Latest news with #AIDiffusionrule
CNBC
a day ago
- Business
- CNBC
‘Sovereign AI' is poised to boom under Trump. Wall Street is sizing up a multibillion-dollar market
Global artificial intelligence deals under the Trump administration this year have thrust "sovereign AI" into the limelight, bringing to life a new engine to power the AI investment cycle. While the first major wave of AI was tied to innovation spurred by private sector tech companies, Wall Street is now looking for opportunities that are tied to the growth from the public sector. These so-called sovereign AI deals involve a nation's capabilities to produce and control its own AI infrastructure, data, regulations and networks. The term isn't new, but analysts and investors are taking notice of the trend given U.S. President Donald Trump 's enthusiasm to strike AI deals between domestic tech giants and select foreign allies. In May, the president scrapped Biden-era U.S. chip export restrictions , known as the "AI diffusion rule." That move helped pave the way for Nvidia and Advanced Micro Devices to supply chips to Saudi company Humain . "Sovereign AI has risen from the ashes of the Biden AI Diffusion rule to now squarely fit with the Trump Administration's goals to reshape allies in the Middle East around American technology," Ben Reitzes, head of technology research at Melius Research, said in a May note to clients. "Nvidia (along with even AMD and Cisco) now have a spring in their step. … This trend is just starting." Some nations have also released long-term frameworks for their AI public and commercial infrastructure strategies. Canada last December launched the Canadian Sovereign AI Compute Strategy, while India kicked off its IndiaAI Mission in March 2024 . Further, last year, a bipartisan group of U.S. senators rolled out a road map for artificial intelligence policy, seeking a $32 billion annual investment in AI research and development by 2026. Underpinning this trend is the belief that AI will be a key tool in influencing global power dynamics. "With c. $16tn annual global economic value on the line, AI is becoming key in setting the new world order," Bank of America analyst Haim Israel said in a May note. "It is not just the race for technological development but also for resources, supply chains, regulation and standards that are on the line," the analyst added. "Whoever controls AI could have an advantage, possibly reshaping the geopolitical balance of power." Bank of America estimates the global sovereign AI market could be worth $50 billion annually and lead to an "AI infrastructure opportunity" ranging from $450 billion to $500 billion, according to another May note to clients. The firm estimates that at least $2.5 trillion in funding toward AI investments and partnerships have been announced so far this year. This includes projects from Big Tech, recent pledges from the Middle East and North Africa region, the Stargate project — which is a joint venture between OpenAI, Oracle and Softbank — and efforts in the European Union . "These should increase the rollout of AI and improve the economics," Bank of America's Israel said. A new driver for leading AI stocks To capture this momentum, major tech companies are building out product portfolios tied to sovereign AI, framing the technology as a catalyst for national economic and industry growth, national security enforcement, and scientific discovery. For starters, Cisco Systems is "the sovereign AI player no one talks about," Reitzes said in a June 2 note to clients. In May, the networking giant announced it would partner with Saudi Arabia's Humain to build AI infrastructure. Reitzes thinks investors are overlooking the value of Cisco's recent deals in the Middle East , as he said greater sovereign demand could lead the company to see a pickup in AI product orders and spur a breakout in shares. He reiterated his buy rating and $78 price target on the stock, which suggests nearly 20% upside from