'How hard can it be?' Nvidia CEO on leap to supercomputing
:: Stanford, California
:: May 29, 2025
:: John Hennessy, Former president, Stanford University
'It's a big leap going from being a graphics company to being a supercomputer company. That's a big leap of different business, it requires a lot more system expertise, more software expertise. So how do you think about assembling a team? Because that really had to change the…"
:: Jensen Huang, Nvidia CEO
'It drives all the same way. So first, you (say) 'Hey we can do this.' And so it always starts with this: 'Hey, guys, this we can do this.' And we can do it. So you start with a dream, and the next thing, the next logical leap is: how hard can it be? I always start every meeting with 'Hey guys we should start building cars.' It's sensible, we have computer vision now, you know, planning algorithms (unintelligible). Let's do it. And then somebody goes, 'we don't know anything about cars.' Well, how hard can it be? And so I think the clincher on that one is my mom can drive.
HENNESSY: 'Yeah, actually, humans are pretty good drivers. Look at it...'
HUANG: 'I know my mom shouldn't, but she can. And so, how hard can it be? And so the supercomputers were all the same thing. You know, we just… A.I., how hard can it be? And so we just kind of just go into it.'
His remarks came the same day the U.S. Department of Energy announced that Nvidia and Dell will supply core technology for 'Doudna,' a new supercomputer to be installed in 2026 at Lawrence Berkeley National Laboratory.
The event also came amid renewed political scrutiny of Nvidia's global operations, including bipartisan concern in Washington over the company's planned R&D facility in Shanghai.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Business Insider
27 minutes ago
- Business Insider
Nvidia AI Chips Worth $1B Were Smuggled to China to Defy Trump, Claims Report
U.S. semiconductor giant Nvidia (NVDA) stock edged up today despite reports that its chips were smuggled into China to beat President Trump's export controls. Elevate Your Investing Strategy: Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence. Beating Controls The Financial Times reported today that Nvidia's high-end B200 processors worth around $1 billion were smuggled to China in the three months after Washington tightened chip export controls at the height of the tariff trade war between the two nations. The FT stated that the processors, banned for sale in China, are widely available on a thriving Chinese black market for U.S. chips. The FT report cited sales contracts, company filings and insider information. It stated that in May, multiple Chinese distributors started selling B200s to suppliers of data centers that serve Chinese AI groups. It added that the most likely source of the restricted chips is Southeastern Asian nations. Nvidia, led by chief executive Jensen Huang, told Reuters that building data centers with smuggled products is inefficient both technically and financially, as the company only offers service and support for authorized products. It has said that there is no evidence of any 'AI chip diversion.' There is no suggestion the company is involved in or has knowledge of its restricted products being sold to China. AI Race Nvidia last week said it would be allowed to resume sales to China after the Trump administration reversed an export restriction on the sales of chips such as H20. The curbs were imposed in April. It comes as the U.S. said it is looking to relax AI export controls and boost infrastructure investment as part of its newly unveiled AI action plan. At the heart of it is the continued battle for global AI supremacy between the U.S. and China. As can be seen below, Nvidia, which relies heavily on the Chinese market, has good reason to find a careful balance between the two. That includes pledges from Nvidia to invest billions of dollars in U.S.-manufactured chips and electronics over the next four years, as well as regular trips to China to reaffirm its commitment to the country. Is NVDA a Good Stock to Buy Now? On TipRanks, NVDA has a Strong Buy consensus based on 34 Buy, 3 Hold and 1 Sell ratings. Its highest price target is $250. NVDA stock's consensus price target is $182.49, implying a 5.52% upside.
