humorous observation on social media: "for the Chinese market. No one can remember the full name."

Frustratingly, Western search engines (at the time of writing) just point you to the non-H equivalents. The "H" designation denotes Chinese market exclusivity; TechPowerUp has covered previous-generation examples in the recent past. When comparing technical details and specification sheets to Western equivalents, VideoCardz noted that there were no apparent differences—platform and packaging are the same (FP8). A cross-reference—of clock speeds, thermal ratings, and core architecture (Zen 5/Zen 5C)—points to spec parity across East and West variants.

Yesterday, Nintendo's Hong Kong office announced the establishment of a new subsidiary company in Taiwan—specifically, in Taipei City. Their official statement mostly outlines upcoming improved service benefits for local customers. The House of Mario has relied on contracted partners to take care of smaller regional markets. Evidently, their Taiwanese audience has relied on a third party company for over a decade. An older subsidiary—Nintendo Phuten—was shuttered back in 2014. The successor is chaired by Hiroyuki Matsumoto—on February 18, the new company representative delivered a message: "Thank you for your long-term support and love, I would like to express my sincere thanks. In order to further strengthen the business foundation in the Taiwan market and improve the service quality of customers, the company will establish a new local legal person 'Taiwan Nintendo Co., Ltd.' as one of the subsidiaries of Nintendo Co., Ltd. Starting from April 1, 2025, we will officially transfer our business in Taiwan to 'Taiwan Nintendo Co., Ltd.', and adhere to the business philosophy of Nintendo Group to continue to promote business development." As reported by Nintendo Life, the veteran video game house has made serious inroads in the region—Taiwanese fans were greeted by a larger than expected (current-gen) Switch console showcase at Taipei Game Show's 2024 edition.

On a surface level, Nintendo's reestablished operation in Taiwan seems to be a customer-focused initiative. Certain gaming news outlets have disclosed more elaborate theories; based on reported problematic market conditions in China. The Chinese Nintendo eShop will be phased out by mid-May 2026, likely in reaction to the government's introduction of new rules that: "limit the encouragement of spending in online games and battle video game addiction among young people." Coincidentally, Shuntaro Furukawa (Nintendo's President) recently new announced "contingency plans"—with a manufacturing model that will become less reliant on Chinese factories. The company chief discussed revised strategies in an interview with Reuters: "Nintendo Switch is not only manufactured in China, but in places such as Vietnam and Cambodia as well. We are predicting various geopolitical risks and establishing ways to respond...While we anticipate a certain impact, the influence on this year's financial results is expected to be minimal. We will continue to observe the trends, and thoroughly consider how to respond."

Loongson Technology has announced its next-generation 3B6600 processor—according to recent Chinese press reports, early details were disclosed in "investor relations activity" documentation. Their eight-core design is still in development, with the fabless company (allegedly) setting some lofty performance goals. According to a MyDrivers report (citing Fast Technology), the 3B6600 processor's single-core and multi-core "high level" performance is touted to match that of Core i5 and i7 offerings from Intel's 12th and 13th Generation portfolios. In the past, Loongson has compared its older products to Western equivalents. Last year, their 3A6000 chip design was mentioned in the same sentence as "AMD Zen 4 and Intel Raptor Lake" processor families. Around November 2023, reports had it approximating Team Blue Core i3-10100 "Comet Lake" CPU performance.

The next-gen 3B6600 CPU is supposedly using its predecessor (3A6000) as a springboard; MyDrivers believes that the same in-house LoongArch ISA design (fabricated on 12 nm/14 nm) will be revisited. The publication highlighted a significant area of optimization: "single-core performance of 3B6600 is expected to be in the world's leading ranks...Loongson 3B6600 will continue to use mature technology, and the architecture core will be upgraded to the new LA864. The same frequency performance is greatly improved by about 30%, compared with the Loongson 3A6000's LA664 architecture." Loongson engineers are reportedly targeting a maximum turboboost frequency of 3.0 GHz, but under normal operation the 3B6600 chip is predicted to offer a main frequency of 2.5 GHz. The new design will integrate a new "LG200 GPGPU" graphics computing core—additionally, supported standards include: DDR5 memory, PCIe 4.0 bus, and HDMI 2.1.

