An anonymous reader quotes a report from Reuters: The United States' practice of installing location trackers in chip shipments at risk of diversion to China reflects the "instincts of a surveillance empire," China's state-run media outlet Xinhua said in a commentary published on Friday. Reuters reported earlier this week that U.S. authorities had secretly placed location tracking devices in targeted shipments of advanced chips to detect diversions to China, which is under U.S. curbs for advanced chip exports. The Xinhua commentary, titled "America turns chip trade into a surveillance game," cited "reports" that Washington had embedded such trackers, accusing the United States of running "the world's most sprawling intelligence apparatus." [...] In its commentary, Xinhua accused the U.S. government of seeing its trading partners as "rivals to be tripped up or taken down," adding that "if U.S. chips are seen as Trojan horses for surveillance, customers will look elsewhere." Further reading: China Urges Firms To Avoid Nvidia H20 Chips After Trump Resumes Sales

Intel's stock jumped 7% after reports that the Trump administration is considering taking a stake in the struggling chipmaker to support U.S.-based manufacturing. CNBC reports: Intel is the only U.S. company with the capability to manufacture the fastest chips on U.S. shores, although rivals including Taiwan Semiconductor Manufacturing Company and Samsung also have U.S. factories. President Donald Trump has called for more chips and high technology to be manufactured in the U.S. The government's stake would help fund factories that Intel is currently building in Ohio, according to the report.

DeepSeek's development of its next-gen R2 AI model was severely delayed after months of failed training attempts on Huawei's Ascend chips, which suffered from unstable hardware, slow interconnects, and immature software. The Register reports: Following the industry rattling launch of DeepSeek R1 earlier this year, the Chinese AI darling faced pressure from government authorities to train the model's successor on Huawei's homegrown silicon, three unnamed sources have told the Financial Times. But after months of work and the help of an entire team of Huawei engineers, unstable chips, glacial interconnects, and immature software proved insurmountable for DeepSeek, which was apparently unable to complete a single successful training run. The failure, along with challenges with data labeling, ultimately delayed the release of DeepSeek R2 as the company started anew, using Nvidia's H20 GPUs instead. The company has reportedly relegated Huawei's Ascend accelerators to inference duty.

An anonymous reader shares a report: U.S. authorities have secretly placed location tracking devices in targeted shipments of advanced chips they see as being at high risk of illegal diversion to China, according to two people with direct knowledge of the previously unreported law enforcement tactic. The measures aim to detect AI chips being diverted to destinations which are under U.S. export restrictions, and apply only to select shipments under investigation, the people said.

They show the lengths to which the U.S. has gone to enforce its chip export restrictions on China, even as the Trump administration has sought to relax some curbs on Chinese access to advanced American semiconductors. The trackers can help build cases against people and companies who profit from violating U.S. export controls, said the people who declined to be named because of the sensitivity of the issue.

An anonymous reader quotes a report from Bloomberg: Beijing has urged local companies to avoid using Nvidia's H20 processors, particularly for government-related purposes, complicating the chipmaker's return to China after the Trump administration reversed an effective US ban on such sales. Over the past few weeks, Chinese authorities have sent notices to a range of firms discouraging use of the less-advanced semiconductors, people familiar with the matter said. The guidance was particularly strong against the use of H20s for any government or national security-related work by state enterprises or private companies, said the people, who asked not to be identified because the information is sensitive. The letters didn't, however, constitute an outright ban on H20 use, according to the people. Industry analysts broadly agree that Chinese companies still covet those chips, which perform quite well in certain crucial AI applications. President Donald Trump said Monday that the processor "still has a market" in the Asian country despite also calling it "obsolete."

Beijing's stance could limit Trump's ability to turn his export control about-face into a windfall for government coffers, a deal that highlighted his administration's transactional approach to national security policies long treated as nonnegotiable. Still, Chinese companies may not be ready to jump ship to local semiconductors. "Chips from domestic manufacturers are improving dramatically in quality, but they might not be as versatile for specific workloads that China's domestic AI industry hopes to focus on," said Homin Lee, a senior macro strategist at Lombard Odier in Singapore. Lee added that he anticipates "strong" demand for the chips the Trump administration is allowing Nvidia and AMD to sell.

