AMD's December 2023 launch of new Instinct series accelerators has generated a lot of tech news buzz and excitement within the financial world, but not many folks are privy to Team Red's MSRP for the CDNA 3.0 powered MI300X and MI300A models. A Citi report has pulled back the curtain, albeit with "projected" figures—an inside source claims that Microsoft has purchased the Instinct MI300X 192 GB model for ~$10,000 a piece. North American enterprise customers appear to have taken delivery of the latest MI300 products around mid-January time—inevitably, top secret information has leaked out to news investigators. SeekingAlpha's article (based on Citi's findings) alleges that the Microsoft data center division is AMD's top buyer of MI300X hardware—GPT-4 is reportedly up and running on these brand new accelerators.

The leakers claim that businesses further down the (AI and HPC) food chain are having to shell out $15,000 per MI300X unit, but this is a bargain when compared to NVIDIA's closest competing package—the venerable H100 SXM5 80 GB professional card. Team Green, similarly, does not reveal its enterprise pricing to the wider public—Tom's Hardware has kept tabs on H100 insider info and market leaks: "over the recent quarters, we have seen NVIDIA's H100 80 GB HBM2E add-in-card available for $30,000, $40,000, and even much more at eBay. Meanwhile, the more powerful H100 80 GB SXM with 80 GB of HBM3 memory tends to cost more than an H100 80 GB AIB." Citi's projection has Team Green charging up to four times more for its H100 product, when compared to Team Red MI300X pricing. NVIDIA's dominant AI GPU market position could be challenged by cheaper yet still very performant alternatives—additionally chip shortages have caused Jensen & Co. to step outside their comfort zone. Tom's Hardware reached out to AMD for comment on the Citi pricing claims—a company representative declined this invitation.

An Open Image Denoise 2.2 release candidate was released earlier today—as discovered by Phoronix's founder and principal writer; Michael Larabel. Intel's dedicated website has not been updated with any new documentation or changelogs (at the time of writing), but a GitHub release page shows all of the crucial information. Team Blue's open-source oneAPI has been kept up-to-date with the latest technologies—not only limited to Intel's stable of Xe-LP, Xe-HPG and Xe-HPC components—the Phonorix article highlights updated support on competing platforms. The v2.2 preview adds support for Meteor Lake's integrated Arc graphics solution, and additional "denoising quality enhancements and other improvements."

Non-Intel platform improvements include updates for Apple's M-series chipsets, AArch64 processors, and NVIDIA CUDA. OIDn 2.2-rc: "adds Metal device support for Apple Silicon GPUs on recent versions of macOS. OIDn has already been supporting ARM64/AArch64 for Apple Silicon CPUs while now Open Image Denoise has extended that AArch64 support to work on Windows and Linux too. There is better performance in general for Open Image Denoise on CPUs with this forthcoming release." The changelog also highlights a general improvement performance across processors, and a fix that resolves a crash incident: "when releasing a buffer after releasing the device."

Palit Microsystems Ltd., a leading graphics card manufacturer, proudly announces the NVIDIA GeForce RTX 3050 6 GB KalmX and StormX Series graphics cards. The GeForce RTX 3050 6 GB GPU is built with the powerful graphics performance of the NVIDIA Ampere architecture. It offers dedicated 2nd gen RT Cores and 3rd gen Tensor Cores, new streaming multiprocessors, and high-speed G6 memory to tackle the latest games.

GeForce RTX 3050 6 GB KalmX: Passive Cooling. Silent Gaming
Introducing the Palit GeForce RTX 3050 KalmX, where silence meets performance in perfect harmony. The KalmX series, renowned for its ingenious fan-less design, redefines your gaming experience. With its passive cooling system, this graphics card operates silently, making it ideal for both gaming and multimedia applications. Available on shelves today—2nd February 2024.

