Latest news with #GraceBlackwell
Time of India
04-07-2025
- Business
- Time of India
Nvidia just dropped a game-changer — CEO Jensen Huang calls it a miracle for AI supercomputing
Nvidia CEO Jensen and co-founder Huang recently unveiled a 'miracle' in AI and technology, the GB200 Grace Blackwell superchip at the GTC Paris Keynote during VivaTech 2025, as reported by Barchart. Nvidia's Technological Milestone He emphasized that it was a milestone achievement for his company and even for the broader tech industry, saying, 'Grace Blackwell systems. It is really a miracle. It's a miracle from a technology perspective,' as quoted in the report. Mass Production on a New Level Huang highlighted that, 'The supply chain that came together to build these GB200 systems, two tons each, we're producing them now, 1000 systems a week,' as quoted in the report. He even said that, 'No one has ever mass-produced supercomputers at this scale before,' and also pointed out at the logistical and engineering feats involved, as reported by Barchart. ALSO READ: Elon Musk flip-flops? Slams Trump's Big Beautiful Bill, then praises him as a peacemaker — what's going on? Meeting a Global AI Demand Nvidia's latest supercomputer technology comes at a time when demand for AI infrastructure is constantly increasing, driven by advancements in machine learning, autonomous systems, and data analytics, according to the report. The CEO even emphasized that, 'We've really taken AI supercomputing to a whole new level,' and also highlighted on how NVIDIA's innovation is not just a technical marvel but also a response to global demand for more powerful computing solutions, as reported by Barchart. Live Events ALSO READ: Shocking tariff list: Trump to hit these countries with 70% duties — even longtime US allies aren't spared Inside the GB200 Grace Blackwell Superchip According to the report, the AI chipmaker's GB200 Grace Blackwell superchip combines two Blackwell B200 GPUs and a Grace CPU, which is interconnected through NVLINK to create a unified memory domain with up to 896GB of accessible memory. This latest architecture lets the system to scale up to 512 GPUs within a single NVLINK domain, which enables massive parallel processing and high throughput for demanding AI workloads, as reported by Barchart. The GB200 is designed for both training and inference of large language models, as it delivers up to 30 times the inference performance of previous generations, along with major gains in power efficiency due to liquid-cooled racks and optimized data processing capabilities, according to the Barchart report. The GB200's ability to process massive amounts of data at lightning speed makes it appropriate for applications ranging from scientific research to enterprise-scale AI, and its mass production at a rate of 1,000 systems per week is reportedly an unprecedented achievement in the supercomputing industry, accoridng to the report. ALSO READ: Good news for some: Child tax credit gets a boost, but millions left behind in Trump's Big Beautiful Bill Nvidia's Market Cap The latest development comes as the chipmaker has become the world's most valuable semiconductor company, hitting a market capitalisation of more than $3 trillion and even surpassing tech giants like Microsoft and Amazon, as reported by Barchart. ALSO READ: 10 key takeaways from Donald Trump's Big Beautiful Bill summarised simply FAQs What did Jensen Huang unveil? Jensen Huang introduced the GB200 Grace Blackwell superchip and called it a technological 'miracle,' as per the report by Barchart. What is GB200 Grace Blackwell? It's Nvidia's latest AI supercomputing chip, a powerful combo of GPUs and CPUs made for massive AI tasks, as per the report by Barchart.