In an interview with The Telegraph, Acer CEO and chairman Jason Chen said that its products made in the PRC will see a price increase of 10 percent as direct results of the new tariffs that the US will levy on electronics. However, Mr Chen is quoted as saying "We think 10 percent probably will be the default price increase because of the import tax." which doesn't mean it will be exactly 10 percent, as it might vary a bit between product segments. That said, what's clear is that Acer and most likely every other company that manufactures hardware in the PRC aren't going to eat any of the tariffs, as the companies appear to be shifting the burden of the new tariffs straight over to the end consumers. Mr Chen also suggested that some companies might be increasing their pricing by more than 10 percent.

The price increase will happen over time, as the new tariffs won't affect products that have left the PRC before the end of February. Alongside Acer, which is the fifth-biggest computer brand in the US market, it's likely that Dell, HP and Lenovo, as well as Apple, are going to hike their prices by the same 10 percent or more. Acer moved the assembly of its desktop computers out of the PRC during Trump's previous term, when a 25 percent tariff was imposed. Now Acer is looking at moving at least some additional parts of its productions out of the PRC and the US is on the table for some of its products. Considering that some 80 percent of all laptops imported to the US are made in the PRC, the Consumer Trade Association is expecting the new tariffs to cost US consumers some US$143 billion, which it assumes will lead to a slump in sales of consumer electronics.

Inside sources—familiar with goings-ons at leading DRAM manufacturing firms—have predicted the end of DDR3 and DDR4 production lines. According to a DigiTimes Asia report (citing Nikkei), industry observers have noticed that the DRAM market is undergoing a so-called "shift." They believe that pricing trends are decreasing due to weak demand. Samsung Electronics, SK Hynix, and Micron are named as major players; allegedly involved in devising new strategies—in reaction to fluid market circumstances. The DigiTimes insider network proposes that the big three: "may phase out DDR3 and DDR4 by 2025...by the end of 2025...anticipating a future focused on advanced memory technologies." Older standards are falling out of favor, with DDR5 and high-bandwidth memory (HBM) on the ascent. Industry watchdogs reckon that possible DDR3 and DDR4 supply shortages could occur "post-summer 2025."

Taiwan's Nanya Technology has predicted that the overall DRAM market will "bottom out" within the first half of 2025. An eventual recovery is envisioned by the second quarter; AI-related demands could help drive up demand by a large margin. Additionally, Nanya points to improved inventory management and global economic stimulus. Taiwanese DRAM production houses are expected to pick up some slack, but an unnamed "key component distributor" anticipates serious after-effects. An anonymous source believes that the: "anticipated halt in production could lead to significant supply constraints, challenging market dynamics and impacting pricing strategies." Nanya Technology and Winbond Electronics produce specialized DRAM-types; therefore are not touted to be great gap fillers. The latter is reportedly reacting to weak demand for "mature" DDR products—DigiTimes commented on this development: "Winbond Electronics is advancing its manufacturing by transitioning to a 16 nm process in the latter half of 2025. This upgrade from the current 20 nm process, primarily used for 4 Gb DDR3 and DDR4, will enable Winbond to produce 8 Gb DDR memory."

Yesterday, the ASUS Chinese office announced a special event—on February 25—dedicated to launching a next-gen AMD APU-powered premium notebook model: "ROG Magic X (or Illusion X) is the first to be equipped with the Ryzen AI MAX+ three-in-one chip, which can efficiently coordinate multiple modes, provide combat power and computing power on demand, and can handle e-sports, creation and AI with one chip!" The manufacturer's Weibo post has generated plenty of buzz; industry insiders reckon that reviews could be published today (February 18)—HXL/9550pro informed VideoCardz with a not so cryptic message: "STX-Halo NDA: Feb 18th 2025."

Western press outlets point out that the ROG Magic X is a local variant of the familiar ROG Flow Z13 design; a 2025 refresh brings in AMD's much anticipated "Strix Halo" APU design. Team Red-authored marketing material and pre-release evaluation leaks have hinted about impressive integrated graphics solution performance; equalling or even exceeding that of previous-gen dGPUs. Well-known North American hardware review outlets have dropped hints (NDA permitting) about AMD's Ryzen Al Max+ 395 and Max 390 processors. Hardware Canucks could barely contain their excitement regarding the potent Zen 4 and RDNA 3.5 combo package; to the point of wish listing a potential direct successor: "Strix Halo is one of the most exciting things launched into the PC space in the last half decade. Full stop...AMD can't keep this as a one-off. If it's followed up with Zen 6 and RDNA 4 next year...watch out." Naturally, Team Red's cutting-edge mobile CPU technology is arriving in devices with high asking prices. The aforementioned ROG Flow Z13 2025 model—configured with top specs—is priced at $2699. Notebookcheck reckons that ASUS has tacked on an extra $500, since an announcement of initial pricing at CES 2025.

