The Trump administration is expanding controls on access to advanced U.S. AI chips, aiming to close a cloud-computing workaround to restrictions on direct sales to China. The Commerce Department’s Bureau of Industry and Security says overseas data centers serving customers in China, Russia, Iran and Cuba must identify end users and check for diversion. The rule also requires case-by-case licenses for advanced chip exports, reexports or transfers to Indonesia, Malaysia, Singapore, Thailand and Brazil. Covered facilities are described as those able to provide more than 12 million total processing performance—roughly the power of 750 Nvidia H100 or H200 chips. Reporting differs on when the measures take effect: some accounts say immediately, while others describe a 90-day period for covered data centers to obtain licenses. Chips for qualifying U.S.-headquartered companies are exempt, as are commercial recommendation systems used in e-commerce and social media. The rule rescinds the Biden-era AI Diffusion Rule, which the department says was never enforced, and follows scrutiny of Nvidia and Singapore-based data-center operator Megaspeed International. Megaspeed says it conducts lawful business and has hired outside counsel to investigate. Reports also say Nvidia, Intel and Micron shares fell after the announcement.
Nvidia faces separate pressure over its leadership and investment strategy. An open letter urging Jensen Huang to return as CEO has drawn more than 165,000 signatures, including Elon Musk, Epic Games’ Tim Sweeney and Uber co-founder Travis Kalanick; more than 7,500 signatories say they work at Nvidia. The company’s board named CFO Colette Kress interim CEO after Huang resigned, and the controversy has reportedly erased about $600 billion from Nvidia’s value. Kress is expected to present the latest results and, according to one report, may rein in computing deals and startup investments Huang backed. More broadly, Mizuho Securities strategist Sean Darby calls the pullback in AI stocks a healthy reset as tighter credit and weaker semiconductor pricing power slow earnings growth, though he expects semiconductor stocks to remain supported through year-end.
AI-related projects are also expanding beyond chips. Databricks plans to invest about $3 billion in Europe and the Middle East over three years, its largest regional commitment outside North America. It expects its EMEA workforce—currently more than 2,000 people—to grow by about 50%, while adding engineering and infrastructure resources for Lakebase, Genie and Unity Gateway. Mercari and Suruga-ya plan to launch a U.S. retail venture in December 2026, with ownership split 51% to Mercari and 49% to Suruga-ya. Starting in Los Angeles, the venture will use AI for appraisal, listings, pricing, inventory and search, drawing on Suruga-ya’s catalog of more than approximately 30 million products. Separately, Solflare is applying AI security reviews across planning, code review and production monitoring. Its MoneyGram integration lets users cash out USDC at participating locations in more than 170 countries and territories; eligible users pay no fees through December 31, 2026. Solflare reports more than 4 million active users and over $15 billion in self-custodied assets.
Google’s education policy is drawing criticism in New Jersey: schools that want to disable AI features on student accounts must also turn off Google Search. Gov. Mikie Sherrill’s office says schools should decide what is right for students, while Google says AI Mode and AI Overviews are intended to guide students through problems and reflect educator feedback. About 150 million students and teachers use Google Classroom nationwide, and students in most of New Jersey’s more than 600 school districts use Google products. The article also notes a lawsuit alleging Google’s Gemini contributed to a 13-year-old girl’s psychological crisis. In other AI applications, Texas A&M Forest Service reports that its tree-canopy mapping model achieved 94.5% average accuracy across 295 Texas municipalities using aerial imagery and LiDAR-based training data. KForge 2.0, meanwhile, brings AI agents that launch and trade tokens to Kaspa’s public testnet; six agents are active, but results are test-only, not live mainnet performance.
Rising demand for GPUs and high-bandwidth memory is also raising concerns about counterfeit or misrepresented hardware entering supply chains. The article recommends supplier and end-user checks, verification of intended use and export-license requirements, and layered inspections such as surface analysis, X-ray imaging and, where necessary, decapsulation—particularly because failures in vehicles, medical equipment, aircraft or space systems can have serious consequences.