Yahoo
27-06-2025
- Business
- Yahoo
Nvidia's stark turnaround is being powered by robots, hyperscalers, and sovereign AI
Nvidia (NVDA) is riding a wave of AI enthusiasm that has sent its stock price soaring to new highs after being battered in the early part of the year on fears of new AI models out of China and tariff threats. On Friday, shares of Nvidia hit $157, up from $94 in April, marking a stunning turnaround for the world's leading AI chip company. The moves can be attributed to a number of factors turning in Nvidia's favor, including the successful ramp of the buildout of its Blackwell chip line, the explosion in interest around so-called sovereign AI, and the company's push into what it calls physical AI, or robotics. Nvidia's Blackwell ramp, the process of building out chips and increasing overall volume, has been a major win for the company. During its GTC Paris event earlier this month, Nvidia said it is now shipping 1,000 Grace Blackwell (GB200) server racks per week and is on track for its transition to its next-generation Blackwell Ultra-powered GB300 servers. Getting its chips and servers out the door as fast as possible has been key for Nvidia, especially after earlier delays as the company began to ramp production of Blackwell chips in late 2024. Nvidia said the GB300 line of chips will be able to slot directly into GB200 servers, which the company claims will make it easier for its customers to upgrade to its latest offerings. Nvidia also continues to benefit from the seemingly insatiable demand for its chips from hyperscalers, including Amazon (AMZN), Google (GOOG, GOOGL), Microsoft (MSFT), Meta (META), and xAI, among others. According to CFO Colette Kress, large cloud service providers accounted for just under 50% of the company's data center sales in its fiscal first quarter. And with hyperscalers expected to spend billions more on their AI buildouts in the coming months, there's little reason to believe that will slow down anytime soon. CEO Jensen Huang is also betting big on the idea of sovereign AI, or AI data centers built in specific countries to power their own AI needs. Huang was on hand during President Trump's visit to the Middle East in May, during which Trump announced that Saudi Arabia and the United Arab Emirates (UAE) will be able to purchase thousands of Nvidia chips for their data centers. One of those projects, the UAE's Project Stargate, could include 100,000 Nvidia GPUs, according to Reuters. Huang also touted sovereign AI plans for Europe during GTC Paris. "The company is establishing AI technology centers in Germany, Sweden, Italy, Spain, UK, and Finland, and working with regional cloud and telco partners in France, UK, and Germany to build new AI datacenters with [tens] of thousands of Grace Blackwell systems and GPUs," Bernstein analyst Stacy Rasgon wrote in an investor note following the event. "They are planning for 20 AI factories to be deployed in Europe, several at gigawatt scale, and sees European compute capacity growing by a factor of 10x over the next 2 years," Rasgon added. Huang and company also continue to push further into physical AI, with the CEO calling it a multitrillion-dollar opportunity. Physical AI is another way of referring to the software and computers needed to power things like humanoid robots and self-driving cars. Nvidia offers both chips and the training information needed to run robots in factories and has been working in the automotive industry for years on self-driving car technologies. But don't expect to have a robot walking around your home anytime soon. Those kinds of bipedal butlers are likely years away from being something you can grab at your local Best Buy. Importantly, Nvidia has managed to shake off fears over the Trump administration's ban on sales of its chips to China. Despite taking a hefty charge on canceled orders, Wall Street analysts are more upbeat on the issue now that the ban is official and they no longer have to question whether one is coming or not. After all, there's comfort in certainty. Of course, Nvidia still faces the prospect of competition from rivals like AMD (AMD). And its own customers, including Amazon, Google, and Microsoft, are building or currently using their own in-house chips to take on Nvidia's offerings, making for a more complicated relationship. For now, however, Nvidia is rolling into summer with the wind at its back. Email Daniel Howley at dhowley@ Follow him on X/Twitter at @DanielHowley. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Mint
11-06-2025
- Business
- Mint
Nvidia teams up with Perplexity and Mistral AI: How will it impact Europe's AI infrastructure push?
Nvidia has announced new collaborations with AI firms Perplexity and Mistral AI as part of a broad initiative to strengthen artificial intelligence infrastructure and language model development across Europe. The announcement was made during an AI conference in Paris, where the US-based chipmaker outlined its strategy to support regional AI growth with both hardware and software solutions. In partnership with search firm Perplexity, Nvidia will work with over a dozen AI companies in Europe and the Middle East to refine their technologies and help distribute them to local businesses. This effort will involve the development of reasoning models—AI systems capable of handling more complex tasks—in various European languages, where training data is often limited. Mistral AI, a French startup, is joining forces with Nvidia to launch a new service called Mistral Compute, which will run on 18,000 of Nvidia's Grace Blackwell chips. The service will be based in the company's data centre in Essonne, France, with plans to expand across the continent. The aim is to facilitate domestic AI computing capabilities and reduce dependence on overseas infrastructure. Jensen Huang, Nvidia's chief executive, said that Europe urgently needs to scale up its AI infrastructure if it wants to stay competitive globally. The firm is supporting this push by partnering with cloud and telecom providers and enabling small-scale access to its AI accelerators—specialised chips used to train and deploy AI models. Kari Briski, Nvidia's vice president for generative AI software, said the company is