Aetina, a leading global Edge AI solution provider, announces the release of its new embedded MXM GPU series utilizing the NVIDIA Ada Lovelace architecture - MX2000A-VP, MX3500A-SP, and MX5000A-WP. Designed for real-time ray tracing and AI-based neural graphics, this series significantly enhances GPU performance, delivering outstanding gaming and creative, professional graphics, AI, and compute performance. It provides the ultimate AI processing and computing capabilities for applications in smart healthcare, autonomous machines, smart manufacturing, and commercial gaming.

The global GPU (graphics processing unit) market is expected to achieve a 34.4% compound annual growth rate from 2023 to 2028, with advancements in the artificial intelligence (AI) industry being a key driver of this growth. As the trend of AI applications expands from the cloud to edge devices, many businesses are seeking to maximize AI computing performance within minimal devices due to space constraints in deployment environments. Aetina's latest embedded MXM modules - MX2000A-VP, MX3500A-SP, and MX5000A-WP, adopting the NVIDIA Ada Lovelace architecture, not only make significant breakthroughs in performance and energy efficiency but also enhance the performance of ray tracing and AI-based neural graphics. The modules, with their compact design, efficiently save space, thereby opening up more possibilities for edge AI devices.

GIGABYTE TECHNOLOGY Co. Ltd, a leading manufacturer of premium gaming hardware, today launches the GeForce RTX 40 EAGLE OC ICE series graphics cards powered by NVIDIA ADA Lovelace architecture. The latest EAGLE OC ICE series presents a white iteration of the well-received EAGLE OC graphics card series. With exceptional performance catering to the diverse requirements of gamers, creators, and AI developers, it features an extensive integration of white materials in its design. This introduces an alternative choice for gamers who appreciate white-themed setups.

GIGABYTE has introduced four models in the GeForce RTX 40 EAGLE OC ICE series, corresponding to the GeForce RTX 4070 Ti SUPER, GeForce RTX 4070 SUPER, GeForce RTX 4060 Ti, and GeForce RTX 4060 GPUs. The EAGLE OC ICE series features a brand-new white exterior design inspired by space technology and incorporates futuristic design elements. The graphics cards come with white covers and backplates, complemented by cosmic-themed graphics, symbols, and geometric shapes, providing a unique and personalized appearance for white-themed desktop setups.

GIGABYTE TECHNOLOGY Co. Ltd, a leading manufacturer of premium gaming hardware, today announced new GeForce RTX 3050 6G graphics cards. The GeForce RTX 3050 EAGLE OC 6G and GeForce RTX 3050 OC Low Profile 6G graphics cards are powered by Ampere- NVIDIA's 2nd gen RTX architecture, featuring dedicated RT cores, AI Tensor cores, and fast GDDR6 graphics memory. They deliver the most realistic ray tracing graphics and advanced AI features with DLSS. In addition to providing gamers with a visually stunning and smooth 1080p gaming experience, creators can also indulge in the joy of creative expression.

The GIGABYTE GeForce RTX 3050 EAGLE OC 6G graphics card incorporates futuristic design elements, presenting distinctive and personalized features through cosmic-themed graphics, symbols, and geometric shapes. The GIGABYTE GeForce RTX 3050 EAGLE OC 6G is equipped with the GIGABYTE WINDFORCE cooling system. Featuring unique blade fans, alternate spinning, and fan stop functionalities, this cooling system efficiently lowers the working temperature of the graphics card, ensuring stable performance even under high-demand operations.

NVIDIA's H800 AI GPU was rolled out last year to appease the Sanction Gods—but later on, the US Government deemed the cutdown "Hopper" part to be far too potent for Team Green's Chinese enterprise customers. Last October, newly amended export conditions banned sales of the H800, as well as the slightly older (plus similarly gimped) A800 "Ampere" GPU in the region. NVIDIA's engineering team returned to the drawing board, and developed a new range of compliantly weakened products. An exclusive Reuters report suggests that Team Green is taking pre-orders for a refreshed "Hopper" GPU—the latest China-specific flagship is called "HGX H20." NVIDIA web presences have not been updated with this new model, as well as Ada Lovelace-based L20 PCIe and L2 PCIe GPUs. Huawei's competing Ascend 910B is said to be slightly more performant in "some areas"—when compared to the H20—according to insiders within the distribution network.