Key Takeaways
- The U.S. is tightening controls on remote access to advanced AI chips through foreign data centers, with new licensing requirements covering Indonesia, Malaysia, Singapore, Thailand and Brazil.
- Covered overseas data centers serving customers in China, Russia, Iran and Cuba must identify end users; the threshold cited is more than 12 million total processing performance, about 750 Nvidia H100 or H200 chips.
- The rule rescinds the Biden-era AI Diffusion Rule. Reports vary on whether it took effect immediately or gives covered operators 90 days to secure licenses.
- Nvidia’s board named CFO Colette Kress interim CEO after Jensen Huang resigned; an open letter calling for Huang’s return has drawn more than 165,000 signatures, and one report says the controversy has cost Nvidia about $600 billion in value.
- Kress is expected to curb some Nvidia computing deals and startup investments, while Mizuho strategist Sean Darby says slower AI earnings growth and weaker semiconductor pricing power are contributing to a healthy stock-market reset.
- Databricks plans roughly $3 billion of investment across Europe and the Middle East over three years and expects its regional workforce, now above 2,000, to grow about 50%.
- Mercari and Suruga-ya plan a U.S. venture for December 2026, using AI for appraisal, pricing, inventory and search and drawing on a catalog of more than approximately 30 million products.
- Solflare is deploying AI security reviews and integrating MoneyGram cash-outs for USDC in more than 170 countries and territories; eligible users pay no fees through December 31, 2026.
- Google’s policy requires schools to disable Search as well as AI features if they want AI off student accounts; New Jersey officials object, while Google says its AI tools guide learning.
- Texas A&M’s AI tree-canopy model reports 94.5% average accuracy across 295 municipalities; KForge 2.0 has six AI token-trading agents on Kaspa testnet, where results do not represent live market performance.
- AI hardware shortages heighten counterfeit-part risks, prompting calls for supplier checks, end-use verification and layered component inspections.
Trump moves to close AI chip export loophole
The Trump administration is tightening export rules to limit China’s access to advanced AI chips, including Nvidia chips. China has been able to get around existing restrictions by renting computing power from data centers in other countries. A new rule, issued Monday and set to take effect in 90 days, adds restrictions for data centers in some countries. They must check customers carefully to make sure they are not helping anyone evade an export ban.
US adds AI chip reexport licenses for five countries
The Commerce Department’s Bureau of Industry and Security introduced a rule to stop countries from using data centers in other countries to get around US limits on sensitive technology. It adds license requirements for chip exports and reexports to five countries, alongside existing rules. The requirements do not apply to chips for companies headquartered in the United States, or with a US-based parent company, if they meet reporting rules. The department will review applications individually, and the rule rescinds the Biden-era AI Diffusion Rule, which it says was never enforced. The rule takes effect when it appears in the Federal Register, and public comments are due 60 days later.
US rule targets remote access to advanced AI chips
The US Commerce Department said on Monday, November 16, that it issued a rule to prevent third countries from remotely accessing advanced artificial intelligence chips. The rule aims to stop existing restrictions on sensitive US technology from being bypassed through data centers in other countries. The department said the change could limit China’s access to advanced technology.
US tightens rules on remote access to AI computing
The Trump administration introduced export controls to stop companies, especially in China, from accessing US AI computing power through data centers in other countries. The rule covers data centers and networks that can provide more than 12 million total processing performance, roughly the computing power of 750 Nvidia H100 or H200 chips. It also requires licenses for advanced chip exports, reexports, or transfers to Malaysia, Singapore, Indonesia, Thailand, and Brazil, which the Commerce Department says pose a higher-than-normal risk of diversion. Data center operators covered by the rule must get licenses within 90 days, while commercial recommendation systems used in e-commerce or social media are exempt. The rule also cancels the Biden administration’s January 2025 AI Diffusion Rule.
