Nvidia marks Paris tech fair with Europe AI push
Nvidia CEO Jensen Huang delivers his keynote address at the Vivatech fair in Paris. (AP pic)
PARIS : Drawing high-powered tech CEOs and a presidential visit, the Vivatech trade fair opened in Paris on Wednesday with a bang as Nvidia boss Jensen Huang announced a major push into Europe.
'In just two years we will increase the amount of AI computing capacity in Europe by a factor of 10,' Huang told a packed hall in a southern Paris convention centre, striding around the stage wearing his trademark leather jacket.
He also announced a multi-billion-dollar partnership with French AI champion Mistral AI.
President Emmanuel Macron hailed the Nvidia-Mistral tie-up as a 'historic' opportunity for France and Europe, urging other local firms to climb aboard.
He had arrived late on Wednesday afternoon for a tour of the show and meetings with European startups about technological sovereignty, a subject dear to his heart.
'We want AI… that's secure, sustainable, humanist,' Macron said.
People from around the globe thronged the halls of Vivatech, crammed with stands in blaring colours showing off the latest innovations from startups, tech giants and more traditional firms and patrolled here and there by gesticulating robots.
Around 14,000 startups and more than 3,000 investors were expected in Paris, while organisers forecast total visitor numbers to at least equal last year's 165,000 people.
Nvidia's Huang took top billing with an opening presentation of almost two hours that drew bouts of rapturous applause from attendees.
The US firm's tie-up with Mistral will see the companies build a cloud computing platform powered by 18,000 of Nvidia's 'Blackwell' high-end chips worth billions.
Speaking in a panel discussion with Huang and Macron, Mistral chief Arthur Mensch said the offering would be 'completely independent' in a nod to the president's sovereignty drive.
'You're no longer relying for your AI workload on certain of the US providers,' he promised the audience.
Macron dubbed the Mistral-Nvidia collaboration a 'game-changer, because it will increase our sovereignty and it will allow us to do much more' with AI.
Europe 'has put its ability to produce things in danger' and 'become more and more dependent on the rest of the world,' he warned.
Aside from Mistral, Nvidia will also intensify work with existing partners like Germany's Siemens and France's Schneider Electric, Huang said.
And it will help build multiple data centres in seven European countries.
Europe is well behind competitors like the US and China in building up the computing power needed to power generative artificial intelligence.
The continent hosts 'less than 5% of global computing power, whereas we consume 20%,' Macron's office said in a press briefing ahead of the leader's visit to Vivatech.
Nvidia has seen export restrictions slapped on its top-performing chips by Washington, with American politicians leery of ceding their country's lead in generative AI.
Remaining high-tech controls on China are at issue in high-stakes trade talks with Beijing.
Huang has warned that the US' superpower rival is nevertheless making swift strides to catch up.
There was little sign of impact from export restrictions on Nvidia's chip sales in its May earnings release.
But the company has warned the braking effect may be larger in the current quarter.
US politics also preoccupies many European tech leaders and policymakers.
Concerns range from Trump's mercurial tariff policy to the continent's ability to stand on its own without US tech giants – and the massive gap in funding for AI development between the two sides of the Atlantic.
'Sovereignty, which wasn't as important in the conversation just a year or two years ago, has become an absolutely strategic priority,' Vivatech managing director Francois Bitouzet told AFP.
Macron's hammering on tech sovereignty followed on from his hyping of French and European openness to AI at a Paris global summit in February.