The leakers reckon that NVIDIA's mainland distributors will be selling H20 models within a price range of $12,000 - $15,000—Huawei's locally developed Ascend 910B is priced at 120,000 RMB (~$16,900). One Reuters source stated that: "some distributors have started advertising the (NVIDIA H20) chips with a significant markup to the lower end of that range at about 110,000 yuan ($15,320). The report suggests that NVIDIA refused to comment on this situation. Another insider claimed that: "distributors are offering H20 servers, which are pre-configured with eight of the AI chips, for 1.4 million yuan. By comparison, servers that used eight of the H800 chips were sold at around 2 million yuan when they were launched a year ago." Small batches of H20 products are expected to reach important clients within the first quarter of 2024, followed by a wider release in Q2. It is believed that mass production will begin around Spring time.

PC power supply manufacturer Seasonic has introduced a new limited-edition variant of its Vertex 1000 W 80+ Gold certified modular power supply featuring an eye-catching design inspired by cherry blossoms. Dubbed the "Vertex Sakura," this specially-themed PSU sports a clean white color scheme with pink Sakura flower graphics and a unique textured paint finish. The Vertex Sakura caters specifically to PC enthusiasts and case modders in Japan who want to build a themed rig coordinated around the colors and aesthetic of traditional Japanese cherry blossoms. It offers the same fully modular cabling as the standard Vertex 1000 W model, allowing builders flexibility in cable management.

The PSU also meets the latest ATX 3.0 specifications, including a 12VHPWR connector to support next-generation high-wattage GPUs. Seasonic showcased the Vertex Sakura 1000 W unit at 30,000 yen (around $200 USD) in Japan, where retailers recommend interested buyers preorder as soon as possible due to limited availability. Specific regional launch details outside of Japan are still forthcoming. But the Vertex Sakura's blend of technical prowess and stunning cherry blossom visual flair will likely attract global attention from PC builders looking to add an extra touch of style to their high-end systems.

GIGABYTE Technology, a pioneer in computer hardware, has taken a significant stride in shaping its European business model. Today, GIGABYTE has broadened its e-commerce platform, shop.gigabyte.eu, by integrating enterprise server and server motherboard solutions into its product portfolio. Being at the forefront of computer hardware manufacturing, GIGABYTE recognizes that it is imperative to expand its presence in the EMEA region to maintain its leadership across all markets. With the introduction of our enterprise-level server and motherboard solutions, we are dedicated to delivering a diverse range of high-performance products directly to our B2B clients.

GIGABYTE offers a complete product portfolio that addresses all workloads from the data center to edge including traditional and emerging workloads in HPC and AI to data analytics, 5G/edge, cloud computing, and more. Our enduring partnerships with key technology leaders ensure that our new products are at the forefront of innovation and launch with new partner platforms. Our systems embody performance, security, scalability, and sustainability. Within the e-commerce product portfolio, we offer a selection of models from our Edge, Rack, GPU, and Storage series. Additionally, the platform provides server motherboards for custom integration. The current selection comprises a mix of solutions tailored to online sales. For more complex solutions, customers can get in touch via the integrated contact form.

Taiwanese tech maker Advantech has converted Intel's Arc A370M mobile GPU into a desktop graphics card named the EAI-3100. The new card utilizes the same Arc A370M mobile GPU based on the Xe-LP architecture chip as found in laptops but adds more robust cooling to enable desktop-level performance. Specifically, the EAI-3100 implements a large aluminium heatsink spanning the entire PCB, paired with a 40 mm fan active cooling fan. This allows the card to operate at up to 60 Watt TGP (total graphics power), a noticeable increase over the A370M's 35-50 Watt mobile power range. Despite the improved cooling, Advantech has not factory overclocked the EAI-3100, leaving its graphics clock speed unchanged at 1,550 MHz. The card also retains the same PCIe 4.0 x8 interface as the mobile A370M. An 8-pin PCIe power connector has been added, giving headroom for user overclocking attempts.

