Latest news with #ProjectDigits
Yahoo
21-05-2025
- Business
- Yahoo
Computex showdown: Nvidia & MediaTek tipped to steal Windows-on-Arm spotlight
When you buy through links on our articles, Future and its syndication partners may earn a commission. There have been rumblings of an Nvidia arm-based CPU for the last few years, but the project just might be revealed next week. At CES in January, Nvidia and MediaTek announced Project Digits, a personal AI supercomputer. Shortly after the announcement, Nvidia rebranded the Project Digits concept as the Nvidia DGX Spark, powered by the Grace Blackwell GB10 chipset. Unfortunately, Nvidia and MediaTek's GB10 superchip is too large for most computers, which is where these rumored new chips come in. According to rumors this month (collected by German outlet ComputerBase), Nvidia and MediaTek are slated to unveil a smaller, more laptop-friendly arm chipset at Computex in Taipei City, Taiwan, next week. But what do we know about these new Nvidia and MediaTek chips so far? The Nvidia/MediaTek N1 chips w direct rival to Qualcomm's Snapdragon X series. Nvidia and MediaTek's new chips are being called N1X and N1. The silicon will power desktops and laptops aimed at general users, unlike the DGX Spark, an AI supercomputer. The N1X and N1 chips appear to combine MediaTek's Arm-based CPUs with Nvidia's Blackwell GPU technology. The N1X and N1 processors are currently expected to feature up to 10 Cortex-X925 high-performance cores and up to 10 Cortex-A725 cores, with less powerful configurations able to address more budget-conscious users. Nvidia and MediaTek's N1X and N1 chips will utilize the Arm instruction set and run a version of Windows OS optimized for Arm-based chips. This will make the Nvidia/MediaTek N1 chips a direct rival to Qualcomm's Snapdragon X series. Due to the inclusion of Nvidia's Blackwell GPU tech, the N1X and N1 could also rival AMD's Ryzen AI Max APU. SemiAccurate indicates that the N1X and N1 CPUs will hit the shelves in early 2026, due to technical hurdles that delayed the ship date. This production timeline would make sense with a Computex announcement, as it gives Nvidia and MediaTek at least 7 months to resolve technical issues before the chips ship. Based on the hardware powering these rumored N1X and N1 chips, we could see them implemented in powerful workstation machines, gaming rigs, and general consumer-level machines. After all, Qualcomm's Snapdragon X chips are used in ultra-light and portable laptops, while AMD's Ryzen AI Max APU has been used for the Asus ROG Flow Z13 gaming tablet and the HP ZBook Ultra 14 G1a mobile workstation. We don't expect to see laptops or desktops with the rumored N1X and N1 processors until 2026. However, we could better understand what systems this chip will ship in at Computex next week, as Nvidia and MediaTek are both holding keynote presentations at the trade show on May 19 and May 20th. And since Qualcomm's keynote presentation focuses more on AI than the highly anticipated second generation of Snapdragon X chips, Nvidia and MediaTek could steal the spotlight as the summer's darling Windows-on-Arm chipset. Nvidia's affordable RTX 5060 is coming soon, but that's not the GPU I have my eyes on Doom The Dark Ages is my favorite kind of nightmare — this is how it runs on an RTX 5090 gaming laptop Who really owns your Switch 2? "It's-a-me!" suggests Nintendo
Yahoo
06-04-2025
- Business
- Yahoo
Nvidia's mini desktop supercomputer is coming this year
When you buy through links on our articles, Future and its syndication partners may earn a commission. Nvidia has unveiled a line of artificial intelligence (AI) supercomputers that can deliver unprecedented processing power in a portable, desktop-friendly chassis. Previously dubbed Project Digits, the powerful machines first revealed at CES 2025 in January have been rebranded as the DGX Spark and DGX Station machines. These computers are powered by Nvidia's Blackwell Ultra platform and promise up to a petaFLOP in processing power — upwards of 1,000 times faster than the best laptops or high-end desktop PCs. Blackwell Ultra is designed for massive-scale AI training and testing, and the DGX machines promise to put that power into the hands of data scientists, AI researchers and students at a relatively affordable price point. It's the equivalent of putting the power of a data center into a computer small enough to fit on your desk. The DGX Spark is a little box smaller than a laptop that you could easily tuck away on a corner of your desk or fit into your bag. It stands just under 2 inches (5 centimeters) in height and slightly under 6 inches (15 cm) in width, and at its core is the