During GTC 2025, NVIDIA unveiled the professional (PRO) side of its "Blackwell" GPU line—headlined by a monstrous GDDR7 96 GB option, that unleashes the full potential of their GB202 die. Industry watchdogs anticipated sky-high pricing, as befits such a potent specification sheet/feature set. As reported by VideoCardz over the past weekend, a North American enterprise PC hardware store—Connection—has populated its webshop with several of Team Green's brand-new RTX PRO Blackwell Series SKUs. The publication received tip-offs from a portion of its readership; including some well-heeled individuals who have already claimed pre-orders. Starting off, the investigation highlighted upper crust offerings: "the flagship model, called the RTX PRO 6000 with 96 GB of VRAM, will launch at $8435 (bulk) to $8565 (box), and this price seemingly applies to both models: the Workstation Edition and a sub-variant called Max-Q. Both are equipped with the same specs, but the latter is capped at 300 W TDP while retaining 88% of the Al performance, claimed NVIDIA."

Connection has removed its RTX PRO 6000 Blackwell and RTX PRO 6000 Blackwell Max-Q product pages, but the rest of Team Green's professional stack is still visible (see relevant screenshot below). The RTX PRO 5000 Blackwell 48 GB card is priced at $4569.24 (or $4439.50 for bulk). The cheapest offering is a $696.54 RTX PRO 2000 Blackwell 8 GB model. Officially, NVIDIA and its main professional series board partner—PNY—only revealed 4500, 5000 and 6000 product tiers. VideoCardz put a spotlight on some of these unannounced options, including: "the RTX 4000 non-SFF version, while this retailer has six listings for such SKUs (two SFF and two non-SFF, both in bulk and box variants). Presumably, this would suggest that NVIDIA may launch a non-SFF version later. However, the company didn't put 'SFF' in the official card's name, so perhaps this information is no longer valid, and there's only one model." According to a GDM/Hermitage AkiHabara Japan press release, a local reseller—Elsa—is preparing NVIDIA RTX PRO 6000 Blackwell Workstation Edition and RTX PRO 6000 Blackwell Max-Q Workstation Edition stock for scheduled release "in May 2025, while the other models are scheduled for release around summer." Additionally, another retailer (ASK Co., Ltd.): "has stated that the price and release date are subject to inquiry."

ASRock Industrial introduces the iEP(F)-9040 Series, a next-generation Robust Edge AIoT Platform designed to deliver high-performance edge AI computing, real-time industrial automation, and scalable expansion for smart manufacturing, intelligent vision systems, autonomous robotics, and edge AI applications. Powered by Intel Core 200S Processors (Bartlett Lake-S), the iEP(F)-9040 Series enables seamless AI inferencing, high-speed data processing, and intelligent edge computing for industrial IoT and AIoT applications. With support for up to 24 cores and 32 threads with enhanced longevity, DDR5 5600 MHz memory up to 96 GB, PCIe Gen 5 expansion, the Series is built to handle complex AI workloads. Designed to thrive in extreme industrial conditions, the series supports up to -40°C~75°C wide operating temperature range and 9 V-36 V DC power input with ignition control, ensuring reliable operation in mission-critical environments. The iEP(F)-9040 Series features rich industrial I/O, supports real-time TSN/TCC, and connectivity options like 5G/4G LTE, Wi-Fi 6E and BT 5.4 for high-speed, low-latency communication. The Series is ideal for various industrial applications such as smart factory, AI-based Inspection, robotics, surveillance, smart transportation, and more, providing a future-ready AIoT solution for next generation edge and industrial intelligence.