In terms of gaming performance, the A370M and, by extension, the EAI-3100 deliver playable frame rates at 1080p resolution with medium image quality settings. The card is comparable to NVIDIA's mobile RTX 3050 GPU. As Intel continues optimizing Arc drivers, more gains are expected. The EAI-3100's dual-slot, 6.61-inch design allows compatibility with most desktop PC cases. Between its small size and the A370M's solid 1080p capabilities, this transformed card represents an interesting budget option for gamers seeking a discounted route to Arc's architecture. Despite the diminutive size, this custom cooling solution keeps the A370M at appropriate temperatures for sustained operation, possibly delivering more than the laptop form factor SKU. For video output, the card features two HDMI 2.0b and two DP 1.4a ports.

DFI, the world's leading brand in embedded motherboards and industrial computers, is targeting the AI application market by launching the embedded system module (SOM) MTH968 equipped with the latest Intel Core Ultra processor. It is the first product integrated with an NPU (Neural Processor Unit) processor, representing the official integration of AI with industrial PCs (IPCs). With the expansion into AI IPC, DFI expects to inject new momentum into the AI edge computing market.

According to the STL Partners report, the potential market value of global edge computing will increase from US$9 billion in 2020 to US$462 billion in 2030, representing a compound annual growth rate (CAGR) of 49%. Therefore, the development of products that utilize the core capabilities of chips to rapidly execute AI edge computing in devices has become a key focus for many major technology companies.

According to a VideoCardz report, NVIDIA is implementing a very last minute time shift with its GeForce RTX 4080 SUPER review program—embargo conditions have been delayed by a day to January 31, which coincides with the official retail launch day. We already know about non-specific sample units reaching reviewers a week (or more) in advance of Team Green's embargo date—thanks to various graphical benchmarks appearing prematurely on the Geekbench Browser database. VideoCardz states the Founders Edition GeForce RTX 4080 SUPER model was not received in a timely manner by a number of media outlets, thus dismissing rumors about driver issues being a main factor behind the sudden rescheduling. Hardware evaluators have been busy this month with trade event coverage, and spending analytical time with Team Green's previous batches of RTX 40 SUPER cards.

We've known since way back in August 2023, that AMD is rumored to be retreating from the enthusiast graphics segment with its next-generation RDNA 4 graphics architecture, which means that we likely won't see successors to the RX 7900 series squaring off against the upper end of NVIDIA's fastest GeForce RTX "Blackwell" series. What we'll get instead is a product stack closely resembling that of the RX 5000 series RDNA, with its top part providing a highly competitive price-performance mix around the $400-mark. A more recent report by Moore's Law is Dead sheds more light on this part.

Apparently, the top Radeon RX SKU based on the next-gen RDNA4 graphics architecture will offer performance comparable to that of the current RX 7900 XTX, but at less than half its price (around the $400 mark). It is also expected to achieve this performance target using a smaller, simpler silicon, with significantly lower board cost, leading up to its price. What's more, there could be energy efficiency gains made from the switch to a newer 4 nm-class foundry node and the RDNA4 architecture itself; which could achieve its performance target using fewer numbers of compute units than the RX 7900 XTX with its 96.

Graid Technology, an industry trailblazer in GPU-based RAID for NVMe, proudly announces the groundbreaking release of SupremeRAID SR-1001. This innovative GPU-based RAID solution is designed to maximize NVMe SSD performance while eliminating CPU cycle consumption and avoiding throughput bottlenecks. Utilizing patented out-of-path RAID protection technology, data travels directly from the CPU to the NVMe SSDs, ensuring unmatched flexibility, unprecedented performance, and overall superior value.