Business Insider
an hour ago
- Business Insider
China's Repair Market for Banned Nvidia (NVDA) AI Chips Is Booming
In China, demand is quickly rising for repair services of Nvidia's (NVDA) advanced artificial intelligence chips, even though these products are banned from being exported to the country by the United States. Several small repair companies in Shenzhen now specialize in fixing the chipmaker's high-end H100 and A100 GPUs. These bans, which began in 2022, were meant to limit China's progress in AI and military technology, but the ongoing repairs suggest that many of these chips continue to enter the country. Elevate Your Investing Strategy: Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence. The repair industry has grown so fast that some companies have created new businesses just to handle the influx of work. This sudden boom has led to worries about large-scale smuggling of banned chips, especially since government and military purchase records have confirmed their presence in China. In response, U.S. lawmakers from both political parties have proposed laws to track these chips after they are sold, an idea that was recently supported by President Donald Trump's administration. Although Chinese tech company Huawei has introduced its own AI chips, Nvidia's products remain the top choice for demanding tasks, such as training large language models. Since many H100 and A100 chips in China have been running nonstop for years, this has led to a higher failure rate and increased the need for repairs. Interestingly, fixing a single chip can cost $1,400 to $2,800, depending on the issue. At the same time, Chinese buyers are now looking to get Nvidia's newest B200 chips, which cost over 3 million yuan (about $415,000) per server, showing just how strong the demand for these processors remains. What Is a Good Price for NVDA? Turning to Wall Street, analysts have a Strong Buy consensus rating on NVDA stock based on 34 Buys, three Holds, and one Sell assigned in the past three months, as indicated by the graphic below. Furthermore, the average NVDA price target of $182.49 per share implies 5% upside potential.
Yahoo
2 hours ago
- Yahoo
As Elon Musk, Mark Zuckerberg And Sam Altman Chase Nvidia AI Chips, Jensen Huang Says 'Just Call Me' — Here's How Allocation Really Works
Benzinga and Yahoo Finance LLC may earn commission or revenue on some items through the links below. Nvidia Corp. (NASDAQ:NVDA) CEO Jensen Huang revealed the company's straightforward chip allocation process during Wednesday's All In Podcast, dismissing concerns about favoritism among tech titans competing for scarce artificial intelligence processors. What Happened: When asked how he manages allocation requests from high-profile clients like Tesla Inc.'s (NASDAQ:TSLA) Elon Musk, Meta Platforms Inc.'s (NASDAQ:META) Mark Zuckerberg and OpenAI's Sam Altman, Huang offered a surprisingly simple answer: 'Place a PO. That's it.' Huang said that the company now discloses its roadmap to partners a year in advance for better planning. Trending: Be part of the breakthrough that could replace plastic as we know it— Huang addressed supply concerns by revealing he 'wrote off $5 billion worth of Hoppers,' adding, 'If anybody would like to have some extras, you got them. Just give me a call.' The allocation process has become critical as demand for NVIDIA's H100 AI chips intensifies. Current Hopper architecture GPUs maintain strong residual value, retaining 75-80% of their original value after one year and approximately 50% after three years. Responding to Musk's projection of deploying '50 million H100 equivalents' within five years, Huang described AI infrastructure as requiring 'factories of AI' that continuously produce tokens. He estimates the industry is 'a couple hundred billion dollars into a multi-trillion dollar infrastructure buildout per year.'Why It Matters: The chip shortage has intensified amid U.S. sanctions affecting Chinese customers. Chinese server maker H3C recently warned of depleted H20 chip inventory, with supply uncertainties extending beyond April. Major Chinese firms, including Alibaba Group Holding Ltd. (NYSE:BABA) and Tencent Holdings Ltd. (OTC:TCEHY) have increased H20 orders following DeepSeek's emergence in the AI market. Previous reports indicate the competitive dynamics at play. Oracle Corp.'s (NYSE:ORCL) Larry Ellison and Musk reportedly 'begged' Huang for additional GPUs during a dinner at Nobu Palo Alto, with Ellison stating they told the CEO to 'please take our money.' Read Next: $100k+ in investable assets? Match with a fiduciary advisor for free to learn how you can maximize your retirement and save on taxes – no cost, no obligation. If there was a new fund backed by Jeff Bezos offering a 7-9% target yield with monthly dividends would you invest in it? Image via Shutterstock This article As Elon Musk, Mark Zuckerberg And Sam Altman Chase Nvidia AI Chips, Jensen Huang Says 'Just Call Me' — Here's How Allocation Really Works originally appeared on