Thursday's close. Cisco shares have risen 8% this year. "Middle Eastern sovereign entities are unproven, and the investment figures seem cartoonish — but Cisco may have achieved the equivalent of adding a hyperscaler to its customer list," Reitzes wrote. "The company boasts an excellent longstanding relationship with the CEO of Saudi's HUMAIN and other leaders of UAE and Qatar entities — who are not short on cash." Cisco has seen rapid growth in its switching business, which has attracted demand from cloud data center customers. In the fiscal third quarter , Cisco's AI infrastructure orders exceeded expectations for the second quarter in a row, Reitzes noted. The company is "getting on the right side of AI, which can help expand its multiple," Reitzes said. "With HUMAIN, UAE and Qatar, these AI orders could multiply — and start to catch investors' attention and contribute to an acceleration in switching later in FY26 and FY27." Analysts have also pointed to chipmakers Nvidia and AMD as the obvious leaders of sovereign AI, given their multiyear AI infrastructure partnerships with Humain. AMD signed a deal with Humain: A $10 billion collaboration to deploy 500 megawatts of AI compute capacity over the next five years. The agreement involves AMD's deployment of its Instinct GPUs, EPYC processors and ROCm open software ecosystem. Nvidia said it is deploying 18,000 of its most advanced GB300 Grace Blackwell chips to Humain. On news of the Humain deal, Bank of America's Vivek Arya reiterated his buy ratings on Nvidia and AMD and lifted his price targets on each stock by $10 to $160 and $130, respectively, citing "upside to long-term AI opportunity." Arya called the Humain partnership "an important win" for AMD. "Conceptually this would be the first time for AMD on a 'similar' footing as NVDA in terms of engagement in large projects," the analyst said in a note. "However, NVDA appears to be getting direct awards, while AMD appears to be engaged in a [joint venture]-like approach (with Cisco) with some unspecified level of investments." The trickle-down impact of sovereign AI investments Analysts and investors think that growth in sovereign AI will have knock-on positive effects on other parts of the AI industry — such as foundry, optical and memory technology companies. In addition to Nvidia and AMD, T. Rowe Price portfolio manager Tony Wang called out Arista Networks and Broadcom as strong plays for exposure to AI leaders as well as to companies building projects abroad. "I think Nvidia would be the most standard way that the world will build on. Broadcom is more custom," he said. Wang manages T. Rowe Price's Science & Technology Fund (PRSCX) , and its holdings include Nvidia, Broadcom, AMD and Arista Networks, as of March 31. Bank of America's Arya named Marvell Technology another beneficiary of the trend and is bullish on the semiconductor company's earnings growth over the next several years. Though Wall Street is growing keen on this trend now, tech CEOs have been calling for the push of government-led AI investments since the early innings of the artificial intelligence boom. Nvidia CEO Jensen Huang and IBM CEO Arvind Krishna have called for countries to continue building out their sovereign AI capabilities. In 2023, Krishna reportedly advocated for governments to set up large language models, national AI computing centers and common data sets for specific use cases. Huang said on Nvidia's May 28 earnings call that "sovereign AI is a new growth engine for Nvidia." During a fireside chat in 2023, he also highlighted "a recognition that every region and every country needs to build their sovereign AI." The call for greater AI investment goes back even further. In 2021, a group of experts chaired by former Google CEO Eric Schmidt warned about increasing competition between the U.S. and China "in the AI era" and suggested a government AI investment plan that wound up very similar to the one U.S. lawmakers proposed in 2024.