NVMe SSDs, with their high-speed performance and low latency, significantly enhance tasks across CAD, video editing, IoT, and gaming. Faster loading times, improved rendering, quick file transfers, smooth playback, efficient data processing, and reduced latency contribute to overall superior performance. But traditional RAID methods introduce bottlenecks, limiting the performance of NVMe SSDs in critical applications.

MSI presented a massive table of GeForce RTX 40 SUPER series custom design graphics cards at CES 2024—TPU spent a lot of time photographing and documenting everything GPU-related at the tech company's booth. The kind-of mysterious GeForce RTX 4080 SUPER 16G EXPERT model seemed to get a lot of attention from online PC hardware communities. The brand new and very substantial EXPERT shroud design integrates a Zero Frozr cooling solution, but folks were quick to link its aesthetic (and vapor chamber setup) to NVIDIA's dual fan Founders Edition cooling solution. MSI has not yet published a dedicated GeForce RTX 4080 SUPER 16G EXPERT product page, while all of the other models from CES are uploaded and active.

A tipster on social media has posted a screenshot of the 16G EXPERT's specification sheet—wxnod has uncovered alleged factory settings ahead of review and launch day embargos, although pricing is still unknown at the time of writing. The leaked core figures include a boost clock of 2610 MHz in GAMING and SILENT modes (60 MHz above reference), while the MSI Center software suite can activate an Extreme Performance mode: 2625 MHz. These figures align with the SUPER SUPRIM model's core clock specs—the SUPRIM X sits above everything else as the fastest card in MSI's RTX 4080 SUPER stable. MSI's official introduction stated that the 16G EXPERT: "features a push-pull airflow design for enhanced cooling efficiency. The enclosure is constructed with aluminium Die-Casting for structural strength, while Core Pipe and a Vapor Chamber work to efficiently dissipate heat. Lastly, a patented fan design provides quiet yet reliable airflow." We hope to see the MSI GeForce RTX 4080 SUPER 16G EXPERT's addition to the TPU GPU database.

LaminiAI appeared to be one of the first customers to receive an initial shipment of AMD's Instinct MI300X accelerators, as disclosed by their CEO posting about functioning hardware on social media late last week. A recent Taiwan Economic Daily article states that the "MI300X is rumored to have begun supply"—we are not sure about why they have adopted a semi-secretive tone in their news piece, but a couple of anonymous sources are cited. A person familiar with supply chains in Taiwan divulged that: "(they have) been receiving AMD MI300X chips one after another...due to the huge shortage of NVIDIA AI chips, the arrival of new AMD products is really a timely rainfall." Favorable industry analysis (from earlier this month) has placed Team Red in a position of strength, due to growing interest in their very performant flagship AI accelerator.

The secrecy seems to lie in Team Red's negotiation strategies in Taiwan—the news piece alleges that big manufacturers in the region have been courted. AMD has been aggressive in a push to: "cooperate and seize AI business opportunities, with GIGABYTE taking the lead and attracting the most attention. Not only was GIGABYTE the first to obtain a partnership with AMD's MI300A chip, which had previously been mass-produced, but GIGABYTE was also one of the few Taiwanese manufacturers included in AMD's first batch of MI300X partners." GIGABYTE is expected to release two new "G593" product lines of server hardware later this year, based on combinations of AMD's Instinct MI300X accelerator and EPYC 9004 series processors.

OpenAI is reported to be initiating R&D on a proprietary AI processing solution—the research organization's CEO, Sam Altman, has commented on the in-efficient operation of datacenters running NVIDIA H100 and A100 GPUs. He foresees a future scenario where his company becomes less reliant on Team Green's off-the-shelf AI-crunchers, with a deployment of bespoke AI processors. A short Reuters interview also underlined Altman's desire to find alternatives sources of power: "It motivates us to go invest more in (nuclear) fusion." The growth of artificial intelligence industries has put an unprecedented strain on energy providers, so tech firms could be semi-forced into seeking out frugal enterprise hardware.