GB10 Grace Blackwell Superchip, capable of delivering 1,000 AI TOPS (trillions of operations per second). Related: The 9 most powerful supercomputers in the world right now It also comes with 128 gigabytes of unified system memory and Nvidia's full stack of AI software, including a number of tools, libraries as well as some pretrained models. This includes things like the CUDA Deep Neural Network (cuDNN) library for enhancing neural network layers during training and inference and a pre-trained SegFormer model. The Nvidia version of the DGX Spark is available to reserve online starting at $3,999, although the company has said other models will soon be available from manufacturers like ASUS, Dell and Lenovo. The DGX Station is the Spark's larger, more powerful sibling and is closer in size to a professional workstation. Built around the GB300 Grace Blackwell Ultra Desktop Superchip, it contains a staggering 748 GB of "large coherent" memory — memory which can be accessed by more than one processor at a time. It also features Nvidia's ConnectX-8 SuperNIC, which enables network connectivity at a blistering rate of up to 800 gigabits a second — fast enough to download approximately five 4K movies in a second. It also uses the NVLink-C2C Interconnect to connect internal components at 900 gigabytes per second. The DGX Station is a powerhouse designed to execute large-scale AI training and inferencing workloads from the comfort of your desktop, without having to access additional resources through the cloud. RELATED STORIES —Supercomputer runs largest and most complicated simulation of the universe ever —World's fastest supercomputer 'El Capitan' goes online — it will be used to secure the US nuclear stockpile and in other classified research —Japan to start building 1st 'zeta-class' supercomputer in 2025, 1,000 times more powerful than today's fastest machines The DGX Station isn't currently available to reserve, but Nvidia has indicated it will be available later in 2025 from partners like Asus, Dell, HP, Lambda and Supermicro. Jensen Huang, Nvidia's founder and CEO, said the new DGX machines represent a natural next step in AI development. "AI has transformed every layer of the computing stack. It stands to reason a new class of computers would emerge, designed for AI-native developers and to run AI-native applications,' he said in a statement. "With these new DGX personal AI computers, AI can span from cloud services to desktop and edge applications."
Yahoo
24-03-2025
- Business
- Yahoo
More Petaflop AI mini supercomputers set to launch in 2025 as Asus, Dell and HPE reveal their plans for DGX Spark sidekicks
When you buy through links on our articles, Future and its syndication partners may earn a commission. Nvidia's DGX Spark, once called Project Digits, is a tiny AI supercomputer Built on GB10, is delivers 1000 TOPS and 200B parameter support Dell, HPE, and Asus will offer GB10-based alternatives with similar performance Nvidia has announced DGX Spark, a Mac Mini-sized AI supercomputer designed to bring advanced model development and inferencing directly to desktops. The mini machine was originally called Project Digits and expected to be priced at $3000, but the change of name has caused the figure to skyrocket as it's now priced at $3999, according to Nvidia's reservation page. Built around the GB10 Grace Blackwell Superchip, DGX Spark features a Blackwell GPU with fifth-generation Tensor Cores, FP4 support, and NVLink-C2C, which enables high-bandwidth memory sharing between the GPU and Grace CPU. The system offers up to 1,000 trillion operations per second of AI compute power and supports models with up to 200 billion parameters. It is designed to handle demanding AI workflows such as fine-tuning, inference, and prototyping without relying entirely on external infrastructure. DGX Spark includes 128GB of LPDDR5x unified memory and up to 4TB of NVMe SSD storage, and delivers performance previously limited to data centers. It's aimed at developers, researchers, data scientists, and students working with increasingly complex AI models locally, so it's not something most people will need. 'AI has transformed every layer of the computing stack. It stands to reason a new class of computers would emerge - designed for AI-native developers and to run AI-native applications,' said Jensen Huang, founder and CEO of Nvidia. 'With these new DGX personal AI computers, AI can span from cloud services to desktop and edge applications.' Some of Nvidia's OEM partners are debuting desktop AI systems based on the same GB10 architecture. Dell's Pro Max with GB10 fits into the company's broader AI workstation portfolio, connecting with the Dell AI Factory with Nvidia to give developers an easy path from deskside development to deployment. HP's ZGX Nano AI Station is another entry, offering comparable capabilities for developers who want performance and scalability without full server infrastructure. Asus has also introduced its GB10 AI super computer, the Ascent GX10. Pricing details have not yet been confirmed, but Nvidia lists it on its DGX Spark pre-order page where it says the GX10 will cost $2999 and come with 1TB of storage. These are the best mobile workstations you can buy right now I am thrilled by Nvidia's cute petaflop mini PC wonder Nvidia's DGX Station is a powerful supercomputer-class workstation PC