Yahoo
29-05-2025
- Business
- Yahoo
Nvidia's CEO blasted Trump policy that will cost the company $10.5 billion in lost revenue—then praised Trump's ‘bold vision' minutes later
It's a tricky time to be a corporate chief executive, as tariffs, trade wars, and the red meat politics of the Trump presidency sow chaos across markets and supply chains. But no CEO may have it more difficult than Nvidia boss Jensen Huang, whose company announced a $4.5 billion hit to inventory on Wednesday as a result of a new U.S. policy on chip exports. As Huang spoke to analysts and investors on the company's earnings call, the CEO demonstrated an impressive feat of gymnastics, walking a fine line to critique the Trump policy that left a massive hole on his company's income statement while being careful not to provoke the president. Huang kicked off the call with sharp criticism of the Trump administration's new export controls which have forced Nvidia to stop selling its H20 chips in China. 'The U.S. has based its policy on the assumption that China cannot make AI chips. That assumption was always questionable, and now it's clearly wrong.' He added, 'Export controls should strengthen U.S. platforms, not drive half of the world's AI talent to rivals.' Nvidia said the new rule will cause it to forfeit $10.5 billion in revenue in the first half of this year. But while his words were pointed, Huang was careful not to call out President Trump by name in his critique, referring only to U.S. policy. When Trump's name did cross Huang's lips, it was to bestow praise on the POTUS. 'President Trump has outlined a bold vision to reshore advanced manufacturing, create jobs, and strengthen national security,' he said later during the call, noting that he was 'honored' to join him in U.S.-United Arab Emirates AI investment projects that include expanding to Nvidia chips. 'President Trump wants U.S. tech to lead.' Later in the call, Huang turned to another piece of government regulation: the so-called AI Diffusion Rule, passed in the last days of the Biden administration to restrict how advanced AI technology and equipment is shared with foreign countries. The Trump administration revoked the rule in January, and in this case, Huang deemed it worthy of crediting the president by name. 'It was really terrific to see that the AI Diffusion rule was rescinded,' Huang said. 'President Trump wants America to win. He also realizes that we're not the only country in the race.' The comments were all the more remarkable since they essentially praised Trump for doing exactly what Huang had moment earlier faulted Trump's administration for not doing. 'He wants the United States to win and recognizes that we have to get the American stack out to the world, and get the world to build on the American stack instead of alternatives,' Huang said. The paradox did not seem to bother investors, who cheered Nvidia's results and sent shares of the stock up roughly 5% in after hours trading on Wednesday. This story was originally featured on 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
Forbes
06-05-2025
- Business
- Forbes
White House Faces Imminent Deadline To Finalize AI Diffusion Rule
The flags of China and the USA are being displayed on a smartphone, with an NVIDIA chip visible in ... More the background, in this photo illustration in Brussels, Belgium, on December 30, 2023. (Photo by Jonathan Raa/NurPhoto via Getty Images) NurPhoto via Getty Images One of the final actions taken by the Biden administration was to release new export controls, commonly referred to as the AI Diffusion rule. The regulations would impose new restrictions on selling certain advanced chips used for artificial intelligence and some 'closed AI model weights.' The rules are intended to increase the U.S.'s control of the use of its technology in the development of AI globally by creating a three-tier system where each level is subjected to different standards. The U.S.'s closest allies and partners were exempted from the rules, while adversaries like Russia and China were effectively blocked from receiving the restricted goods. This left a wide swath of countries in the middle facing restrictions, which could be lessened by agreeing to specific safety and security measures led by the U.S. While the primary target for the new restrictions was China, imposing global limits was seen as necessary to cut down on Chinese companies' ability to rely on the use of subsidiaries in countries not facing American export controls, which could then pass the restricted chips back into China, as documented by The Wall Street Journal . Although the rule would likely not have solved this issue entirely, the Biden administration believed it would make this process more difficult by expanding the countries facing license requirements for shipments of these advanced AI chips. The rule was released as an interim final rule, meaning it can take effect immediately, but compliance with some parts was delayed while public comment was taken. The most significant looming date is May 15, when the export restrictions will take effect. The fast-approaching deadline has led some companies, including Nvidia and Oracle, to pull orders forward to enable customers to receive the potentially restricted goods shipped before the new rules come into force, according to The Information . The proposal has seen significant pushback from tech companies and foreign leaders, as reported by Bloomberg . Still, the Trump administration is not expected to eliminate it. How exactly the White House will finalize the regulations has yet to be settled despite the fast-approaching date for compliance. If no changes are made, the rule would simply go into effect as written. However, the Trump administration has signaled an interest in simplifying and tightening the regulations. However, senior officials have not spoken on the record about the rule much, creating significant uncertainty about what will happen in the coming days. The White House is primarily considering simplifying the rules by removing the three tiers created in the initial proposal. Instead, according to Reuters , this would be replaced 'with a global licensing regime with government-to-government agreements.' This change would appeal to President Donald Trump's preferences for bilateral negotiations, and these deals could be incorporated into the ongoing discussions over reciprocal tariffs. What is uncertain under this plan is what the limits on countries would be until these deals are reached, and whether there would be no restrictions or all countries face the tightest limits until agreements are secured. The other significant change that the Trump administration is considering is to tighten the rules , according to Bloomberg . The main way of doing so would be to lower the threshold for which licenses are required, according to Reuters . Under the current proposal, licenses are required to sell computing power equivalent to about 1,700 of Nvidia's H100 chips. The change under consideration would lower this to the equivalent of about 500 H100 chips. This would be a significant reform and is motivated by being better able to crack down on the issue of China receiving otherwise restricted semiconductors through other countries. Although there is generally an expectation that the White House will act to adjust these rules before the deadline, the regulations may instead be allowed to take effect as written. The expected revisions would then come later. This approach could solve the issue of how to handle the time before bilateral deals are reached and provide a runway before the standards are tightened. However, it would not immediately address the Trump administration's goal of simplifying the regulations, which could make this strategy a nonstarter.