Macron, Mensch and Huang were set to dine together behind closed doors at the president's Elysee Palace residence on Wednesday evening.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Sun
an hour ago
- The Sun
Trend Micro Enhances AI Safety and Security in NVIDIA Enterprise AI Factories
HONG KONG SAR - Media OutReach Newswire - 12 June 2025 - Trend Micro Incorporated (TYO: 4704; TSE: 4704), a global cybersecurity leader, today announced its adoption of the NVIDIA Agentic AI Safety blueprint, enhancing foundational security to help ensure customers' AI systems are protected across their full development and deployment lifecycle. To read more about how Trend Micro enhances security across NVIDIA Enterprise AI factory deployments, please visit: Kevin Simzer, COO at Trend Micro: 'Global organizations are racing to innovate with agentic AI systems, and there's a critical need to ensure the safety and security of these systems. The NVIDIA Agentic AI Safety blueprint provides an important enabling technology that works in conjunction with Trend's threat intelligence to support safety across all phases of the AI lifecycle – from model adoption, deployment, and runtime protection — allowing customers to innovate with AI faster.' The 'Trend Secure AI Factory' is built around unified cybersecurity platforms Trend Vision OneTM and Trend Vision One – Sovereign Private Cloud. It's designed to help customers mitigate risk at every layer of the AI factory, encompassing data, models, microservices, infrastructure, networks, and users. Pat Lee, VP of Strategic Enterprise Partnerships at NVIDIA: 'Embedding real-time, autonomous threat detection into enterprise AI factories empowers organizations to confidently scale innovation without compromising on protection. By integrating advanced cybersecurity directly into AI factories with Trend Micro and NVIDIA Agentic AI blueprints, enterprise data, models, and workloads can remain resilient and trusted —unlocking the full potential of AI in a secure, accelerated environment.' To achieve this goal, Trend is deepening its AI-driven capabilities through the integration of Trend Cybertron, its cybersecurity-specific large language model. Built to detect and respond to evolving threats in real-time, Trend Cybertron is now deployable via NVIDIA universal LLM NIM microservices, enabling scalable, secure inference across cloud, hybrid, and on-premise environments. This addition supports the alliance's shared goal of delivering intelligent, production-ready infrastructure for the AI era. Thanks to capabilities including data security, infrastructure posture management, API guardrails, and CI/CD validation, the Trend Secure AI Factory directly supports and reinforces NVIDIA Agentic AI Safety blueprint by: • Integrating with the NVIDIA NeMo model assessment, training, and customization framework to ensure model safety mechanisms scale reliably and securely across enterprise environments. • Safeguarding model integrity against poisoning and misuse during the training and evaluation phases • Securing the deployment environment, including microservices and infrastructure (e.g., NVIDIA NIM and AI agents), with Trend Container Security. This helps to prevent adversarial manipulation or resource-based attacks. • Protect sensitive datasets with Data Risk Posture Management (DSPM), leveraging NVIDIA AI Enterprise, which includes NVIDIA Morpheus, NVIDIA RAPIDS, and the NVIDIA AI Safety Recipe for evaluations and post-training, to adhere to privacy and compliance standards. • Providing guardrails, network protection, and secure AI agent interactions with application users, via Trend Zero Trust Secure Access (ZTSA) AI Service Access • Strengthening sovereign AI with trusted security controls, via Trend Vision One – Sovereign Private Cloud Justin Vaïsse, Director General at the Paris Peace Forum: 'As AI becomes increasingly embedded in critical systems, its safety and security must be treated as global priorities. We welcome the role of companies like Trend Micro in advancing responsible AI by contributing tangible, scalable solutions to multi-actor partnerships. This kind of cross-sector collaboration is essential to fostering trust and resilience in the technologies shaping our shared future.'
New Straits Times
2 hours ago
- New Straits Times
Nvidia to exclude China from forecasts amid US chip export curbs
KUALA LUMPUR: Nvidia will no longer include the China market in its revenue and profit forecasts following stringent US trade restrictions on chip sales to the region, CEO Jensen Huang told CNN on Thursday. When asked if the United States would lift the export controls after the trade discussions with China in London this week, Huang said he was not counting on it. "But, if it happens, then it will be a great bonus. I've told all of our investors and shareholders that, going forward, our forecasts will not include the China market," he said. Huang reiterated his criticism of US chip export curbs during his CNN interview, building on his earlier remarks about the restrictions from April that prevented Nvidia from selling its H20 chip made for China. "The goals of the export controls are not being achieved," he said. "And so I think, with all export controls, the goals have to be well-articulated and tested over time." "Beyond next quarter, if Nvidia is not able to resume sales into China, we believe there may be some downside to expectations for calendar year 2026," said D.A. Davidson analyst Gil Luria. Nvidia was still evaluating its "limited options for the China market", it said on Thursday. "Until we settle on a new product design and receive approval from the US government, we are effectively foreclosed from China's US$50 billion data center market," the company said. The export limits cost it US$2.5 billion in sales during the first quarter and it expects an US$8 billion sales hit in the second quarter. "By zero-basing China, Nvidia removes a volatile variable that neither Wall Street nor the Commerce Department can reliably handicap," said Michael Ashley Schulman, CIO of Running Point Capital, adding that any China sales would come as an upside surprise. The company reported US$4.6 billion in first-quarter revenue from H20 sales as customers stockpiled the chips before the export controls set in, with the China business accounting for 12.5 per cent of overall revenue.