Euronews
19-03-2025
- Business
- Euronews
GTC 2025: 5 key takeaways from Nvidia's developer conference
Nvidia's CEO and founder Jensen Huang took to the stage at the semiconductor company's annual software conference on Tuesday to announce new products. While giving his keynote speech, Huang also expanded on his thoughts surrounding the future of artificial intelligence (AI) which he said was 'at an inflexion point'. In case you missed it, here are five key takeaways from GTC 2025 in San Jose, California. 'AI agents will be everywhere,' Huang told the audience at his keynote, referring to the technology that is designed to take autonomous actions to assist humans without their input. 'How they run, what enterprises run, and how we run it will be fundamentally different. And so we need a new line of computers,' he continued. The answer to this large computing requirement, he said, is two new computers, called DGX Spark (previously called Project Digits) and DGX Station. DGX Spark is already available and allows developers to prototype, fine-tune, and inference (the process that machines use to conclude new data) the latest generation of reasoning AI models, such as DeepSeek, Meta, Google, and others. Nvidia says that it can also 'seamlessly deploy' to data centres or the cloud. It is available in a desktop-friendly size. The much bigger DGX Station is in the works, but Huang said 'this is what a PC should look like and this is what computers will run in the future. And we have a whole lineup for enterprise now, from little, tiny ones to workstation ones'. It resembles a tower desktop and like its smaller cousin, will allow users to prototype fine-tune, and run AI models. DGX Station is marketed toward enterprise customers that would need to run heavy AI workloads. It features Nvidia's so-called 'super chip,' the Grace Blackwell Ultra. Nvidia announced several technologies to supercharge humanoid robot development. One includes a very Star Wars-esque-style robot that is being developed in partnership with Disney Research and Google DeepMind. It is called Newton and is Nvidia's physics engine, a computer software that helps simulate robot behaviour, to develop robots. One of the small Disney robots made its debut on stage next to Jensen. The entertainment company hopes to bring these robots to its theme parks next year. Nvidia said Newton is supposed to help robots be more 'expressive' and 'learn how to handle complex tasks with greater precision'. The company also announced NVIDIA Isaac GR00T N1, a technology it says is the world's 'first open, fully customisable foundation model for generalised humanoid reasoning and skills'. Huang said that the future will see a move from traditional data centres - buildings dedicated to handling general-purpose computing - to AI factories, which handle AI specifically. The European Commission has made AI factories one of its priorities for the bloc's AI development. 'AI workloads aren't static. The next wave of AI applications will push power, cooling, and networking demands even further,' Nvidia states in a blog post on its website. Huang announced Omniverse Blueprint, which lets engineers design, test, and optimise a new generation of intelligence manufacturing data centres using digital twins. One of the uses of Blueprint is to ensure AI factories are ready by predicting how changes in AI workloads will affect power and cooling at the data centre scale as well. Huang announced a partnership with General Motors (GM), the largest US carmaker, to integrate AI chips and software into its autonomous vehicle technology and manufacturing. The partnership will see Nvidia's platforms train GM's AI models to develop its next-generation vehicles, factories, and robots. AI is now going out 'to the rest of the world,' in robotics and self-driving cars, factories, and wireless networks,' Huang said. 