Yahoo
21-04-2025
- Business
- Yahoo
Nvidia stock falls as China's Huawei reportedly boosts AI chip production after Trump's export ban
Nvidia (NVDA) stock fell as much as 3.6% premarket Monday after a report from Reuters said Chinese tech giant Huawei is set to begin shipping advanced AI chips as soon as next month after Trump's new export rules effectively banned the company from selling its H20 chips in China. Huawei's new 910C chips are reportedly competitive with Nvidia's H100 AI chips, which are two generations behind its latest Blackwell chips. The H100 GPUs (graphics processing units) were banned from export to China by the US government in 2022. Though Nvidia has adapted to multiple new US trade rules by creating chips specifically for the Chinese market, the ever-tightening restrictions mean those chips are far less powerful than their mainstream counterparts. The report of the latest Huawei chip comes nearly one week after Nvidia said the US government had effectively banned sales of its H20 graphics processing units — a version of its H100 AI chips made for Chinese consumers to comply with US trade rules. The House Select Committee on China also released a statement saying it was 'demanding answers' from Nvidia about whether it had previously complied with trade restrictions. Fellow chip stocks Advanced Micro Devices (AMD) (also subject to the new US export rules), Qualcomm (QCOM), and Broadcom (AVGO) also fell Monday, each dropping around 2% premarket. Nasdaq futures dipped 1.5%. Nvidia disclosed last week that it will record a $5.5 billion hit from lost inventory and contracts in the first quarter due to the change in trade policy. JPMorgan analyst Harlan Sur projected that, overall, Nvidia will lose as much as $16 billion in the current fiscal year from the H20 ban. Bernstein analyst Stacy Rasgon said last week in a note to investors following the news: 'Banning the H20 makes little sense to us…a ban essentially simply hands the Chinese AI market over to Huawei." China accounted for $17 billion, or 13%, of Nvidia's revenue in its fiscal year 2025, Rasgon noted. Shares of Nvidia fell 7% on the day after the restriction was announced. The company saw roughly 8%, or $230 billion, erased from its market cap last week. Nvidia CEO Jensen Huang visited China and met with trade officials in Beijing last week, while the company is simultaneously looking to expand its domestic manufacturing footprint, pledging $500 billion to the US AI supply chain buildout last week. Nvidia has said it follows trade rules 'to the letter,' and the company was critical of tightening US trade policy under the prior administration. When former US President Biden passed the AI Diffusion rule — set to go into effect in May — capping AI chip exports on a country basis and limiting US companies' data center expansion abroad, Nvidia essentially argued that tighter rules would only bolster foreign competition. 'Rather than mitigate any threat, the new Biden rules would only weaken America's global competitiveness, undermining the innovation that has kept the U.S. ahead,' said Nvidia vice president of government affairs Ned Finkle. Trump set the stage to tariff semiconductors last week by opening a probe into US exports of the products. The president has promised that most electronics will still get wrapped up in his tariffs, which could hurt Nvidia and other Big Tech firms. "We are taking a look at Semiconductors and the WHOLE ELECTRONICS SUPPLY CHAIN in the upcoming National Security Tariff Investigations," he posted last week on Truth Social. Sign in to access your portfolio
Miami Herald
29-03-2025
- Business
- Miami Herald
An event that may save Nvidia is coming May 15