The Financial Times has followed up on last week's Bloomberg report of OpenAI courting investment partners in the Middle East. FT's news piece alleges that Altman is in talks with billionaire businessman Sheikh Tahnoon bin Zayed al-Nahyan, a very well connected member of the United Arab Emirates Royal Family. OpenAI's leadership is reportedly negotiating with TSMC—The Financial Times alleges that Taiwan's top chip foundry is an ideal manufacturing partner. This revelation contradicts Bloomberg's recent reports of a potential custom OpenAI AI chip venture involving purpose-built manufacturing facilities. The whole project is said to be at an early stage of development, so Altman and his colleagues are most likely exploring a variety of options.

Intel has released the latest version of its Arc GPU Graphics Drivers, version 101.5234 WHQL. This appears to be a major update as it brings Game On Driver support for Arc A-Series Graphics for plenty of new games including Enshrouded, Suicide Squad: Kill the Justice League, Like a Dragon: Infinite Wealth, Tekken 8, and the recently released and quiet popular Palworld, as well as Game On Driver support on Intel Core Ultra CPUs with Intel Arc Graphics for Like a Dragon: Infinite Wealth, Tekken 8, and Palworld.

The new drivers also bring plenty of game performance improvements for a rather extensive list of games for both Arc A-series Graphics and Intel Core Ultra CPUs with Arc Graphics. These improvements range anywhere from 4 percent, up to 268 percent, and include both DirectX 11 and DirectX 12 titles. You can check out the full list below. Intel also fixed a couple of issues, including Alan Wake 2 white corruption issue on reflective surfaces, Sons of the Forest corruption on the item text issue, and issues with Intel Smooth Sync in some DirectX 11 titles. It also fixes issues with on Intel Core Ultra CPUs with Arc Graphics where The Talos Principle may experience an application crash with certain upscaling presets, Call of Duty Modern Warfare III application crash issues, and Blackmagic Fusion application crash during render operations.

DOWNLOAD: Intel Arc GPU Graphics Drivers 101.5234 WHQL

We are well aware that NVIDIA GeForce RTX 4080 SUPER graphics cards are next up on the review table (January 31)—TPU's W1zzard has so far toiled away on getting his evaluations published on time for options further down the Ada Lovelace SUPER food chain. This process was interrupted briefly by the appearance of custom Radeon RX 7600 XT models, but today's attention soon returned to another batch of GeForce RTX 4070 Ti SUPER cards. Reviewers are already toying around with driver-enabled GeForce RTX 4080 SUPER sample units—under strict confidentiality conditions—but the occasional leak is expected to happen. The appropriately named Benchleaks social media account has kept track of emerging test results.

The Geekbench Browser database was updated earlier today with premature GeForce RTX 4080 SUPER GPU test results—one entry highlighted by Benchleaks provides a quick look at the card's prowess in three of Geekbench 5.1's graphics API trials: Vulkan, CUDA and OpenCL. VideoCardz points out that all of the scores could be fundamentally flawed; in particular the Vulkan result of 100378 points—the regular (non-SUPER) GeForce RTX 4080 GPU can achieve almost double that figure in Geekbench 6. The SUPER's other results included a Geekbench 5 CUDA score of 309554, and an achievement of 264806 points in OpenCL. A late morning entrant looks to be hitting the right mark—an ASUS testbed (PRIME Z790-A WIFI + Intel Core i9-13900KF) managed to score 210551 points in Geekbench 6.2.2 Vulkan.

Samsung, SK hynix, and Micron are considered to be the top manufacturing sources of High Bandwidth Memory (HBM)—the HBM3 and HBM3E standards are becoming increasingly in demand, due to a widespread deployment of GPUs and accelerators by generative AI companies. Taiwan's Commercial Times proposes that there is an ongoing shortage of HBM components—but this presents a growth opportunity for smaller manufacturers in the region. Naturally, the big name producers are expected to dive in head first with the development of next generation models. The aforementioned financial news article cites research conducted by the Gartner group—they predict that the HBM market will hit an all-time high of $4.976 billion (USD) by 2025.