producing synthetic data in low-resource European languages to help train more robust local models. "We're doing a lot of synthetic data generation and translating our reasoning datasets so they can be used in training," she said. Nvidia's collaboration includes firms in France, Germany, Italy, Poland, Spain and Sweden. Perplexity will support the deployment and accessibility of these models, ensuring they can be used by local enterprises for tasks such as content research and data analysis. The company's CEO, Aravind Srinivas, noted that Germany is already its second-largest market by revenue, underscoring the growing demand for regionally tailored AI services. As part of the broader rollout, Nvidia also confirmed that its AI Lepton service, which connects developers to essential compute infrastructure, will now include participation from AWS and Mistral. The company added that 'AI factories'—large-scale data centres for training and operating AI models—are being developed in over 20 European locations, several of which will be 'gigafactories' housing more than 100,000 chips. In the UK, companies such as Nebius Group and Nscale Global Holdings are also set to deploy thousands of Nvidia's AI semiconductors, while other countries, including Italy and Armenia, are preparing to upgrade their hardware systems. Despite Europe having over 1.5 million AI developers and nearly 10,000 businesses involved in Nvidia's Inception startup programme, the continent still lags behind the US in terms of infrastructure investment. Nvidia said it plans to triple the region's AI hardware capacity next year and increase total computing power tenfold. (With inputs from Bloomberg and Reuters)
Yahoo
11-06-2025
- Business
- Yahoo
Nvidia Teams Up With Startup Mistral as Part of European AI Push
(Bloomberg) -- Nvidia Corp. announced a raft of projects aimed at bolstering artificial-intelligence infrastructure across Europe, including an expanded partnership with French startup Mistral AI. Trump's Military Parade Has Washington Bracing for Tanks and Weaponry NY Long Island Rail Service Resumes After Grand Central Fire NYC Mayoral Candidates All Agree on Building More Housing. But Where? Senator Calls for Closing Troubled ICE Detention Facility in New Mexico California Pitches Emergency Loans for LA, Local Transit Systems Chief Executive Officer Jensen Huang made the announcements during a joint Nvidia-VivaTech event in Paris, part of a globe-trotting campaign to promote the adoption of AI and his company's products. A data center buildout is needed in Europe to help countries there catch up in deploying the technology, he said. The chipmaker is trying to expand the market for AI accelerators — the processors used to develop and run artificial intelligence models. Nvidia is pushing for countries to deploy technology on a national level and trying to make it easier for individual companies to get the benefits from AI. In France, Nvidia will team up with Mistral to use local AI computing to run the startup's services. An offering called Mistral Compute will tap 18,000 new Grace Blackwell chips from Nvidia. It will be developed in Mistral's data center in Essonne, France, and the company plans to roll it out to other locations in Europe. In the UK, AI firms Nebius Group and Nscale Global Holdings Ltd. will use 'thousands' of such semiconductors for their own platforms. Other countries, including Italy and Armenia, also are installing new hardware, Nvidia said. In Europe, Nvidia is working with 1.5 million developers and 9,600 businesses, as well as 7,000 startups in what the company calls its inception program. 'The only thing that's missing is infrastructure,' Dion Harris, Nvidia's director of data center and high-performance computing, said in a briefing ahead of the presentations. Nvidia is working with cloud and telecommunications companies across Europe, he said. Europe has lagged behind the US in developing the infrastructure for AI and hasn't matched the spending promised in other regions. Huang said at an event in London on Monday with UK Prime Minister Keir Starmer that a lack of infrastructure was holding back growth in a country that otherwise had the expertise and startups to be a global competitor in AI. Huang said at the GTC-VivaTech event that more than 20 so-called AI factories are being planned and built across Europe in the next two years, with 'several' of them being 'gigafactories.' The larger facilities will be home to over 100,000 chips. It calculates that AI hardware capacity in Europe will grow by three times next year. 'We will increase the amount of AI computing capacity in Europe by a factor of 10,' he said. Santa Clara, California-based Nvidia has transformed its fortunes over the last three years and now gets almost as much revenue per quarter as Intel Corp., its longtime nemesis, gets in a year. Much of that money come from AI accelerator chips, which are used by a cadre of giant companies to develop AI software and services. That group, which includes Microsoft Corp. and Meta Platforms Inc., provides about half of Nvidia's total sales. The chipmaker is looking to reach a wider market by promoting the use of smaller-scale systems by companies and countries. Adding to a previous announcement, Nvidia said Inc.'s AWS, Mistral and others are joining the chipmaker's Lepton service, which helps connect AI developers with the computing hardware they need. Nvidia said that European countries need help to get AI models deployed that are based on local languages and data. It's providing software and services that will accelerate those efforts. Separately, Nvidia said that vehicles using its chips and software are starting to appear on the road — the result of years of work. Mercedes-Benz Group AG's CLA models and forthcoming vehicles from Volvo and Jaguar will rely on its Drive platform. --With assistance from Rachel Metz. New Grads Join Worst Entry-Level Job Market in Years The Spying Scandal Rocking the World of HR Software American Mid: Hampton Inn's Good-Enough Formula for World Domination Cavs Owner Dan Gilbert Wants to Donate His Billions—and Walk Again The SEC Pinned Its Hack on a Few Hapless Day Traders. The Full Story Is Far More Troubling ©2025 Bloomberg L.P.