Nvidia (NVDA) has been in a slump for weeks, reaching concerningly low levels. The artificial intelligence (AI) leader has battled extreme volatility over the past month but failed to generate sustainable momentum. As of this writing, shares are down more than 12% for the month and 8% for the week. Despite the momentum generated by last week's GTC (GPU Technology Conference) 2025, market momentum still seems to be working against NVDA and many of its big tech peers. Get expert insights and actionable trade alerts from veteran investing experts and hedge fund managers. Join TheStreet Pro today and get the first month FREE The future may appear uncertain right now, as ongoing pressure from tariffs against Canada, Mexico, and China continues to push down entire sectors and prompt concern about a bear market. However, a major announcement is coming soon, and it has the potential to significantly impact Nvidia and many of its peers, particularly those in the chipmaking space. The catalyst the industry needs may finally be in sight. For the past few months, both investors and consumers have anxiously waited to learn more about how the Trump administration will handle AI policy. Vice President JD Vance has discussed it at multiple events but offered little direct context as to upcoming policies. However, the White House is scheduled to release its so-called AI Diffusion Rule on May 15, described as a "regulatory framework that aims to control foreign countries' access to high-tech chips from the US." It will provide investors with an in-depth look at what to expect regarding AI over the next few years. Related: JD Vance offers jolting preview of U.S. AI policy Whenever he discusses AI policy under Trump, Vance highlights the need for less regulation. Therefore, the AI Diffusion rule will likely mean fewer rules for companies to follow, including where AI chips can be sold. Such a decision would be in stark contrast to the AI Diffusion Rule that former President Joe Biden implemented before leaving office in January 2025. It included severe curbs on AI chip exports and garnered significant backlash from companies in the space, including Nvidia. Wall Street seems optimistic about what this new AI policy framework will mean for Nvidia. One Bank of America BAC strategist describes the date as "liberation day," as it will likely be the growth-driving catalyst the company needs. Per Business Insider: Nvidia stock isn't responding well to this news so far, but that's likely because of the current AI market slump noted by Bank of America. Fellow leading AI chip stocks such as Broadcom (AVGO) and Advanced Micro Devices (AMD) have also spent the past week trending downward amid unstable market conditions. Several AI leaders are considering a deal that could save IntelFormer big tech CEO issues grave warning on the impact of tariffsMicrosoft takes action to fight for major tech policy change Bank of America's team notes that this volatility may continue as May 15 draws closer due to rising geopolitical concerns. However, its analysts are confident that Nvidia will benefit from the new framework, maintaining a bullish $200 price target that implies 75% upside potential. When a booming new market like AI loses momentum, some investors will invariably speculate that a bubble is slowly starting to burst. However, the upcoming AI Diffusion Rule release could provide the sector with the catalyst it needs to start making up the ground it has lost in 2025. Related: AI CEO issues grave warning about the future of Nvidia With environmental curbs from Beijing's government posing new concerns, Nvidia is facing an uphill battle with momentum working against it. But Wall Street isn't souring on its growth prospects, despite its recent struggles. Thirty-nine out of forty-two analysts currently rate it as a Buy. Right now, all eyes are on Nvidia as investors look to the AI leader to help pull the sector out of its slump. Investing Ed Ponsi believes that it can do just that, though. "The tech market needs leadership," he reports. "With Nvidia in the dumps, the Nasdaq 100 will have a difficult time regaining its mojo. But if the chipmaker can break a few technical barriers, tech stocks could resume their ascent." If investors can wait until May 15, or a date close to it, Nvidia may pull it off. It will likely help lead the charge as AI chipmakers shake off the current slump and capitalize off a regulatory environment that favors them. Related: Veteran fund manager unveils eye-popping S&P 500 forecast The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