The Star
4 hours ago
- The Star
The China trade deal can't go soft on chips
THE Donald Trump administration's most recent round of trade talks with China has ended, but the great export-control debate is just getting underway. Export controls were at the centre of the negotiations in London this week; for years, they have anchored America's strategy for economic warfare with Beijing. China wants few things more than a relaxation of curbs on its access to high-end semiconductors; so do corporate players and analysts who argue that they do more harm than good. Although the details are hazy, the Trump administration reportedly agreed to lift some recent controls as part of a deal to deescalate the trade war. Let's hope the administration doesn't go much further than that. Ditching or dramatically rolling back the measures that preserve American's technological advantage – and that have been imposed, with bipartisan backing, over the past several years – would be strategic self-sabotage of the highest form. The export controls at issue pertain to the most advanced computer chips and the inputs, both hardware and software, required to make them. During Trump's first term, the United States hit specific Chinese firms, namely Huawei, with targeted bans. Under President Joe Biden, Washington went big: It intensified the tech war by expanding their application to cover China as a whole. Since then, Washington and Beijing have been playing cat-and-mouse, as China has sought to evade tech controls while the United States gradually tightens them. Export controls can seem arcane, but their purpose is strategically profound. The goal, as then-National Security Adviser Jake Sullivan said in 2022, was to maintain 'as large a lead as possible' in key sectors like artificial intelligence (AI) and advanced computing, because those sectors will shape the future economic and military balance. Leaps in AI will affect everything from the lethality of warfare in the Western Pacific to the productivity of US and Chinese workers. Export controls, designed to keep China from accessing the crown jewels of US and allied innovation – and to exploit Washington's unique position in key nodes of the semiconductor supply chain – are vital to staying ahead in this fateful race. Doubts remain Not everyone agrees. The debut of DeepSeek's R1 reasoning model in late 2024 raised questions about the dominance of US AI companies, just as the unveiling of a next-generation Huawei smartphone in 2023 sparked speculation that the company was shrugging off US bans. Last month, Jensen Huang, the CEO of Nvidia – which makes top-of-the-line chips and would surely love to sell them in China – claimed that US controls have backfired by causing Beijing to push harder for self-sufficiency. But that position ignores a great deal of evidence. Any honest evaluation starts with a realistic standard for success. The goal of US policy was never to comprehensively halt Chinese innovation. It was to slow Beijing in key areas that matter most to the future distribution of technological advantage and global power. If you want to know whether US controls have had that effect, just ask Beijing. DeepSeek's CEO has admitted that the chip ban is his company's number one obstacle. Chinese leader Xi Jinping and his country's Politburo have acknowledged the pressures US restrictions create. Xi repeatedly pushed for the easing of export controls in his interactions with Biden. China wants those restrictions gutted because its leaders understand their power all too well. This is why Beijing has been wielding its own limits on the sale of critical minerals: Imitation is the surest sign China appreciates the leverage export controls convey. Other critiques falter under scrutiny. US tech restrictions can't have spurred China's drive for self-reliance: 'Made in China 2025,' Xi's bid to dominate key areas of innovation, was unveiled years before the United States got serious about export controls. To suggest that a Leninist party-state needs US encouragement to seek technological primacy is to misunderstand how determined, how power-minded, Chinese leaders are. The major problem with US export controls is the degree to which they have been evaded and undermined. Nvidia has built chips for export that just barely skirt performance thresholds the United States government has established. Huawei has used shell companies and other illicit measures for getting prohibited chips. US policy has also been imperfect: The Biden administration's deliberate approach to building bureaucratic and allied consensus around new controls often gave Beijing time to stockpile and blunt their impact. And if Trump's deal with Xi entails an agreement not to impose new export controls in the future, it will — thanks to the cat-and-mouse dynamic — severely erode America's ability to keep even its existing restrictions effective and up to date. Leading the way Export controls can't preserve US advantage by themselves. Such a strategy also requires running faster, by investing in America's production and innovation ecosystem, and by ensuring this country remains a magnet for global talent. Trump's approach to those issues is, sometimes, self-defeating: Witness his hostility to foreign students (not just Chinese students) and his attacks on the elite universities that have long been the envy of the world. But countries don't win fierce tech competitions simply by being the most innovative versions of themselves. In the Cold War, Washington and its allies used broad export controls and technological restrictions to keep the Soviet bloc well behind. The specific measures will vary, but a similar mindset is required today. A hot war between the United States and China threatens; the new cold war is well underway. America will struggle to win the technological race that may well determine the outcome if it doesn't slow the enemy down. — Bloomberg Hal Brands is a Bloomberg Opinion columnist and the Henry Kissinger Distinguished Professor at Johns Hopkins University's School of Advanced International Studies. The views expressed here are the writer's own.