Sapphire Technology has introduced a new NITRO+ motherboard model in China—weekend reports pointed to a listing on JD.com, where the latest addition resides alongside various previous-gen B650 options. This appears to be another region specific release—priced at 1299 RMB (~$179 USD). The NITRO+ B850M Wi-Fi model seems to share its overall aesthetic design with the preceding B650M-equipped equivalent. As surmised by VideoCardz, the Sapphire NITRO+ board's VRM patterning reminded them of the leaked premium-tier Radeon NITRO+ RX 9070 graphics card series.

The generational jump from AMD B650 to B850 chipset—as expected—yields only minor benefits (see below). The older NITRO+ B650M Wi-Fi model is a significantly cheaper prospect; coming in at 899 RMB ($124 USD). VideoCardz's weekend detective work settled on a nice improvement: "based on our research, the memory overclocking support has increased from 7600 MT/s+ to 8000 MT/s. Keep in mind that these figures may be higher for both motherboards; this is just what Sapphire guarantees when paired with a supported memory kit." Industry watchdogs believe that ASRock is likely involved in the contract manufacturing of Sapphire's motherboard offerings—the former has been long-involved in pumping out NZXT-badged products.

The notion of designing tailor-made AI-crunching chips is nothing new; several major organizations—with access to big coffers—are engaged in the formulation of proprietary hardware. A new DigiTimes Asia report suggests that DeepSeek is the latest company to jump on the in-house design bandwagon. The publication's insider network believes that the open-source large language model development house has: "initiated a major recruitment drive for semiconductor design talent, signaling potential plans to develop its proprietary processors." The recent news cycle has highlighted DeepSeek's deep reliance on an NVIDIA ecosystem, despite alternative options emerging from local sources.

Industry watchdogs believe that DeepSeek has access to 10,000 of sanction-approved Team Green "Hopper" H800 AI chips, and (now banned) 10,000 H100 AI GPUs. Around late January, DeepSeek's Scale AI CEO—Alexandr Wang—claimed that the organization could utilize up to 50,000 H100 chips for model training purposes. This unsubstantiated declaration raised eyebrows; given current global political tensions. Press outlets have speculated that DeepSeek is in no rush to reveal its full deck of cards, but they appear to have a competitive volume of resources; when lined up against with Western competitors. The DigiTimes news article did not provide any detailed insight into the rumored in-house chip design. DeepSeek faces a major challenge; the Chinese semiconductor industry trails behind market leading regions. Will local foundries be able to provide an advanced enough node process for required purposes?

NDA-busting renders of the much-anticipated and rumored Nintendo Switch 2 console appeared online late last year—reports suggested that a member of the Xiahongshu forum was involved in this dramatic leak. The source seemingly had access to a 3D CAD model—possibly procured from a manufacturing partner. Industry insiders believed that top Nintendo brass were incensed by the pre-Christmas 2024 leaks. Around mid-January, an official unveiling of the next-gen gaming handheld arrived online, courtesy of a relatively short teaser video. Early April public showcases are on the calendar; press outlets and regular punters will be participating in hands-on experiences (invite-only) at venues across North America, Europe, Asia and Australia. The flow of early 2025 Switch successor leaks have seemingly slowed down, but renewed activity on Reddit points to early samples emerging via black market outlets.

The aforementioned Xiahongshu forum member is reportedly considering a purchase of pre-release Switch 2 units—previous boasts have indicated that they already own one example. Screenshots of alleged interactions between the seller and potential client were uploaded to Chinese discussion boards, and then shared on the Nintendo Switch 2 Subreddit. The anonymous black market dealer reckons that stock will be available in the near future—claiming by "next week"—with an asking price of 290,000 RMB per package (~$39,780 USD). This steep demand—allegedly—grants early access to the core console tablet, Joy-Con controllers, and dock. Online community debates have produced several theories, regarding the leaker's motivations. Given their history of selling CAD models to accessory makers; this working relationship could develop into a—presumably more profitable—supply of actual working hardware.