New US licenses target remote access to AI chips
The Trump administration’s new export controls aim to prevent companies in China and other restricted countries from remotely using AI infrastructure in other countries. The rule covers data centers or networks that can provide more than 12 million total processing performance, about the computing power of 750 Nvidia H100 or H200 chips. It also requires licenses for advanced chip exports, reexports, or transfers to Malaysia, Singapore, Indonesia, Thailand, and Brazil, but exempts chips intended for US-based companies. Covered data center operators must obtain licenses within 90 days, and commercial recommendation systems such as e-commerce and social media services are not covered. The rule also rescinds the Biden administration’s AI Diffusion Rule from January 2025.
US Moves to Close Loophole in China’s Access to AI Chips
The Trump administration announced a rule to limit China’s ability to use advanced U.S. AI chips through data centers overseas. The rule requires overseas data centers serving companies in China, Russia and Iran to check their customers. It also requires export licenses for advanced chips sent to Indonesia, Malaysia, Singapore, Thailand and Brazil. The change takes effect in 90 days and comes days before President Trump’s planned trip to China to meet Xi Jinping. Chip stocks fell Monday, including Nvidia, Intel and Micron.
Commerce Department Seeks to Limit Remote Access to AI Chips
The Commerce Department’s Bureau of Industry and Security issued a temporary rule on Monday to restrict remote access to advanced semiconductors. The rule aims to stop foreign nations from using data centers in third countries to get around U.S. technology restrictions. It requires licenses to export, re-export or transfer advanced semiconductors to Brazil, Indonesia, Malaysia, Singapore and Thailand. The department said cloud computing has made it easier to access U.S. AI computing power from abroad.
US Rule Restricts Remote Access to Advanced AI Chips
The U.S. Commerce Department introduced a temporary rule restricting remote access to advanced AI chips and server hardware. Exporting or re-exporting advanced chips to Brazil, Singapore, Indonesia, Malaysia and Thailand now requires a license. The department said the rule is meant to stop data centers in third countries from helping users get around restrictions on sensitive U.S. technology. The action follows increased attention to chip shipments after an alleged espionage episode tied to Nvidia.
New US Rules Target China’s Indirect Access to AI Chips
The Trump administration issued rules to restrict remote access to advanced U.S. computing chips and servers, aiming to limit China’s ability to use them to train AI models. The rules took effect immediately, according to the Commerce Department’s Bureau of Industry and Security. Overseas data centers serving companies in China, Russia, Iran and Cuba must identify their end users. Sales of chips to Indonesia, Malaysia, Singapore, Thailand and Brazil will also face extra case-by-case licensing requirements because these countries pose a higher-than-normal risk of diversion.
US Rules Limit China’s Indirect Access to American AI Technology
The Trump administration issued rules to restrict remote access to advanced U.S. computing chips and servers, aiming to limit China’s ability to use them to train AI models. The rules followed the resignation of Nvidia co-founder Jensen Huang as CEO amid scrutiny of his ties to a Chinese executive who once led Singapore-based Megaspeed International. The company is suspected of diverting advanced Nvidia chips to China and provides computing power to Chinese companies through chips in Southeast Asian data centers, including facilities in Malaysia. Overseas data centers serving companies in China, Russia, Iran and Cuba must identify end users, while chip sales to Indonesia, Malaysia, Singapore, Thailand and Brazil face extra case-by-case licensing requirements.
US tightens rules on remote access to advanced AI chips
The Trump administration issued rules to restrict remote access to advanced U.S. chips and servers, aiming to stop China from using them to train AI models. The rules took effect immediately and require overseas data centers serving companies in China, Russia, Iran and Cuba to identify their end users. Sales of chips to Indonesia, Malaysia, Singapore, Thailand and Brazil will face extra case by case licensing because officials see a higher risk of diversion. The U.S. already bans sales of its most advanced AI chips to China, while lawmakers have warned that cloud computing can help companies bypass those limits.