Rosenblatt Securities analyst Kevin Cassidy said he doesn't anticipate that Nvidia's processor sales to China will be affected because "Chinese companies are going to want to use the best chips available." Nvidia and AMD's chips are superior to local alternatives, he said. Beijing asked companies about that issue in some of its letters, according to one of the people, posing questions such as why they buy Nvidia H20 chips over local versions, whether that's a necessary choice given domestic options, and whether they've found any security concerns in the Nvidia hardware. The notices coincide with state media reports that cast doubt on the security and reliability of H20 processors. Chinese regulators have raised those concerns directly with Nvidia, which has repeatedly denied that its chips contain such vulnerabilities.

The Financial Times reported that some Chinese companies are planning to decrease orders of Nvidia chips in response to the letters. Right now, the people said, China's most stringent chip guidance is limited to sensitive applications, a situation that bears similarities to the way Beijing restricted Tesla vehicles and Apple iPhones in certain institutions and locations over security concerns. China's government also at one point barred the use of Micron Technology Inc. chips in critical infrastructure. It's possible that Beijing may extend its heavier-handed Nvidia and AMD guidance to a wider range of settings, according to one person with direct knowledge of the deliberations, who said that those conversations are in early stages.

In an unusual arrangement to secure export licenses, Nvidia and AMD have agreed to give the U.S. government 15% of revenue from certain chip sales to China. The Associated Press reports: The Trump administration halted the sale of advanced computer chips to China in April over national security concerns, but Nvidia and AMD revealed in July that Washington would allow them to resume sales of the H20 and MI308 chips, which are used in artificial intelligence development. President Trump confirmed the terms of the unusual arrangement in a Monday press conference while noting that he originally wanted 20% of the sales revenue when Nvidia asked to sell the "obsolete" H20 chip to China. The president credited Nvidia CEO Jensen Huang for negotiating him down to 15%. "So we negotiated a little deal. So he's selling a essentially old chip," Trump said.

Nvidia did not comment about the specific details of the agreement or its quid pro quo nature, but said they would adhere to the export rules laid out by the administration. "We follow rules the U.S. government sets for our participation in worldwide markets. While we haven't shipped H20 to China for months, we hope export control rules will let America compete in China and worldwide," Nvidia wrote in a statement to the AP. "America cannot repeat 5G and lose telecommunication leadership. America's AI tech stack can be the world's standard if we race."

"Benchmarking firm Artificial Analysis found that only five of the top 15 AI models are open source," reports the Washington Post, "and all were developed by Chinese AI companies...."

"Now some American executives, investors and academics are endorsing a plan to make U.S. open-source AI more competitive." A new campaign called the ATOM Project, for American Truly Open Models, aims to create a U.S.-based AI lab dedicated to creating software that developers can freely access and modify. Its blueprint calls for access to serious computing power, with upward of 10,000 of the cutting-edge GPU chips used to power corporate AI development. The initiative, which launched Monday, has gathered signatures of support from more than a dozen industry figures. They include veteran tech investor Bill Gurley; Clement Delangue, CEO of Hugging Face, a repository for open-source AI models and datasets; Stanford professor and AI investor Chris Manning; chipmaker Nvidia's director of applied research, Oleksii Kuchaiev; Jason Kwon, chief strategy officer for OpenAI; and Dylan Patel, CEO and founder of research firm SemiAnalysis...

The lack of progress in open-source AI underscores the case for initiatives like ATOM: The U.S. has not produced a major new open-source AI release since Meta's launch of its Llama 4 model in April, which disappointed some AI experts... "A lot of it is a coordination problem," said ATOM's creator, Nathan Lambert, a senior research scientist at the nonprofit Allen Institute for AI who is launching the project in a personal capacity... Lambert said the idea was to develop much more powerful open-source AI models than existing U.S. efforts such as Bloom, an AI language model from Hugging Face, Pythia from EleutherAI, and others. Those groups were willing to take on more legal risk in the name of scientific progress but suffered from underfunding, said Lambert, who has worked at Google's DeepMind AI lab, Facebook AI Research and Hugging Face.