Optimized AI and Compute Performance with Intel Core 200S Processors
The iEP(F)-9040 Series harness the power of Intel Core 200S Processors (Bartlett Lake-S) to deliver superior computing performance and AI acceleration with long-term support. With up to 24 (8P+16E) cores, delivering up to 32 threads, the series enables real-time AI inferencing, industrial automation, and intelligent edge computing. Compared to previous generations, Intel Core 200S Processors (Bartlett Lake-S) achieve up to 1.06x faster single-threaded and 1.08x faster multi-thread performance. Supporting dual DDR5 5600 MHz SO-DIMM memory up to 96 GB, the iEP(F)-9040 Series significantly boosts machine vision, AI model deployment, and automation. Integrated Intel UHD Graphics provides up to 4K high-resolution AI-driven imaging, with DisplayPort 1.4a, HDMI 2.0b, and VGA outputs for flexible industrial display configurations.

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

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

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

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

Mini PCs that prioritize performance over anything else are far from hard to find. Most such systems opt for high-end laptop chips instead of desktop ones, which is to be expected considering that there is quite a lot of overlap in terms of thermal expectations between the two segments - in other words, both have to deal with limited thermal headroom. The newly launched K10 mini PC from the well-known brand GMKtec is no exception, and packs the Intel Core i9-13900HK CPU paired with up to 96 GB of memory. Pricing details are under wraps as of this writing, and the system is available for pre-order in China. A global release in not in sight, although GMKtec does like to bring many of its products to the global market.

The 14-core, 20-thread Core i9-13900HK CPU may be over two years old at this point, but is far from a slouch. With performance that is almost neck-and-neck with the Core Ultra 9 185H, the high-end Raptor Lake-H is more than sufficient for the vast majority of intensive workloads that one might throw its way. Dual SODIMM slots are present, which can accommodate up to 96 GB of DDR5-5200 memory, Moreover, the K10 sports a whopping three M.2 2280 slots, which allows the system to pack up to 12 TB of solid-state storage. A pretty decent cooling system is also present, the performance of which, however, can only be established once independent reviews arrive.

The NVIDIA GeForce RTX 5070 Ti graphics cards—built on the NVIDIA Blackwell architecture—are out now, ready to power generative AI content creation and accelerate creative performance. GeForce RTX 5070 Ti GPUs feature fifth-generation Tensor Cores with support for FP4, doubling performance and reducing VRAM requirements to run generative AI models.

In addition, the GPU comes equipped with two ninth-generation encoders and a sixth-generation decoder that add support for the 4:2:2 pro-grade color format and increase encoding quality for HEVC and AV1. This combo accelerates video editing workflows, reducing export times by 8x compared with single encoder GPUs without 4:2:2 support like the GeForce RTX 3090. The GeForce RTX 5070 Ti GPU also includes 16G B of fast GDDR7 memory and 896 GB/sec of total memory bandwidth—a 78% increase over the GeForce RTX 4070 Ti GPU.

The mini PC market has grown substantially in the past few years, in no small part thanks to the commendable performance and efficiency improvements brought to the table by the x86 giants. Gigabyte has unveiled its fair share of such systems, with its latest offering sporting an Intel Core Ultra 5 225H CPU paired with up to a whopping 96 GB of memory. As of right now, we do not have any concrete information regarding when the system will be made available to the public, and at what price. But given the system specifications and Gigabyte's track record, we can expect the system to command an affordable price tag.

As mentioned, the mini PC, dubbed the GB-BRU5-225H, is powered by the Intel Core Ultra 5 225H CPU including four high-performance "Lion Cove" cores, eight high-efficiency "Skymont" cores, as well as two additional "low power" cores for a total of 14 cores and 14 threads. The 12.6 x 3.4 x 11.9 cm mini PC has no room for discrete graphics, and has to rely on the integrated Arc 130T iGPU with eight Xe cores. The system can be equipped with up to 96 GB of DDR5-6400 memory courtesy of its dual SO-DIMM slots. Considering the specifications, the mini PC should be sufficiently potent for the vast majority of moderately intensive workloads thrown its way.