Last week, AMD China and involved board partners launched region-exclusive Radeon RX 7650 GRE graphics card models; initial publicity produced a mountain of promotional imagery, but only a minority of AIBs published GPU specification sheets. On Tuesday (February 11), new Team Red documentation revealed a slightly confusing change in nomenclature—as of 2025, GRE stands for "Great Radeon Edition," rather than "Golden Rabbit Edition." On launch day, ASRock China showcased their Navi 33 GPU-based Radeon RX 7650 GRE Challenger 8 GB OC Edition card. A product page has appeared on the company's local website, complete with technical information. We are looking at 8 GB of GDDR6 VRAM, 2048 stream processors, a game clock frequency of 2539 MHz, and a boost clock of "up to 2810 MHz." The ASRock official store price is listed as 2049 RMB (~$281 USD).

Another model, the ASRock Radeon RX 7650 GRE Challenger OC+ Edition has appeared online; its official product page looks almost identical to the OC (non-plus) Edition's. The tacked-on "+" denotes key differences in game clock and boost clock factory settings—ASRock has implemented higher frequencies: 2400 MHz and 2725 MHz (respectively). After analyzing promotional imagery, the slightly fancier card appears to sport a chunkier shroud design. The new "Great Radeon Edition" OC+ Edition is not listed on ASRock's official e-tail outlet. Amusingly, the two product pages contain the same error; their "main specification" sections point to a non-existent AMD Radeon RX 7600 GRE GPU. Team Red's Radeon RX 7650 GRE GPU seems to be a slightly tweaked variant of the Radeon RX 7600 (Navi 33 XL) 8 GB model.

China's dominance in semiconductor equipment procurement is expected to face its first setback since 2021, with spending projected to decrease from $41 billion to $38 billion in 2025, according to semiconductor research firm TechInsights. This 6% decline marks a significant shift for the world's largest buyer of wafer fabrication equipment, whose purchases represented 40% of global sales in 2024. The downturn reflects mounting pressures from both market dynamics and geopolitical constraints. US export controls targeting advanced semiconductor capabilities have intensified while domestic chipmakers grapple with overcapacity in mature node segments. SMIC, China's leading foundry, has already signaled concerns about oversupply risks in this sector, where Chinese manufacturers have rapidly expanded their market share against Taiwanese competitors.

Despite these headwinds, Chinese equipment manufacturers have notably advanced domestic capability development. Naura Technology Group has emerged as the seventh-largest global equipment manufacturer, while AMEC continues to expand its international presence. However, critical gaps persist in China's semiconductor equipment ecosystem, particularly in lithography systems, where dependence on foreign suppliers like ASML remains high. TechInsights data reveals that Chinese companies supplied only 17% of testing tools and 10% of domestic assembly equipment in 2023. The spending reduction comes after a period of aggressive stockpiling prompted by US sanctions to limit Beijing's access to advanced chipmaking capabilities, especially those applicable to artificial intelligence and military applications. However, Chinese manufacturers have demonstrated resilience, with SMIC and Huawei successfully producing advanced chips through alternative, albeit more costly, manufacturing methods.

Late last week, Colorful Technology revealed a brand-new series of "CVN" motherboards—the manufacturer's Bilibili account hinted at an upcoming launch and uploaded three highly-stylized promotional images. As reported by ITHome, Colorful's debut AMD X870 chipset-based board will be: "equipped with Colorful's first-generation graphics card quick-release device." Western press outlets have covered the functional validity of similar systems in recent times, but Colorful's teaser seems to be poking fun at a rival's current predicament. The incoming CVN X870 series will sport a unique selling point; its GPU removal system features an "ON or OFF" switch. Colorful advertises this proprietary design as being "really safe."

CVN X870 or X870E product pages have not appeared online, but ITHome reckons that this new range will sit in-between Colorful's familiar iGame and Tomahawk motherboard lines. The intended target audience appears to be "mainstream DIY" builders. According to one of the teaser images, Colorful reckons that its silver/white-tinted CVN X870 boards will "live up to expectations." The Chinese manufacturer has devised a new "Frost Armor" structure that covers a board's VRM sections—sporting "composite heat dissipation grooves and heat pipes." Lower-end CVN B840 and B850 models could be in the pipeline, but Colorful is likely testing the waters with its initial launch of X870-equipped fare.

AMD's China customers are now surprised by AMD's claim that the Radeon RX 7650's "GRE" nomenclature actually translates into "Great Radeon Edition" and not into "Golden Rabbit Edition" as we had previously thought. The Radeon RX 7900 GRE was the first "GRE" GPU to debut in the Chinese market. Back in 2023, China celebrated the Year of the Golden Rabbit; hence, AMD made a reference to that and named its GPUs appropriately to appeal to Chinese gamers. However, with 2025 being the year of the snake, AMD decided to just find a new meaning for its GRE branding instead of replacing it altogether. Now, AMD's official documents point to AMD Radeon RX 7650 Great Radeon Edition (GRE).