The other problem? The hefty cost of top-performing AI. Lambert estimates that getting access to 10,000 state-of-the-art GPUs will cost at least $100 million. But the funding must be found if American efforts are to stay competitive, he said.
The initiative's web page is seeking signatures, but also asks visitors to the site to "consider how your expertise or resources might contribute to building the infrastructure America needs."

Intel's chief executive Lip-Bu Tan has hit out at "misinformation" over his career after U.S. President Donald Trump alleged the semiconductor industry veteran was "highly conflicted" and should resign. From a report: In a letter to Intel staff published late on Thursday, Tan said that Intel was "engaging" with the Trump administration "to address the matters that have been raised and ensure they have the facts."

"There has been a lot of misinformation circulating about my past roles...I want to be absolutely clear: Over 40+ years in the industry, I've built relationships around the world and across our diverse ecosystem -- and I have always operated within the highest legal and ethical standards," Tan wrote.

Tan's move to reassure staff at Intel, the only US-headquartered company capable of manufacturing advanced chips, came hours after Trump had demanded his resignation in a post on Truth Social. Trump did not detail Tan's alleged conflicts of interest but the U.S. president's broadside followed a letter from Tom Cotton, the Republican head of the Senate intelligence committee, to Intel's chair expressing "concern about the security and integrity of Intel's operations" and Tan's ties to China.

EreIamJH writes: German tech newsletter Notebookcheck is reporting that the unexpectedly high 20% tariff the U.S. recently imposed on Taiwan is intended to pressure TSMC to buy a 49% minority stake in Intel -- including an IP transfer and to spend $400 billion in the U.S., in addition to the $165 billion previously planned.

Without providing specifics, President Trump said on Wednesday that he will impose a 100% tariff on imports of semiconductors and chips, but not for companies that are "building in the United States." CNBC reports: "We're going to be putting a very large tariff on chips and semiconductors," Trump said in the Oval Office on Wednesday afternoon. "But the good news for companies like Apple is if you're building in the United States or have committed to build, without question, committed to build in the United States, there will be no charge," he said. "So in other words, we'll be putting a tariff on of approximately 100% on chips and semiconductors. But if you're building in the United States of America, there's no charge." The remarks follow a recently announced commitment by Apple to invest another $100 billion in the U.S. over the next four years to boost manufacturing in the U.S.

Nvidia's chief security officer has published a blog post insisting that its GPUs "do not and should not have kill switches and backdoors." From a report: It comes amid pressure from both sides of the Pacific, with some US lawmakers pushing Nvidia to grant the government backdoors to AI chips, while Chinese officials have alleged that they already exist.

David Reber Jr.'s post seems pointedly directed at US lawmakers. In May a bipartisan group introduced the Chip Security Act, a bill that would require Nvidia and other manufacturers to include tracking technology to identify when chips are illegally transported internationally, and leaves the door open for further security measures including remote kill switches. While Nvidia is expecting to be granted permits to once again sell certain AI chips in China, its most powerful hardware is still under strict US export controls there and elsewhere.

An anonymous reader quotes a report from Ars Technica: Apple released a new developer beta build of macOS 26 Tahoe today, and it came with another big update for a familiar icon. The old Macintosh HD hard drive icon, for years represented by a facsimile of an old spinning hard drive, has been replaced with something clearly intended to resemble a solid-state drive (the SSD in your Mac actually looks like a handful of chips soldered to a circuit board, but we'll forgive the creative license).

The Macintosh HD icon became less visible a few years back, when new macOS installs stopped showing your internal disk on the desktop by default. It has also been many years since Apple shifted to SSDs as the primary boot media for new Macs. It's not clear why the icon is being replaced now, instead of years ago -- maybe the icon had started clicking, and Apple just wanted to replace it before it suffered from catastrophic icon failure -- but regardless, the switch is logical (this is a computer storage pun). Apple's iconic Macintosh HD hard drive icon was first introduced in a 2000 Mac OS X beta and remained largely unchanged for over two decades, with only subtle updates in 2012 and 2014.

