Nvidia to resume H20 GPU chip sales to Beijing, launches China-compliant model
According to a company blog post, Nvidia is filing applications to resume H20 sales with the U.S. government and expects to get the licenses soon. Deliveries are expected to begin shortly thereafter.
The company announced a new RTX Pro GPU designed specifically for China. Nvidia described the model as "fully compliant" and suitable for digital twin AI applications in sectors such as smart factories and logistics.
The chipmaker's CEO Jensen Huang met with U.S. President Donald Trump and policymakers in Washington and later with officials in Beijing, as part of efforts to promote AI cooperation and highlight Nvidia's support for open-source research and global AI development, the company said.
In May, Reuters reported that the company was planning to release a downgraded version of its H20 artificial intelligence chip for China following U.S. export restrictions on the original model.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
UPI
a few seconds ago
- UPI
More people are considering AI lovers, and we shouldn't judge
People are falling in love with their chatbots. There are now dozens of apps that offer intimate companionship with an AI-powered bot, and they have millions of users. A recent survey of users found that 19% of Americans have interacted with an AI meant to simulate a romantic partner. The response has been polarizing. In a New Yorker article titled "Your A.I. Lover Will Change You," futurist Jaron Lanier argued that "when it comes to what will happen when people routinely fall in love with an A.I., I suggest we adopt a pessimistic estimate about the likelihood of human degradation." Podcaster Joe Rogan put it more succinctly -- in a recent interview with Sen. Bernie Sanders, the two discussed the "dystopian" prospect of people marrying their AIs. Noting a case where this has already happened, Rogan said: "I'm like, oh, we're done. We're cooked." We're probably not cooked. Rather, we should consider accepting human-AI relationships as beneficial and healthy. More and more people are going to form such relationships in the coming years, and my research in sexuality and technology indicates it is mostly going to be fine. When surveying the breathless media coverage, the main concern raised is that chatbots will spoil us for human connection. How could we not prefer their cheerful personalities, their uncomplicated affection and their willingness to affirm everything we say? The fear is that, seduced by such easy companionship, many people will surely give up their desire to find human partners, while others will lose their ability to form satisfying human relationships even if they want to. It has been less than three years since the launch of ChatGPT and other chatbots based on large language models. That means we can only speculate about the long-term effects of AI-human relationships on our capacity for intimacy. There is little data to support either side of the debate, though we can do our best to make sense of more short-term studies and other pieces of available evidence. There are certain risks that we do know about already, and we should take them seriously. For instance, we know that AI companion apps have terrible privacy policies. Chatbots can encourage destructive behaviors. Tragically, one may have played a role in a teenager's suicide. The companies that provide these apps can go out of business, or they can change their terms of service without warning. This can suddenly deprive users of access to technology that they've become emotionally attached, with no recourse or support. Complex relationships In assessing the dangers of relationships with AI, however, we should remember that human relationships are not exactly risk-free. One recent paper concluded that "the association between relationship distress and various forms of psychopathology is as strong as many other well-known predictors of mental illness." This is not to say we should swap human companions for AI ones. We just need to keep in mind that relationships can be messy, and we are always trying to balance the various challenges that come with them. AI relationships are no different. We should also remember that just because someone forms an intimate bond with a chatbot, that doesn't mean it will be their only close relationship. Most people have lots of different people in their lives, who play a variety of different roles. Chatbot users may depend on their AI companions for support and affirmation, while still having relationships with humans that provide different kinds of challenges and rewards. Meta's Mark Zuckerberg has suggested that AI companions may help solve the problem of loneliness. However, there is some (admittedly very preliminary data) to suggest that many of the people who form connections with chatbots are not just trying to escape loneliness. In a recent study (which has not yet been peer reviewed), researchers found that feelings of loneliness did not play a measurable role in someone's desire to form a relationship with an AI. Instead, the key predictor seemed to be a desire to explore romantic fantasies in a safe environment. Support and safety We should be willing to accept AI-human relationships without judging the people who form them. This follows a general moral principle that most of us already accept: we should respect the choices people make about their intimate lives when those choices don't harm anyone else. However, we can also take steps to ensure that these relationships are as safe and satisfying as possible. First of all, governments should implement regulations to address the risks we know about already. They should, for instance, hold companies accountable when their chatbots suggest or encourage harmful behavior. Governments should also consider safeguards to restrict access by younger users, or at least to control the behavior of chatbots who are interacting with young people. And they should mandate better privacy protections -- though this is a problem that spans the entire tech industry. Second, we need public education so people understand exactly what these chatbots are and the issues that can arise with their use. Everyone would benefit from full information about the nature of AI companions but, in particular, we should develop curricula for schools as soon as possible. While governments may need to consider some form of age restriction, the reality is that large numbers of young people are already using this technology, and will continue to do so. We should offer them non-judgmental resources to help them navigate their use in a manner that supports their well-being, rather than stigmatizes their choices. AI lovers aren't going to replace human ones. For all the messiness and agony of human relationships, we still (for some reason) pursue other people. But people will also keep experimenting with chatbot romances, if for no other reason than they can be a lot of fun. Neil McArthur is director of the Center for Professional and Applied Ethics at the University of Manitoba. This article is republished from The Conversation under a Creative Commons license. Read the original article. The views and opinions in this commentary are solely those of the author.