New US rules target cloud access to advanced AI chips
The Trump administration issued rules on Monday to restrict remote access to advanced U.S. chips and servers, aiming to limit China’s use of them to train AI models. The rule followed reports about Nvidia co-founder Jensen Huang and Megaspeed International, a company suspected of diverting advanced Nvidia chips to China through data centers in Southeast Asia. Overseas data centers serving companies in China, Russia, Iran and Cuba must identify their end users. Chip sales to Indonesia, Malaysia, Singapore, Thailand and Brazil will need extra case by case licenses because of a higher risk of diversion. The rule took effect immediately, and a bill to extend export controls to remote access has passed the House.
US rule targets cloud workarounds for AI chip export controls
The Commerce Department’s Bureau of Industry and Security issued a rule to prevent companies from bypassing physical export limits by using cloud infrastructure. Overseas data centers will face new requirements to limit access to services that use U.S. technology, especially when serving companies headquartered in higher risk countries such as China, Russia and Iran. A license will also be required to export advanced chips to companies headquartered in Indonesia, Malaysia, Singapore, Thailand or Brazil. The requirements are set to take effect in 90 days.
Commerce adds new limits on overseas AI chip sales
The Commerce Department announced new export rules to stop U.S. companies from using foreign data centers to get around limits on AI chip sales. The rules restrict exports, reexports and transfers to data centers in countries under U.S. arms embargoes, including China, Cuba, Iran, Libya, North Korea and Sudan. Data centers, foreign foundries and chip exporters must check customers, including their indirect ownership or control by people in those countries. New licenses are also required for advanced AI chip sales to Indonesia, Malaysia, Singapore, Thailand and Brazil, with companies given 90 days to obtain them. Commerce Secretary Howard Lutnick identified those five countries as having a higher than normal risk of chip diversion.
US seeks to stop China from renting access to AI chips
The Trump administration is moving to restrict foreign data centers that let China remotely use high end chips needed to train AI systems. The Commerce Department said the change closes a loophole that allowed countries such as China, Russia and Iran to rent access to advanced chips despite U.S. export limits. The move follows an Nvidia scandal involving allegations that China accessed the company’s AI chips without proper oversight. The United States is concerned that access could help China advance its AI research and compete in the global market. The article says officials expect to finalize the rule in the coming weeks.
US requires licenses for AI chip sales to five countries
The Commerce Department now requires data center customers in Indonesia, Malaysia, Singapore, Thailand and Brazil to get a special license before buying advanced AI chips from US companies such as Nvidia and AMD. The rule also requires licenses for overseas data centers that provide significant computing power to companies based in countries including China, Russia and Cuba. Officials said the change addresses concerns that Chinese companies could remotely use chips in foreign data centers despite limits on direct sales to China. The rule also formally cancels the Biden administration’s AI Diffusion Rule.
US tightens rules on remote access to advanced AI chips
The Commerce Department’s Bureau of Industry and Security introduced a rule to limit foreign companies’ remote access to advanced US computing technology through overseas data centers. Since October 2022, US companies have generally needed a license to sell certain advanced computing chips to China, including Hong Kong and Macau, but the new measure targets remote access that could bypass those limits. The announcement came amid attention on Nvidia and Megaspeed, a Singapore-based data center company that also operates facilities in Malaysia. Megaspeed said it conducts lawful business and has hired outside legal counsel to investigate. The rule follows the Trump administration’s decision not to enforce a Biden administration framework issued in January 2025.
Nvidia shares face pressure as supporters urge Jensen Huang’s return
An open letter calling for Jensen Huang to return as Nvidia CEO has attracted more than 165,000 signatures, including those of Elon Musk, Epic Games chief Tim Sweeney, Y Combinator president Garry Tan and Uber co-founder Travis Kalanick. More than 7,500 signatories say they work at Nvidia, about 18% of the company’s workforce if accurate. Nvidia’s board named chief financial officer Colette Kress interim CEO after Huang resigned last Thursday. The controversy has cut about $600 billion from Nvidia’s value so far. Kress is due to present the company’s latest financial results the following afternoon.