'One of the earliest industries AI went into was autonomous vehicles… We build technology that almost every single self-driving car company uses,' he added. Huang also announced Nvidia was building a quantum computing research centre in Boston. The NVIDIA Accelerated Quantum Research Center, or NVAQC, will provide 'cutting-edge technologies to advance quantum computing,' according to the company. 'Quantum computing will augment AI supercomputers to tackle some of the world's most important problems, from drug discovery to materials development,' Huang said in his keynote. 'Working with the wider quantum research community to advance CUDA-quantum hybrid computing, the NVIDIA Accelerated Quantum Research Center is where breakthroughs will be made to create large-scale, useful, accelerated quantum supercomputers'. A group of 35 EU lawmakers has called upon Henna Virkkunen, the European Commissioner in charge of technology, to require member states to exclude high-risk vendors from 5G telecom infrastructure in the wake of a probe into alleged bribery of MEPs involving Huawei. 'We call upon you to make the EU toolbox for 5G security a legally binding instrument, ensuring its full and uniform application to high-risk vendors,' they said in a letter seen by Euronews. The lawmakers include Aura Salla (Finland/EPP), Svenja Hahn (Germany/Renew), Bart Groothuis (Netherlands/Renew), and Alexandra Geese (Germany/Greens-EFA). In 2020 the Commission adopted a so-called 5G Cybersecurity Toolbox to protect 5G networks from cyber threats and risks. EU member states agreed to apply restrictions for suppliers considered to be high risk – such as China's Huawei and ZTE – including necessary exclusions, following security concerns. To date, only a limited number of countries have taken concrete steps to ban the companies. Euronews reported last year that fewer than half the 27 EU member states have used legal powers to impose restrictions on such telecom suppliers. Wednesday's letter says that making the toolbox binding is 'necessary to establish a uniform and robust defence against foreign interference'. 'The time for voluntary measures has passed,' they add. Earlier this month in a telecom congress in Barcelona, Virkkunen said that she will explore ways to speed up the implementation and enforcement of 5G security measures taken by member states to protect critical communication networks from foreign interference. Huawei lobbyists were banned from entering the Parliament's premises on Friday, following allegations of bribery linked to the Chinese company's lobbying activities. The ban is temporary and will remain in place until the authorities conclude their investigation. Huawei said in a statement that it has "a zero-tolerance policy towards corruption or other wrongdoing, and we are committed to complying with all applicable laws and regulations at all times".
Yahoo
28-01-2025
- Business
- Yahoo
AMD Just Got Hit Again--And This Time, It's Not Because of China
Advanced Micro Devices (NASDAQ:AMD) just cant catch a break. After getting slammed in the recent AI chip selloff, the stock took another hitthis time from Melius Researchs Ben Reitzes, who downgraded it to Hold from Buy. His reason? Its not just the DeepSeek panic. The real problem is Nvidia (NASDAQ:NVDA) coming straight for AMDs turf in both AI and PC chips. Nvidias new Arm-based CPUs and AI-powered desktops could shake up the traditional PC market, putting AMD in a tough spot. Reitzes also slashed his price target from $160 to $129, signaling that AMDs premium valuation might not hold up much longer. Warning! GuruFocus has detected 3 Warning Signs with NVDA. And the competition is only heating up. Nvidias latest play, Project Digits, introduces its GB10 Grace Blackwell Superchip, a direct challenge to AMD and Intel (NASDAQ:INTC) in high-performance computing. If Nvidia succeeds in pushing its integrated AI solutions, AMDs CPU growth could slow even furtherReitzes now expects just 3% growth in 2027, a steep cut from his previous 9% forecast. Even AMDs once-dominant server market is looking shaky, with cloud giants like Amazon turning to custom chips like Graviton, and Qualcomm (QCOM) eyeing the space. If that trend continues, AMDs stronghold in data centers could erode faster than expected. Still, AMD isnt out of the game. Its Turin server chip is making strides, and it remains a key player in AI GPUs. But expectations are shiftinganalysts have revised their 2025 data-center GPU revenue estimates from $10 billion to $7 billion, a sign that Nvidias grip on the market is tightening. With competition ramping up on all fronts, investors are left wondering: Can AMD defend its ground, or is it in for a long battle to stay relevant? This article first appeared on GuruFocus. Sign in to access your portfolio