G.Skill International Enterprise Co., Ltd., the world's leading brand of performance overclock memory and PC components, is announcing the release of high-speed, low-latency DDR5-6800 CL32 memory specification with large kit capacity of up to 96 GB (2x 48 GB), as well as DDR5-6400 CL28 32 GB (2x 16 GB) high-performance memory specification. These new memory specifications will be available under the G.SKILL Trident Z5 Royal, Trident Z5 RGB, and Ripjaws M5 RGB series, with Intel XMP 3.0 memory overclock profile support.

As G.SKILL continues to push memory performance to higher limits, an ideal memory solution for overclocking enthusiasts who are looking for both large capacity and high performance is born at DDR5-6800 CL32-42-42 at up to 96 GB (2x 48 GB). This new specification has been validated on both Intel Z790 and Z890 platforms. Please refer to the screenshots below for Memtest validation on the ASUS ROG Maximus Z890 Hero motherboard with the Intel Core Ultra 7 265K desktop processor, and the ASUS ROG Maximus Z790 Dark Hero motherboard with the Intel Core i9-14900K desktop processor.

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

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

G.SKILL International Enterprise Co., Ltd., the world's leading brand of performance overclock memory and PC components, is announcing the release of low latency DDR5-6400 CL30-39-39-102 memory specification in high-capacity kit configurations, up to 96 GB (2x 48 GB) kit capacity. This high performance memory kit specification will be available under the Trident Z5 RGB and Trident Z5 Royal series, with Intel XMP 3.0 memory overclock profile support.

The G.SKILL DDR5-6400 CL30 memory kit at 96 GB (2x 48 GB) combines high-capacity and low-latency to enable higher performance for memory intensive platforms and uses. The screenshot below shows this 96 GB (2x 48 GB) memory kit operating at DDR5-6400 CL30-39-39-102 with the ASUS ROG Maximus Z890 Hero motherboard and Intel Core Ultra 7 265K desktop processor.

ASRock Industrial unveils the 4X4 BOX AI300 Series, a compact powerhouse designed to redefine AI computing with AMD Ryzen AI 300 Series processors. Featuring the Ryzen AI 7 350 and Ryzen AI 5 340 processors, this series delivers up to 50 TOPS of NPU performance, supporting Copilot+ PC experiences and AI-enhanced multitasking. With up to 96 GB dual-channel DDR5 5600 MHz memory, quad display up to 8K, and versatile connectivity, including dual LAN ports, Wi-Fi 6E, five USB ports with two USB4, it combines exceptional AI performance with scalability and efficiency.

Equipped with Ryzen AI as its dedicated AI engine, the 4X4 BOX AI300 Series ensures responsive AI processing paving the way for next generation productivity. Living up to its slogan, "Tiny Box AI Rocks," the 4X4 BOX AI300 Series is engineered for AI PCs, gaming, content creation, tech enthusiasts, prosumers, and AIoT applications.

Today at CES 2025, future business leaders will be wowed by new and powerful AI PCs and solutions from HP Inc. that empower them to collaborate and lead like never before.

Working with Purpose to Drive Impact
As the work landscape evolves, so do expectations for technology. That's why HP meticulously designed its latest lineup of commercial next-gen AI PCs to give professionals the right tools for their unique work experiences. With HP's newest EliteBooks, fast presentation creation, personalized emails, and a built-in recording studio are possible with just a simple click of a button, alongside built-in security and intelligence thanks to Wolf Security.

GMKTec is a relatively well-known brand of mini PCs that ships internationally from its official online store. The brand's latest offering, the K11, is an RGB-fied mini PC that sports adequately powerful internals for almost any task. At its core, the K11 is powered by the Ryzen 9 8945HS "Hawk Point" APU with eight Zen 4 cores and 16 threads. The system comes pre-installed with up to 32 GB of memory and a 2 TB SSD, both of which can be upgraded down the line.

The memory can be upgraded to a whopping 96 GB, whereas the storage can be upgraded to 8 TB via the dual M.2 slots. The integrated Radeon 780M GPU is decently powerful for most non-intensive workloads, but pushing it any further might be futile. For such cases, the OCuLink port will definitely come in handy, offering speedy eGPU connections that far outshine USB-based solutions. The cooling solution appears to be sufficient on paper, allowing the system to draw up to 70 watts.

