Supermicro just backed the new NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs for many of their servers and workstations. They designed these systems for the Blackwell PCIe GPUs to help more businesses use fast computers for AI tasks like LLM work, agent AI, better graphics, and sharing resources. A bunch of these Supermicro systems have NVIDIA Certification, which means they work perfectly with NVIDIA AI Enterprise software. This makes building and using AI much easier for companies.
Charles Liang is the president and CEO of Supermicro. He says his company leads everyone else with its many GPU servers that fit different needs and ship fast. Adding support for the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU gives customers even...
AMD Radeon RX 9000 graphics cards flew off shelves in record time. The company reported selling 200,000 Radeon RX 9070 and RX 9070 XT cards during their first release wave. This information came straight from AMD at the AI PC Innovation Summit held in Beijing. You'll have trouble finding these cards at stores right about everywhere because they sell out fast.
According to supply predictions, stores might see normal stock levels early next quarter. Many card makers charge extra for their custom designs—some RX 9070 XT versions cost $200 more than AMD's suggested price. AMD has asked sellers to stick with recommended prices but doesn't force pricing rules on its partners. They accept that premium parts often lead to higher costs.
The...
GIGABYTE just hit an amazing overclocking record with their Z890 AORUS TACHYON ICE motherboard matched with the newest Intel Core Ultra Series 2 chip. The secret sauce behind this record? Their special LN2 SWITCH tech plus V-COLOR Manta XFinity RGB DDR5 memory. Famous overclocker HiCookie pushed these parts to reach DDR5-12752, a completely new world record. This proves GIGABYTE leads the pack for extreme overclocking results.
The Z890 AORUS TACHYON ICE stands out as a board made exactly for overclocking fans. It comes packed with an advanced all-digital power system that helps users break limits easily. The board includes handy shortcut keys, toggle switches, voltage-checking features, and a complete overclocking toolkit. All these...
PNY just added five new graphics cards to what they sell. These include the NVIDIA RTX PRO 6000 Blackwell Workstation Edition, RTX PRO 6000 Blackwell Max-Q Workstation Edition, RTX PRO 5000 Blackwell, RTX PRO 4500 Blackwell, and RTX PRO 4000 Blackwell. All these cards help professionals with their work. The new NVIDIA RTX PRO Desktop GPUs use the brand new NVIDIA Blackwell design. They work really fast for jobs that mix AI with making art, designing things, and engineering tasks.
These cards have the newest Tensor Cores and RT Cores and up to 96 GB of super-fast GDDR7 memory. This lets them do amazing things with AI, ray tracing, and neural graphics. They can make workstations much more powerful for the future when more jobs will use...
Dell just marked one year working with NVIDIA on their AI Factory. They brought out fresh AI computers, equipment, and programs and helped businesses speed up their AI work at any level. Companies need good AI to stay ahead, but they face problems like putting systems together and not having the right skills. This slows down what they can do with AI. More than 75% of companies want their equipment makers to help them through all steps of using AI, creating a high demand for simple AI setups that can grow.
Dell works hand in hand with NVIDIA as the top maker of AI equipment. They give customers the same great experience across all parts of AI—hardware, programs, and support services. This makes Dell a one-stop shop for companies that...
Giga Computing showed up at NVIDIA GTC 2025, bringing their best GPU systems for AI, media, and language models. These guys run under GIGABYTE and make cutting-edge cooling tech for AI servers. At booth #1409, they displayed GIGAPOD, their complete AI package that comes in both air and liquid-cooled versions for the NVIDIA HGX B300 NVL16 system. They also brought a piece of their brand new NVIDIA GB300 NVL72 rack system, plus two servers that work with the just-announced NVIDIA RTX PRO 6000 Blackwell Server Edition.
Giga Computing mixed hardware, software, and design services to create GIGAPOD. This system uses GIGABYTE GPU servers with NVIDIA GPU baseboards and runs on GIGABYTE POD Manager software. The air-cooled version fits nine...
NVIDIA teamed up with TSMC to create new co-packaged optics technology for their upcoming network switches. They put silicon photonics right on the switch chips, cutting power use by 3.5 times compared to regular systems. The signal loss drops from 22 dB to just 4 dB with this setup. This fixes major power and connection problems in big GPU systems, especially those with more than 10,000 GPUs working together.
Their design places laser sources inside the switch frame, using only 2 watts per port instead of the usual 10 watts needed by plug-in modules. The built-in optical engines need just 7 watts versus 20 watts for standard signal processors. These changes slash total power needs from around 72 megawatts down to 21.6 megawatts when...
NVIDIA revealed its next flagship AI chip at GTC 2025. It is named Rubin Ultra to replace its current Blackwell Ultra. NVIDIA plans to launch this chip in late 2027. Each Rubin Ultra package connects four AI chips directly together, letting them share cache memory.
NVIDIA packed a massive 1 TB of HBM4e memory into the package. The company claims each chip will reach 100 petaFLOPs of FP4 performance. NVIDIA also introduced their new Vera CPU to replace the older Grace model. The Vera uses 88 custom Arm CPU cores NVIDIA created themselves rather than using standard designs.
Thanks to SMT technology, these special cores run 176 processes simultaneously. The chip transfers data extremely fast with its 1.8 TB/s NVLink C2C connection...
SK Hynix has sent samples of its new 12-layer HBM4 memory chips to major customers ahead of all competitors. These super-fast DRAM chips for AI reached customers early because SK Hynix knows what it's doing with both design and manufacturing. The company has already started testing these chips with its customers. They aim to be ready for full production in the second half of this year, which will make them even stronger in the AI memory business.
These new 12-layer HBM4 chips have the most storage and fastest speeds of any AI memory you can find. The chips process more than two terabytes of data each second, something no other product can do. Think about watching 400 HD movies all at the same time - that's how much data these chips...