This estimate is almost double that of projected revenues (just over $2 billion) generated by the HBM market in 2023—the explosive growth of generative AI applications has "boosted" demand for the most performant memory standards. The Commercial Times report states that SK Hynix is the current HBM3E leader, with Micron and Samsung trailing behind—industry experts believe that stragglers will need to "expand HBM production capacity" in order to stay competitive. SK Hynix has shacked up with NVIDIA—the GH200 Grace Hopper platform was unveiled last summer; outfitted with the South Korean firm's HBM3e parts. In a similar timeframe, Samsung was named as AMD's preferred supplier of HBM3 packages—as featured within the recently launched Instinct MI300X accelerator. NVIDIA's HBM3E deal with SK Hynix is believed to extend to the internal makeup of Blackwell GB100 data-center GPUs. The HBM4 memory standard is expected to be the next major battleground for the industry's hardest hitters.

NVIDIA's intrepid CEO, Jensen Huang, has spent a fair chunk of January travelling around China—news outlets believe that Team Green's leader has conducted business meetings with very important clients in the region. Insiders proposed that his low-profile business trip included visits to NVIDIA operations in Shenzhen, Shanghai and Beijing. The latest updates allege that a stopover in Taiwan was also planned, following the conclusion of Mainland activities. Photos from an NVIDIA Chinese new year celebratory event have been spreading across the internet lately—many were surprised to see Huang appear on-stage in Shanghai and quickly dispense with his trademark black leather jacket. He swapped into a colorful "Year of the Wood Dragon" sleeveless shirt for a traditional dance routine.

It was not all fun and games during Huang's first trip to China in four years—inside sources have informed the Wall Street Journey about growing unrest within the nation's top ranked Cloud AI tech firms. Anonymous informants allege that leadership, at Alibaba Group and Tencent, are not happy with NVIDIA's selection of compromised enterprise GPUs—it is posited that NVIDIA's President has spent time convincing key clients to not adopt natively-developed solutions (unaffected by US Sanctions). The short hop over to Taiwan is reported not to be for R&R purposes—insiders had Huang's visiting key supply partners; TSMC and Wistron. Industry experts think that these meetings are linked to NVIDIA's upcoming "Blackwell" B100 AI GPU, and "supercharged" H200 "Hopper" accelerator. It is too early for the rumor mill to start speculation about nerfed versions of NVIDIA's 2024 enterprise products reaching Chinese shores, but Jensen Huang is seemingly ready to hold diplomatic talks with all sides.

LaminiAI appears to be one of AMD's first customers to receive a bulk order of Instinct MI300X GPUs—late last week, Sharon Zhou (CEO and co-founder) posted about the "next batch of LaminiAI LLM Pods" up and running with Team Red's cutting-edge CDNA 3 series accelerators inside. Her short post on social media stated: "rocm-smi...like freshly baked bread, 8x MI300X is online—if you're building on open LLMs and you're blocked on compute, lmk. Everyone should have access to this wizard technology called LLMs."

An attached screenshot of a ROCm System Management Interface (ROCm SMI) session showcases an individual Pod configuration sporting eight Instinct MI300X GPUs. According to official blog entries, LaminiAI has utilized bog-standard MI300 accelerators since 2023, so it is not surprising to see their partnership continue to grow with AMD. Industry predictions have the Instinct MI300X and MI300A models placed as great alternatives to NVIDIA's dominant H100 "Hopper" series—AMD stock is climbing due to encouraging financial analyst estimations.

According to the recent listing on Goofish, discovered by VideoCardz, Chinese companies have begun selling mobile versions of NVIDIA's latest RTX 40-series GPUs as desktop graphics cards. Initially designed for gaming laptops, the GeForce RTX 4080M and RTX 4090M are now being marketed in China as more affordable alternatives to their official desktop counterparts. This development is no surprise to industry observers who recall similar adaptations with the RTX 20 and 30 series. These companies are leveraging the lower cost of mobile GPUs, combined with budget cooling solutions and simpler PCB designs, to offer more affordable desktop GPU options. The mobile GPUs, which are capped at a power consumption of 175 Watts, are being repurposed without official sanction, with NVIDIA seemingly disregarding this practice. Despite the lack of official endorsement, these modified GPUs are finding their way into the market, providing gamers a cost-effective alternative to the more expensive desktop versions.