As a reminder, this RX 7650 GRE "great" SKU carries a Navi 33 GPU with 32 CUs translating into 2048 SPs. Clocked at 2.69 GHz, its 8 GB GDDR6 memory configuration positions it in the low-middle-class gaming segment. Scheduled for a launch in February, it will be priced at 2,099 RMB or 287 USD at the time of writing.

Mid-way through January, insider sources reckoned that AMD's China-exclusive Radeon RX 6750 GRE 10 GB GPU was on the way out—a rumored successor was seemingly in line to replace this popular budget-friendly RDNA 2-based model. According to VideoCardz, Team Red's latest "Golden Rabbit Edition" design has launched—exclusively for a China-based buying audience. Team Red and its local board partners have produced a plethora of Radeon RX 7650 GRE 8 GB custom models—reports suggest that a reference model (MBA) does not exist. Several manufacturers have models readied for launch, including: ASRock, ASUS, PowerColor, Sapphire, Yeston and VASTARMOR. Official MSRP for the region is 2099 yuan (~$289 USD). In the West, original launch pricing for the Radeon RX 7600 was $269. The fancier Radeon RX 7600 XT was priced at $329 on day one.

Roughly a year ago, AMD decided against releasing its (RDNA 3) Radeon RX 7600 XT 16 GB GPU in China. Their Radeon RX 7600 (non-XT) 8 GB model did make it to China, but it was largely overshadowed by older tech—mainly Team Red's compellingly priced Radeon RX 6750 GRE 12 GB and 10 GB cards. Spec-wise, the Radeon RX 7650 GRE sits somewhere in-between its two "Hotpink Bonefish" Navi 33 siblings. All three cards share the same number of compute units (32 CUs), but differ in terms of boost clocks and thermal design power ratings. The GRE features a max. boost clock speed of 2695 MHz and a board TDP of 170 W—just a nudge over the Radeon RX 7600's 2655 MHz boost capability and 165 W board power rating. Both cards are frugal enough to rely on a single 8-pin power connector. By contrast, the hungrier 190 W TDP-rated Radeon RX 7600 XT sports two 8-pin inputs—this GPU can boost up to 2755 MHz. VideoCardz has kindly assembled a comparison chart (see below).

ASRock told PCMag it plans to move some of its production out of China. "We need time to shift the manufacturing of GPU cards and other products hit by the 10% tariff to different countries," they said. This week, the White House put a 10% tax on all Chinese imports to the US, this tax applies on top of any other taxes the US already had on certain Chinese goods. ASRock also said, "While we move from making things in China to making them elsewhere, we might take on some of the cost and raise prices a bit to show the higher costs." But they added, "It's not easy to raise prices because the market is still very competitive." ASRock also told PCMag that it already pays a 25% tax on its power supplies made in China. "For items like PSUs that already have an extra 25% tax, makers will keep doing what they've been doing," the company said.

If Trump administration doesn't follow through with his threats of huge tariffs against Taiwan, the PC gaming industry will primarily feel the effects on companies like ASRock and MSI (which makes its motherboards in Shenzhen, China). These are the component and peripheral makers that have part or all of their manufacturing processes in China. ASRock's announcement isn't a huge surprise, as we saw hints of this trend in late 2024 when PC Partner (second-biggest graphics card maker, producing PCBs for brands such as Inno3D and Zotac) moved its headquarters from China to Singapore. It will be no surprise if other top-tier brands such as GIGABYTE, MSI, and ASUS take similar actions sooner or later.

Priced at the equivalent of $116, the China-exclusive Ryzen 5 7400F is a 6-core/12-thread processor powered by the older "Zen 4" microarchitecture, but which is based on the "Raphael" MCM, giving it a comprehensive PCIe Gen 5 I/O (as opposed to designing such a chip based on the "Phoenix Point" monolithic silicon with its PCIe Gen 4 I/O). The chip benefits from the full 32 MB on-die L3 cache being enabled on the "Zen 4" CCD, besides the full 1 MB per core L2 cache, and that I/O, but misses out on the iGPU. At its price, the 7400F is attracting a segment of value-conscious gamers.