Minisforum has decided to expand the series of motherboards used in its own mini-PCs with two new models, namely the BD795M (M-ATX) and BD795i SE (Mini-ITX). In principle, these motherboards are almost identical to those in the BD790i series, keeping the included AMD Ryzen 9 7945HX 16 cores processor with frequency of up to 5.4 GHz and integrated AMD Radeon 610M graphics. However, both have some minor physical differences. As for specifications, they offer two M.2 2280 SSD slots (one PCIe 5.0), support for DDR5 SODIMM memory with speeds up to 5200 MT/s, and a maximum capacity of 96 GB. The motherboards offer a full-size PCIe 5.0 x16 slot for users who want to add a dedicated video card.

Unlike the BD790i model, Minisforum decided not to include a CPU cooler for the BD795M model, leaving the cooling solution to the users' choice, a decision that has the collateral effect of a lower price. The motherboards even include an AIO pump connector, so users aren't limited to air cooling. Minisforum confirmed that the CPU has a TDP of 100 W and can also operate at 120 W. Also in the series of minor changes, the BD795M motherboard lacks the SSD heatsink and SSD fan connector but includes an internal antenna adapter.

Framework, the company known for making consumer-friendly, repairable, upgradeable laptops, has officially announced the first expansion bay for the Framework 16, its AMD-powered gaming laptop. The new storage module, which slots into the Expansion Bay, has dual M.2 slots for up to 16 TB total additional storage for the Framework 16. Part of the idea behind the storage expansion seems to be turning what is essentially a gaming laptop into a capable workstation. Crucially, upgrading the storage with the expansion bay requires removing the discrete Radeon 7700s GPU, although the Framework 16 already has dual M.2 slots on the motherboard, so this expansion isn't really intended for gamers, anyway.

One of the major selling points for the Framework 16 was that it offered PCIe expansion via a modular interface, and this is Framework's first real foray into expanding that ecosystem for its largest laptop. In addition to the storage expansion, Framework also announced a new Mystery Box system for its US and Canada Outlets to offload spare parts, like returned modules and components that it doesn't want to relegate to the e-waste pile but also cannot financially justify sorting through and refurbishing. These Mystery Boxes each contain at least three items and come with a warning that reads "Note that these don't come with a warranty and are non-returnable, so only get it if you want random scrap to play with!"

ARBOR Technology has unveiled the boundaries of computing with its latest NUC-sized PC, featuring 34 TOPS of AI computing power driven by Intel Core Ultra processors and Intel Arc Graphics, offering unparalleled performance and efficiency in a compact design.

Experience AI-Accelerated Performance
Experience unparalleled performance and AI acceleration with the latest Intel Core Ultra processors. These cutting-edge chips feature a hybrid architecture that seamlessly blends CPU, GPU, and NPU capabilities, delivering up to a 14% boost in CPU productivity. Immerse yourself in stunning visuals and smooth computing experiences with integrated Intel Arc Graphics. These discrete graphics cards offer exceptional performance and energy efficiency.

The upcoming Strix Point Halo processors from AMD now have a new name - Ryzen AI Max - and come with big promises of impressive power. This rumor, first reported by VideoCardz and originating from Weibo leaker Golden Pig Upgrade, reveals key details about the first three processors in this lineup, along with their specifications.

The leaker claims AMD might roll out a new naming system for these processors branding them as part of the Ryzen AI Max series. These chips will run on the anticipated Strix Halo APU. This series includes three models, with the top-end version boasting up to 16 Zen 5 cores and 40 Compute Units (CUs) for graphics. This setup is expected for the best model contrary to earlier rumors that AMD would drop such a variant. In fact, word has it that at least two of the models in this lineup will come with 40 RDNA 3.5 Compute Units. The leaker also hints that Strix Halo will handle up to 96 GB of video memory suggesting AMD aims to make this processor work with its ROCm (Open Compute Platform) system.