Nvidia expected to pull back on startup funding under interim CEO
Nvidia’s interim CEO, Colette Kress, is expected to reduce the company’s spending on computing deals and startup investments. In his final months as CEO, Jensen Huang aggressively backed these deals, often with little scrutiny, according to the article. Kress, Nvidia’s finance chief before taking the role, had already been cautious about some deals Huang made. Startups that relied on Nvidia’s support now expect the company to pull back.
Solflare adds AI security checks and expands MoneyGram link
Solflare CEO and co-founder Vidor Gencel said the company uses AI-powered security reviews across its development and production work. Solflare has also adopted a “bunker mode” process that requires extensive internal approvals before deployment. Gencel said AI can find and connect vulnerabilities that human reviewers may miss. The Solana-based wallet is adding real-world assets and financial services, including trading, earning and spending features. Solflare also launched a privacy mode with Helios and is integrating with MoneyGram so users in more than 170 countries can convert USDC to cash at MoneyGram locations.
Solflare combines AI security work with MoneyGram cash outs
Solflare says it is adding AI security reviews across product planning, code reviews and production monitoring. CEO Vidor Gencel said AI can help find weaknesses and connect smaller flaws into larger attack paths. After entering “bunker mode” in early November 2026, Solflare announced MoneyGram Ramps on November 16 during the Breakpoint event. The service lets users cash out USDC into local currency at participating MoneyGram locations in more than 170 countries and territories, with no fees for eligible users through December 31, 2026. Solflare reports more than 4 million active users and over $15 billion in self-custodied assets, while MoneyGram operates more than 480,000 locations.
Mercari and Suruga-ya plan AI-powered US retail venture
Mercari and Suruga-ya plan to launch a US venture in December 2026, with Mercari holding 51% and Suruga-ya holding 49%. The company plans to begin US operations in Los Angeles, selling products through Mercari US and the Mercari Global App, though launch dates are not set. It will use AI in processes such as appraisal, product registration, pricing, inventory management and product search. The venture will combine Mercari’s US distribution and technology with Suruga-ya’s specialist inventory and retail expertise, including a catalog of more than approximately 30 million products. The companies aim to use US transaction data to improve search, recommendations and authenticity checks.
Strategist sees a healthy reset in AI stocks as earnings growth cools
Sean Darby of Mizuho Securities called the pullback in AI-related stocks a healthy reset. He said tighter credit conditions and weaker semiconductor pricing power are slowing earnings growth. Darby expects semiconductor stocks to remain supported through the end of the year. However, he believes the strongest growth in AI earnings is likely behind the market.
AI hardware shortages raise supply chain safety risks
Rising demand for GPUs, high-bandwidth memory and other AI hardware is putting pressure on global supply chains and creating more chances for counterfeit parts to enter the market. Some parts are mislabeled, altered, recycled or sold as newer and higher-performing products than they are. The article warns that failures in vehicles, medical equipment, aircraft and space systems can have serious consequences, making reliable part histories important. It recommends supplier checks and layered inspections, including surface analysis, X-ray imaging and decapsulation when needed. Companies should also check suppliers, end customers, intended uses and export license needs before shipping advanced hardware.
Databricks commits $3 billion to Europe and Middle East expansion
Databricks plans to invest approximately $3 billion across Europe and the Middle East over three years, its largest regional commitment outside North America. The data and AI company expects its Europe, Middle East and Africa workforce to grow by about 50% from its current total of more than 2,000 employees. It will add resources to engineering centres in Amsterdam, Berlin, Belgrade and Aarhus and expand regional infrastructure and technical support for Lakebase, Genie and Unity Gateway. Databricks has also invested in companies including France’s Tsuga, UK partner Advancing Analytics and UK startup Applied Computing. Its regional customers include ABN Amro, Adidas, Danone, Heineken, HelloFresh, Mercedes-Benz and Nationwide.