To bring this chip at its price point, AMD had to cut costs somewhere, and Chinese PC enthusiast "Melon Master" soon found out where. On de-lidding (removing IHS) of the Ryzen 5 7400F, it was found that the chip does not use a soldered TIM like the other Ryzen 7000 series "Raphael" processors, instead using more cost-effective thermal paste, as is common in entry-level chips from both AMD and Intel. Harukaze5719 comments that this could be the reason the chip has a fairly noticeable Cinebench R23 performance delta against the Ryzen 5 7500F that's globally available, and has STIM, but is priced at $145. The 7400F and 7500F share the same base frequency of 3.70 GHz, but while the former boosts up to 4.70 GHz, the latter goes all the way up to 5.00 GHz.

A South Korea-based hardware enthusiast—Harukaze5719—has discovered a curious listing of ZOTAC's GeForce RTX 5090 SOLID OC model on a Chinese auction site (asking price: ~$4175 USD). This finding was shared with their audience, via a social media post—a screenshot was accompanied by harukaze5719's short and succinct message: "OMG..." Officially, NVIDIA and its board partners cater to the Chinese hardware market with a restricted variant of the flagship "Blackwell" GPU—GeForce RTX 5090D. Despite under-the-hood nerfing, this region-exclusive model still offers enviable performance (when pushed).

The "full fat" GeForce RTX 5090 GPU appears to be an even hotter commodity—with demand (at launch) exceeding far exceeding supply. ZOTAC's South Korean branch warned potential customers about difficult conditions almost two weeks ahead of Blackwell's market debut on January 30. Harukaze5719 has likely expressed semi-personal frustration over apparent South Korea market-destined ZOTAC stock turning up for sale in China. So-called "Chinese GPU shopping tourists" could have made the journey to a South Korean PC hardware store, with their sole objective being the acquisition of GeForce RTX 5090 graphics cards. Tom's Hardware has gathered various related tidbits from Japanese and Taiwanese news sources—where local TV coverage explored the events of last week's launch event.

Japan's central government will adjust its control over cutting-edge chips and quantum computer-related technology—a Japan Times news article proposes that new regulations will come into effect by the end of May. The report suggests that the nation's governing body is expanding its list of export-controlled items to include: "advanced chips, lithography equipment and cryo-coolers needed for the manufacture of quantum computers." The publication has gathered this information from revised foreign exchange laws. According to economy ministry officials, companies will be required to apply for external trade licenses—extra measures are being put in place to prevent the export of cutting-edge items to foreign military organizations. The updated terms are viewed as another step in tightening the supply of advanced semiconductor products to mainland China. Recent global events have paved the way for a new wave of AI chip-related restrictions.

Naturally, China has expressed concern regarding upcoming changes—they anticipate problems affecting supply chains and normal commercial exchanges between enterprises. According to the Japan Times report, the Ministry of Commerce in Beijing: "hopes Japan will make sure the measures don't hinder the economic and trade development between the two countries." The two nations have enjoyed a cordial semiconductor-centric trade relationship, skewed more in favor of Japan. Industry watchdogs believe that Chinese manufacturers have generated significant demand for Japanese-made production equipment. Japan's Ministry of Economy, Trade and Industry (METI) has updated its documents. According to the latest report, 42 new entities worldwide have been added to a list of: "foreign companies and organizations that would be subject to export oversight on any dual-use items. The additions come into effect on Feb. 5, it said. A total of around 110 Chinese companies, research institutions and other entities are on the list."

Today, Bloomberg reported that the US government under Trump administration is probing whether Chinese AI company DeepSeek circumvented export restrictions to acquire advanced NVIDIA GPUs through Singaporean intermediaries. The investigation follows concerns that DeepSeek's AI model, R1—reportedly rivaling leading systems from OpenAI and Google—may have been trained using restricted hardware that is blocked from exporting to China. Singapore's role in NVIDIA's global sales has surged, with the nation accounting for 22% of the chipmaker's revenue in Q3 FY2025, up from 9% in Q3 FY2023. This spike coincides with tightened US export controls on AI chips to China, prompting speculation that Singapore serves as a pipe for Chinese firms to access high-end GPUs like the H100, which cannot be sold directly to China.