Vecow has launched its TGS-1000 series of mini PCs, featuring stackable expansion options for networking, wireless mobility, and additional capabilities. The Vecow TGS-1000 Series is an ultra-compact, fanless, stackable embedded computer lineup that includes the TGS-1000 and TGS-1500 models, powered by Intel Core Ultra Meteor Lake processors with integrated CPU, GPU, and NPU. This series supports up to 96 GB of DDR5 memory. Optimized for edge AI applications, it delivers up to a 14% boost in CPU performance and enhanced graphics capabilities.

The TGS-1000 Series can drive up to five independent displays through two HDMI and three DisplayPort (DP) ports. It features a variety of I/O connections, including up to five USB 3.0 ports (four Type-A and one Type-C) and a 2.5GbE LAN port with TSN support. Its modular design allows for flexible expansion with options for USB, isolated DIO, COM, LAN, or 4G/LTE, making it suitable for AI, smart retail, office communication, and gaming. The TGS-1500 model adds support for MXM graphics cards up to RTX 5000 Ada.

G.SKILL International Enterprise Co., Ltd., the world's leading brand of performance overclock memory and PC components, is excited to announce an ultra-low latency specification of DDR5-6000 CL28-36-36-96 in 32 GB (2x16GB) and 64 GB (2x32GB) kit capacities, and DDR5-6000 CL28-38-38-96 in 48 GB (2x24GB) and 96 GB (2x48GB) kit capacities under the new Trident Z5 Royal Neo series, designed for compatible AMD AM5 platforms. Including AMD EXPO technology for an easy memory overclock experience in BIOS, this overclock performance memory kit with low timing is the ideal DDR5 memory solution for enthusiasts and overclockers.

Optimization with Memory Timing
For enthusiasts and overclockers, memory timing or latency is a key factor in squeezing performance out of a memory kit. Since memory timing is the delay between specific actions, a lower latency is desired; and performance may be improved by finding the best mix of memory speed and latency. Compared to a standard DDR5 memory speed and latency of DDR5-4800 CL40, this new DDR5-6000 CL28 memory specification aims to deliver a more optimized combination on compatible AMD AM5 platforms.

G.SKILL International Enterprise Co., Ltd., the world's leading brand of performance overclock memory and PC components, is announcing the release of Ripjaws M5 RGB, a high-performance RGB DDR5 memory series designed for the latest DDR5-enabled Intel platforms. At launch, initial specifications will be offered up to DDR5-6400 CL32-39-39-102 96 GB (2x48GB). Available in matte black or matte white aluminium heatspreaders and features customizable RGB lighting, the Ripjaws M5 RGB is an ideal memory kit for a wide variety of PC builds and themes.

Designed for Stylish Performance
Available in matte white or matte black aluminium heatspreaders and standing at 41 mm tall, the Ripjaws M5 RGB series is designed for a minimalistic look and features customizable RGB lighting to match a wide variety of PC build themes.

Today, Lenovo launched the Lenovo ThinkPad P14s Gen 5 designed for professionals who need top-notch performance in a portable 14-inch chassis. Featuring a stunning 16:10 display, this mobile workstation is powered by AMD Ryzen PRO 8040 HS-Series processors. These processors are ultra-advanced and energy-efficient, making them perfect for use in thin and light mobile workstations. The AMD Ryzen PRO HS- Series processors also come with built-in Artificial Intelligence (AI) capabilities, including an integrated Neural Processing Unit (NPU) for optimized performance in AI workflows.

The Lenovo ThinkPad P14s Gen 5 is provided with independent software vendor (ISV) certifications and integrated AMD Radeon graphics, making it ideal for running applications like AutoCAD, Revit, and SOLIDWORKS with seamless performance. This mobile workstation is ideal for mobile power users, offering advanced ThinkShield security features and passes comprehensive MIL-SPEC testing for ultimate durability.