CNBC
a few seconds ago
- CNBC
OpenAI expands Oracle data center deal, says parts of Stargate 1 in Texas are operational
Oracle and OpenAI will develop an additional 4.5 gigawatts of Stargate data center capacity in the U.S., the companies announced on Tuesday, kicking off an expansion of an already massive infrastructure project. Stargate is a joint venture between OpenAI, Oracle and Softbank that aims to invest up to $500 billion in artificial intelligence infrastructure in the U.S. over the next four years. President Donald Trump announced Stargate shortly after his inauguration in January. OpenAI said the additional data center capacity will create over 100,000 jobs across construction and operations roles in the U.S., and pushes the total Stargate AI data center capacity under development to over 5 gigawatts, according to a blog post. The update from OpenAI comes one day after the Wall Street Journal reported that the Stargate project has struggled to get off the ground. The report said the companies have scaled back their near-term plans and were striving to open a small data center by the end of the year. "Together, we're committed to delivering 10GW of new compute capacity through Stargate — one of the most ambitious infrastructure projects in U.S. history," OpenAI and SoftBank said in a statement to CNBC. "We're moving with urgency on site assessments and reimagining how data centers are designed to power advanced AI and make its benefits widely accessible." OpenAI said Tuesday that construction of the Stargate I in Abilene, Texas, is underway, and that parts of the facility are already up and running. Oracle started delivering the first Nvidia GB200 racks to the facility last month, the blog post said. Stargate I has created "thousands" of jobs, including specialized positions for electricians, equipment operators, and technicians, OpenAI said. "Easy to throw around numbers, but this is a _gigantic_ infrastructure project," OpenAI CEO Sam Altman wrote in a post on X on Tuesday.