DeepSeek has not disclosed hardware details for R1 but revealed its earlier V3 model was trained using 2,048 H800 GPUs (2.8 million GPU hours), achieving efficiency surpassing Meta's Llama 3, which required 30.8 million GPU hours. Analysts suggest R1's performance implies even more powerful infrastructure, potentially involving restricted chips. US authorities, including the White House and FBI, are examining whether third parties in Singapore facilitated the transfer of controlled GPUs to DeepSeek. A well-known semiconductor analyst firm, SemiAnalysis, believes that DeepSeek acquired around 50,000 NVIDIA Hopper GPUs, which includes a mix of H100, H800, and H20. NVIDIA clarified that its reported Singapore revenue reflects "bill to" customer locations, not final destinations, stating most products are routed to the US or Western markets.

ChangXin Memory Technology (CXMT) is playing catch up with competitors based in neighboring nations—as reported earlier today, this Chinese memory module manufacturer is making advancements in the field of commercial DDR5 products. Reports from the region suggest that a well-known motherboard manufacturer has embraced some of ChangXin new portfolio. According to a recent ITHome article, MSI has implemented a multitude of optimizations for new-ish Intel "Arrow Lake" Core Ultra Series 2 CPU-oriented motherboards—notably the MAG B860 TOMAHAWK WIFI, MAG B860M MORTAR WIFI, and PRO B860M-A WIFI models.

The publication provided evidence of MSI enabling transfer rates of 6800 MT/s across four engaged modules—a welcome step up from the usual twin-stick setup. ITHome also points out that the motherboard specialist has prepared a new BIOS version for B860 chipset mainboards: "specially optimized for the recently popular DDR5 domestic memory particles of ChangXin Storage, which undoubtedly gives the gaming community (in China) more choices." MSI appears to be the first major international vendor to enable support for CXMT's DDR5 modules—thus, representing a significant achievement for the Chinese memory industry.

Earlier in the month, AMD introduced multiple Ryzen 200 series CPU models—these mainstream laptop and desktop offerings are based on well established architectures. Team Red is once again treading familiar ground with its selection of "Hawk Point" hardware—featuring "Zen 4" CPU cores, as well as (RDNA 3) Radeon 700M series integrated graphics solutions. AMD's latest mild re-branding of existing designs—notably the Ryzen 8000 mobile APU series—are expected to arrive within the second quarter of 2025. Several of these "new" products will sport NPUs, but marketing/press material is surprisingly free of all things AI-related.

Yesterday, a Lenovo China product manager revealed that the entire Ryzen 200 product stack will adopt the FP8 chip package—formerly exclusive to the region-exclusive Ryzen 7 8745H APU. The Beijing-based company representative disclosed—via a Weibo blog post—that the incoming Ryzen 200 processor series will not use the "FP7 and FP7r2" formats. Additionally, they also introduced a previously unannounced SKU: "the FP8 package version of Ryzen 7 8745H will be launched in 2025 as Ryzen 7 255H." This model is likely another Chinese exclusive—the Weibo comment section provides further evidence of new Lenovo ThinkBook products being based on this Ryzen 7 255H. AMD's official listings have quietly confirmed that their Ryzen 200 series will be FP8-exclusive. Team Green is likely making matters easier for itself, by streamlining its packaging options for this affordable-tier of laptop-oriented chips.

Intel tends to release exclusive desktop processor SKUs for the Chinese market, which tend to come with core configurations and other specs not found in regular SKUs available in the West. The latest such SKU is the Core Ultra 5 230F "Arrow Lake." This chip comes with a unique IHS (integrated heat spreader) design that sees the bulge over the die contracted to a corner. It still has the metal flanks that receive pressure from the socket's retention module. The chip is positioned a notch above the Core Ultra 5 225F, but with an identical core configuration of 6P+4E.

The China-exclusive Core Ultra 5 230F comes with clock speeds of 3.40 GHz base and 5.00 GHz boost for the P-cores, while its E-cores tick at 2.90 GHz base, and 4.40 GHz boost. In comparison, the generally available Core Ultra 5 225F does 3.30 GHz base and 4.90 GHz boost for the P-cores, and 2.70 GHz base with 4.40 GHz boost for the E-cores. Both the 230F and 225F lack integrated graphics. As non-K SKUs, both lack unlocked base frequency multipliers. Each of the six "Lion Cove" P-cores comes with 3 MB of dedicated L2 cache; while the sole "Skymont" E-core cluster shares 4 MB of L2 cache among its cores. The P-cores and E-core cluster share 20 MB of L3 cache. What's with the oddly shaped IHS? We have a theory. It's likely that the chip underneath has the same compact Compute tile as the "Arrow Lake-H" mobile processor, which physically only has 6P+8E cores and 24 MB of L3 cache. The modular nature of "Arrow Lake" also makes it possible for the Graphics tile to be physically absent on F-SKUs. We can't wait for someone to delid this chip.