While not officially supported by NVIDIA, these cards utilize the mobile GPU dies paired with custom cooling solutions and PCBs to work in desktop PCs. According to reports, the RTX 4080M desktop variant offers 7424 CUDA cores and 12 GB GDDR6 memory, representing a 24% reduction in cores and 4 GB less memory versus the desktop RTX 4080. The desktop RTX 4090M is even more cut-down, with 9728 cores and 16 GB memory—a 40% drop in cores and 8 GB less memory than the flagship RTX 4090 desktop card. Pricing falls between $420 and $560 for the RTX 4080M and exceeds that of even the desktop RTX 4090 for the 4090M variant. Performance and longevity still need to be determined for these unofficial cards. While they present a cheaper RTX 40-series option for Chinese gamers, the reduced specifications come with tradeoffs. Still, their availability indicates the ongoing demand for next-gen GPUs and the lengths some vendors go to to meet that demand.

ASRock showcased customized Radeon RX 7600 XT 16 GB GPU offerings at CES 2024—only a couple days after AMD's official unveiling of its expanded lower mid-range RDNA 3 line. ASRock was among a select few Team Red board partners with finalized units (based on Navi 33 XT) on display—it seems that the Taiwanese manufacturer is preparing for a retail launch of its Radeon RX 7600 XT Steel Legend 16 GB OC and Challenger 16 GB OC graphics card models. ASRock's website has been updated with product pages for the latest Radeon RX 7000-series entries, but press material for an imminent product launch has not been published (at the time of writing).

ASRock's mid-tier triple-fan Steel Legend and entry-level dual-fan Challenger designs are a familiar sight across the company's Radeon RX 7000 and 6000 product lines—last September, customized Radeon RX 7800 XT and Radeon RX 7700 XT models were unveiled as sporting these shrouds, along with higher-end Phantom Gaming OC options. A slightly overclocked Radeon RX 7600 XT GPU is not expected to be a heat producing monster, so expensive cooling solutions are not a necessity for a cost-conscious audience—likely targeting a decent level of 1080p gaming performance. The ASRock Radeon RX 7600 XT Challenger 16 GB OC model is expected to launch at an MSRP of $329 (AMD's official guide SEP), while the fancier Steel Legend OC is believed to be only marginally more expensive.

Mark Zuckerberg has shared some interesting insights about Meta's AI infrastructure buildout, which is on track to include an astonishing number of NVIDIA H100 Tensor GPUs. In the post on Instagram, Meta's CEO has noted the following: "We're currently training our next-gen model Llama 3, and we're building massive compute infrastructure to support our future roadmap, including 350k H100s by the end of this year -- and overall almost 600k H100s equivalents of compute if you include other GPUs." That means that the company will enhance its AI infrastructure with 350,000 H100 GPUs on top of the existing GPUs, which is equivalent to 250,000 H100 in terms of computing power, for a total of 600,000 H100-equivalent GPUs.

The raw number of GPUs installed comes at a steep price. With the average selling price of H100 GPU nearing 30,000 US dollars, Meta's investment will settle the company back around $10.5 billion. Other GPUs should be in the infrastructure, but most will comprise the NVIDIA Hopper family. Additionally, Meta is currently training the LLama 3 AI model, which will be much more capable than the existing LLama 2 family and will include better reasoning, coding, and math-solving capabilities. These models will be open-source. Later down the pipeline, as the artificial general intelligence (AGI) comes into play, Zuckerberg has noted that "Our long term vision is to build general intelligence, open source it responsibly, and make it widely available so everyone can benefit." So, expect to see these models in the GitHub repositories in the future.