CNBC
a few seconds ago
- CNBC
How's the consumer doing? Capital One will soon give the Street a fresh look
Club holding Capital One will unveil its second-quarter earnings report after the bell Tuesday, offering Wall Street a crucial look into the financial pulse of the U.S. consumer. Capital One's financials don't just give investors an understanding of the company's performance. As one of America's largest credit card issuers — with a more diverse client base than affluent-focused American Express — Capital One's release is also a barometer for assessing the impact of ongoing economic uncertainty on its customers' spending habits. Insights like these are especially important as the market continues to speculate over President Donald Trump's tariffs, the Federal Reserve's next monetary policy decision, persistent conflict overseas and the start of another quarterly earnings season. But a mixed bag of recent economic data has only complicated things further. On one hand, inflation picked up in June as tariffs begin to slowly work their way through the economy. Despite the Fed's best efforts, the rise in both headline and core CPI showed investors last week that inflation still isn't fully under control. Meanwhile, the labor market has proved to be more resilient as U.S. payrolls increased more than expected last month. There's also been a five-week decline in initial jobless claims, suggesting that layoffs aren't happening at a concerning pace. At the same time, continuing claims do remain elevated. That could indicate those who do lose their jobs are having difficulties landing new roles. The stock market has taken the murky backdrop in stride, though. The S & P 500 and tech-heavy Nasdaq Composite both registered record closes Monday. The S & P 500 was basically flat Tuesday, while the Nasdaq was slightly lower. Capital One's upcoming release, however, should give investors a chance to cut through all the noise. Here are four ways the second-quarter results can provide a snapshot of U.S. consumer health. COF YTD mountain Capital One (COF) year-to-date performance Late payments Each quarter, management shares delinquency rates for its credit card and consumer banking businesses. Capital One reports its 30-day delinquency rates – known as the percentage of total outstanding credit card balances that are at least 30 days past due. If this rate rises, it can be a sign of financial stress for the U.S. consumer. CEO Richard Fairbank said so himself during an April conference call. He described delinquencies as "the best leading indicator" for a customers' financial health. "Our delinquencies were stable on a seasonally adjusted basis throughout most of 2024," Fairbank said. "And ... they improved relative to our seasonal expectation over the last six months." In the first quarter of 2025, Capital One's 30-day delinquency rate for its credit card division came in at 4.27%, down from 4.54% the quarter prior. Delinquency rates are important for Capital One shareholders like us, too. That's because the credit card business is the company's biggest money maker – making up roughly 70% of its overall revenues last quarter – from interest-based income and various fees on their offerings. Some on Wall Street seem optimistic about Capital One's credit quality. Ahead of quarterly earnings, TD Cowen upgraded the financial stock to a buy rating from hold. "Delinquencies have continued to demonstrate stability, and card losses have been outperforming expectations," the analysts wrote in a July 8 research note. TD Cowen also praised Capital One's recently completed $35 billion acquisition of Discover — a key reason we first initiated a position in the stock. Potential losses Additionally, if Capital One is worried that more customers are going to struggle paying back their loans in the future, they will park more money away to cover those potential losses. The amount the company sets aside is called its credit loss provision, and they report the number reach quarter. If management raises these, it could serve as an early warning that consumers can't pay back their debts. Provision for credit losses came in lower than analysts expected at $2.3 billion last quarter, down from roughly $2.7 billion in the quarter prior. Car market Third, there's auto loan originations. If auto lending increases, it can suggest consumers feel confident enough in the economy to make larger purchases. This barometer for consumer health, however, has been complicated with tariffs. Some may have felt rushed to buy vehicles as the deadline for larger levies on top U.S. trading partners have loomed. In the case of Capital One, the company reported a 22% year over-year-increase in auto loan originations last quarter. Auto lending is a relatively small business for Capital One, though. It's one of several reported within Capital One's consumer banking division, which makes up roughly 20% of the company's overall revenues. Executive color Finally, it's more than just the raw numbers on press releases and financial supplements. Management's commentary during the conference call can also provide more insight into consumer behavior. CEO Fairbank, for example, talked about Capital One's revolve rates last quarter. That's the percentage of a customer's credit balance not paid off in full each month. The executive reassured investors when he said this rate has "stabilized over the past year" and remains "below pre-pandemic levels for our major products and segments." While higher revolve rates can benefit Capital One because it means more interest-based revenues, it can also mean customers are relying more on credit due to economic pressures like stagnant wages or inflation. Fairbank, a widely respect financial industry veteran, may also just share his thoughts on consumer health outright. He said the U.S. consumer "remains a source of strength in the economy" during the company's last earnings call. Although he acknowledges that "some pockets [of them] are feeling the pressure," on the whole he described U.S. customers as being in "good shape." In June, the executive struck a positive tone again. "Despite all the noise out there and the tariff news and everything, even when we look at the very latest daily data on things like spending data or anything related to consumer behavior, we just don't see an effect," Fairbank said during Morgan Stanley's U.S. financials conference on June 10. He added, "It's as if our consumers aren't really reading the same newspaper that we are. So, I'm cautiously optimistic about what I see." Let's just hope this trend continues into the rest of 2025. (Jim Cramer's Charitable Trust is long COF. See here for a full list of the stocks.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust's portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