Yahoo
11-06-2025
- Business
- Yahoo
Nvidia Teams Up With Startup Mistral as Part of European AI Push
(Bloomberg) -- Nvidia Corp. announced a raft of projects aimed at bolstering artificial-intelligence infrastructure across Europe, including an expanded partnership with French startup Mistral AI. Trump's Military Parade Has Washington Bracing for Tanks and Weaponry NY Long Island Rail Service Resumes After Grand Central Fire NYC Mayoral Candidates All Agree on Building More Housing. But Where? Senator Calls for Closing Troubled ICE Detention Facility in New Mexico California Pitches Emergency Loans for LA, Local Transit Systems Chief Executive Officer Jensen Huang made the announcements during a joint Nvidia-VivaTech event in Paris, part of a globe-trotting campaign to promote the adoption of AI and his company's products. A data center buildout is needed in Europe to help countries there catch up in deploying the technology, he said. The chipmaker is trying to expand the market for AI accelerators — the processors used to develop and run artificial intelligence models. Nvidia is pushing for countries to deploy technology on a national level and trying to make it easier for individual companies to get the benefits from AI. In France, Nvidia will team up with Mistral to use local AI computing to run the startup's services. An offering called Mistral Compute will tap 18,000 new Grace Blackwell chips from Nvidia. It will be developed in Mistral's data center in Essonne, France, and the company plans to roll it out to other locations in Europe. In the UK, AI firms Nebius Group and Nscale Global Holdings Ltd. will use 'thousands' of such semiconductors for their own platforms. Other countries, including Italy and Armenia, also are installing new hardware, Nvidia said. In Europe, Nvidia is working with 1.5 million developers and 9,600 businesses, as well as 7,000 startups in what the company calls its inception program. 'The only thing that's missing is infrastructure,' Dion Harris, Nvidia's director of data center and high-performance computing, said in a briefing ahead of the presentations. Nvidia is working with cloud and telecommunications companies across Europe, he said. Europe has lagged behind the US in developing the infrastructure for AI and hasn't matched the spending promised in other regions. Huang said at an event in London on Monday with UK Prime Minister Keir Starmer that a lack of infrastructure was holding back growth in a country that otherwise had the expertise and startups to be a global competitor in AI. Huang said at the GTC-VivaTech event that more than 20 so-called AI factories are being planned and built across Europe in the next two years, with 'several' of them being 'gigafactories.' The larger facilities will be home to over 100,000 chips. It calculates that AI hardware capacity in Europe will grow by three times next year. 'We will increase the amount of AI computing capacity in Europe by a factor of 10,' he said. Santa Clara, California-based Nvidia has transformed its fortunes over the last three years and now gets almost as much revenue per quarter as Intel Corp., its longtime nemesis, gets in a year. Much of that money come from AI accelerator chips, which are used by a cadre of giant companies to develop AI software and services. That group, which includes Microsoft Corp. and Meta Platforms Inc., provides about half of Nvidia's total sales. The chipmaker is looking to reach a wider market by promoting the use of smaller-scale systems by companies and countries. Adding to a previous announcement, Nvidia said Inc.'s AWS, Mistral and others are joining the chipmaker's Lepton service, which helps connect AI developers with the computing hardware they need. Nvidia said that European countries need help to get AI models deployed that are based on local languages and data. It's providing software and services that will accelerate those efforts. Separately, Nvidia said that vehicles using its chips and software are starting to appear on the road — the result of years of work. Mercedes-Benz Group AG's CLA models and forthcoming vehicles from Volvo and Jaguar will rely on its Drive platform. --With assistance from Rachel Metz. New Grads Join Worst Entry-Level Job Market in Years The Spying Scandal Rocking the World of HR Software American Mid: Hampton Inn's Good-Enough Formula for World Domination Cavs Owner Dan Gilbert Wants to Donate His Billions—and Walk Again The SEC Pinned Its Hack on a Few Hapless Day Traders. The Full Story Is Far More Troubling ©2025 Bloomberg L.P. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