A few days ago, we reported on the US reading a massive 500 billion US Dollar package called "Stargate Project" to build AI infrastructure on American soil. However, China is also planning to stay close behind, or even overlap the US in some areas, with a one trillion Yuan "AI Industry Development Action Plan". Translating into around 138 billion US Dollars at the time of writing, the Chinese AI plan is similar to the US Stargate project: develop AI infrastructure through data center expansion and AI accelerator scale-up. Unlike the Stargate project, led by private initiatives and OpenAI at the helm, the Chinese AI Industry Development Action Plan is an entirely state-sponsored initiative that will fund firms like Baidu, ByteDance, Alibaba, and DeepSeek with additional AI accelerators (or sanction-abiding GPUs) to create more advanced AI systems.

Over the past few days, DeepSeek, a branch of a Chinese hedge fund, has not only open-sourced its R1 reasoning model but made it entirely free to use for everyone. This has challenged moat of Western competitors like OpenAI, pushing its CEO Sam Altman to offer an O3-mini reasoning model for up to 100 queries per day for the ChatGPT Plus user tier. Not only did DeepSeek provide a model equally intelligent to OpenAI's best, but it also offered it completely free. This has stirred up the tech community quite a bit and showed that Chinese companies are not much behind Western competitors. With this AI action plan, the Chinese government wants to push its domestic AI makers even further ahead and allow them to overtake cutting-edge model development potentially. Of course, getting GPUs for these projects remains an intricate task, but with export control loopholes and domestic AI accelerator development, the AI arms race is picking up quite a bit.

Yes, the title is correct. One of NVIDIA's GeForce RTX 5090D "China" edition GPUs, not the regular RTX 5090, managed to run at 3.4 GHz under liquid nitrogen. With a staggering 575 W default TDP, Tony Yu, ASUS China's general manager, has performed physical modifications that allow the card to run up to 1000 W TDP. The RTX 5090D is a China-exclusive variant with virtually no difference from the regular RTX 5090, just limited general AI capability due to US export regulations. ASUS China used its top-end Astral OC variant for this stunt, which, as we proved in our review of the regular ASUS RTX 5090 Astral OC, has some pretty good chip binning, allowing the card to reach the highest overclock. We pushed the regular RTX 5090 Astral OC GPU on air to 3086 MHz, a +277 MHz over the stock boost setting. However, the RTX 5090D equivalent under LN2 manages to reach 3,390 MHz at peak loads, which is a +581 MHz difference.

For memory, the overclock is equally impressive with 34 Gbps. Regarding performance, the LN2-overclocked RTX 5090D surpassed stock performance by approximately 16%. During benchmark tests, the GPU outperformed multiple previous-generation graphics cards, including a dual RTX 3090 Ti configuration in Port Royal and a quad GTX 1080 Ti setup in Fire Strike. Power consumption figures indicate that 1,760 W was used in total for a rig with ASUS ROG Astral RTX 5090D, which is paired with the Ryzen 7 9800X3D on the ASUS ROG X870E Hero motherboard. This roughly yields a 1,000 W power consumption by the card, which has seen its PCB get physical modifications to output such high power.

HXL/9550pro and HardwareLuxx's Andreas Schilling shared evidence of the ASUS PCIe Q-Release Slim—the manufacturer's latest ejection mechanism—causing damage to graphics card PCIe connectors. Recent feedback suggests that repeated usage can scrape or grind off material present on a card's interface. HXL gathered critiques from multiple sources (owners of Intel 800 and AMD 800 series boards), and linked a relevant Bilibili video. The footage presents a damaged GALAX RTX 4070 Ti HOF OC LAB model, following sixty quick release cycles—paired with a ROG Strix B850-A Gaming Wi-Fi S motherboard. Tony Wu, ASUS China's general manager, eventually weighed in on community discussions—stating that he will investigate this matter and report back with his findings.

Schilling expressed similar frustrations—his chosen platform is producing unwanted results: "I'm not happy with the solution either. We use the Strix X870E-E Gaming for testing the graphics cards. So I have had to remove graphics cards from the slot a few dozen times. This didn't always go smoothly and very often the (GeForce RTX 5090) card got stuck in the slot. First damage visible." Press outlets have picked up on the recent surge in Q-Release Slim user feedback—several publications have gathered additional examples of the new mechanism inflicting damage on a variety of graphics card models.