New Google policy puts New Jersey schools between AI and search
Google says schools that want to keep AI features off student accounts must also turn off Google Search. New Jersey Gov. Mikie Sherrill’s office called the policy unfair and said local schools should decide what is right for students. Google says its AI Mode and AI Overviews will guide students through problems instead of giving quick answers, and that the changes respond to educators’ feedback. About 150 million students and teachers use Google Classroom nationwide, and students in most of New Jersey’s more than 600 school districts use Google products. Some critics worry students could use Search AI like a chatbot, and the article also reports a lawsuit alleging Google’s Gemini contributed to a 13-year-old girl’s psychological crisis.
Texas A&M model maps urban tree cover with 94.5% accuracy
Texas A&M Forest Service developed an artificial intelligence model to map tree canopy in communities across Texas. The deep learning model uses publicly available aerial images and reached an average accuracy of 94.5% across 295 municipalities. It was trained with LiDAR-based tree canopy data from four cities, helping map areas that lack LiDAR coverage. The maps can help communities plan tree planting, development and forest management, as well as track changes in tree cover. Trees can provide shade, improve air quality, manage stormwater and reduce building energy needs.
KForge 2.0 brings AI token trading agents to Kaspa testnet
KForge released version 2.0 of its decentralized token launchpad on Kaspa’s public testnet on Monday, November 16. The upgrade lets AI agents launch and trade tokens on Kaspa Layer 1, and users can run them through a Claude Desktop extension or a local terminal script. Six agents are currently trading on the testnet, with results shown on a public leaderboard. Users keep private keys locally, and the platform includes spending and withdrawal limits. The agents are not trading on mainnet, so leaderboard results reflect test conditions rather than live market performance.
Sources
- Trump closes an AI loophole
- U.S. is requiring licenses to reexport AI chips to five countries
- US issues rule to protect advanced AI chips from being accessed remotely
- US tightens AI chip export controls to deter Chinese companies from accessing compute remotely
- US tightens AI chip export controls to deter Chinese companies from accessing compute remotely
- Trump Administration Moves to Close A.I. Chip Loophole Ahead of China Trip
- After Nvidia scandal, Commerce Department moves to curb remote access to AI chips
- US Limits Remote Access to Advanced AI Chips
- Trump administration moves to block China from indirect access to American AI technology
- Trump administration moves to block China from indirect access to American AI technology
- Trump administration moves to block China from indirect access to American AI technology
- Trump administration moves to block China from indirect access to American AI tech
- New US rules aim to close loophole, restrict Chinese firms’ remote access to AI servers and infrastructure — seeks to combat cloud rental workarounds to export controls
- Commerce moves to place new restrictions on sale of AI chips abroad
- US moves to cut China’s remote access to AI chips
- U.S. closes loophole that let Chinese companies use top AI chips in foreign data centers
- U.S. tightens controls on remote access to AI technology amid Nvidia fallout
- AI sell-off resumes as Jensen Huang ‘discusses comeback as Nvidia CEO’
- Nvidia’s Days as an Open Checkbook for the AI Economy Are Numbered
- Solflare CEO says AI security reviews are essential, touts 'bunker mode' and MoneyGram deal
- Solflare CEO Vidor Gencel pitches AI security alongside MoneyGram cash-out deal
- Mercari Plans AI-Led US Venture for Japanese Pop Culture Goods
- AI trade faces a healthy reset as earnings growth cools: Strategist
- As AI Hardware Demand Surges, Supply Chain Trust Becomes a Safety Issue
- Databricks plans $3bn investment across Europe and Middle East
- Google just told schools to use AI or lose search. N.J.’s governor isn’t having it.
- Texas A&M Forest Service develops new technology to expand urban tree canopy mapping
- KForge 2.0 Launches on Kaspa Testnet, Letting AI Agents Launch and Trade Tokens
Comments
Please log in to post a comment.