The first SSD-equipped Mac was in 2008, "when the original MacBook Air came out," notes Ars. "By the time 'Retina' Macs began arriving in the early 2010s, SSDs had become the primary boot disk for most of them; laptops tended to be all-SSD, while desktops could be configured with an SSD or a hybrid Fusion Drive that used an SSD as boot media and an HDD for mass storage. Apple stopped shipping spinning hard drives entirely when the last of the Intel iMacs went away."

According to two sources Reuters spoke with, Intel is struggling with low yields in its next-gen 18A chip manufacturing process for its next PC chip, Panther Lake. Internal data suggests the company is far from reaching commercially viable production levels, leading some insiders to describe the effort as a "Hail Mary." Reuters reports: For months, Intel has promised investors it would increase manufacturing using a process it calls 18A. It spent billions of dollars developing 18A, including the construction or upgrades of several factories, with the goal of challenging Taiwan's chipmaking heavyweight, TSMC. Intel wants to round out its business designing chips that it largely makes in-house and TSMC helps it produce, with a contract manufacturing business that can compete with this key supplier. But whether Intel revives advanced chip production in the U.S. and gets its contract foundry on solid footing depends on closing the technology gap with TSMC. Early tests disappointed customers last year, but Intel has said its 18A is on track to make its "Panther Lake" laptop semiconductors at high volume starting in 2025, which include next-generation transistors and a more efficient way to deliver power to the chip.

The chipmaker has hoped that producing such an advanced in-house chip would grow external interest in its foundry, at a time when new CEO Lip-Bu Tan has explored a major shift to course-correct that fledgling business, Reuters previously reported. Yet only a small percentage of the Panther Lake chips printed via 18A have been good enough to make available to customers, said the two people, who were briefed on the company's test data since late last year. The sources spoke on condition of anonymity because Intel did not authorize them to disclose such information. This percentage figure, known as yield, means Intel may struggle to make its high-end laptop chip profitably in the near future. [...]

Intel in the past has aimed for a yield north of 50% before ramping production because starting any earlier risked damaging its profit margin, three of the sources said. Intel typically does not make the lion's share of its profit until yields reach roughly 70% to 80%, key for a chip as small as Panther Lake where many defects would make it a tough sell, the three people said. Profit also flows from market expansions and building up factory output, Intel said. An immense yield increase would be a tall task by Panther Lake's fourth-quarter launch, the two people with knowledge of Intel's manufacturing operation said. But without such a jump, Intel may have to sell some chips at a lower profit margin or at a loss, the two sources briefed on test data said. The company has warned it could exit leading-edge manufacturing entirely if it does not land external business for 14A, which is 18A's next-generation successor.

Standard DDR4 DRAM prices doubled between May and June 2025, with 8-gigabit units reaching $4.12 and 4-gigabit units hitting $3.14 -- the latter's highest level since July 2021, according to electronics trading companies cited by Nikkei Asia. The unprecedented single-month doubling follows speculation that Chinese manufacturer ChangXin Memory Technologies has halted DDR4 production to shift factories toward DDR5 memory for AI applications.

DDR4 currently comprises 60% of desktop PC memory while DDR5 accounts for 40%, per Tokyo-based BCN research. Samsung Electronics, SK Hynix, and Micron Technology controlled 90% of the global DRAM market in Q2 2025.

Beijing has summoned Nvidia over alleged security issues with its chips, in a blow to the US company's push to revive sales in the country after Washington granted approval for the export of a made-for-China chip. From a report: China's cyber regulator on Thursday said it had held a meeting with Nvidia over what it called "serious security issues" with the company's artificial intelligence chips.

It said US AI experts had "revealed that Nvidia's computing chips have location tracking and can remotely shut down the technology." The Cyberspace Administration of China requested that Nvidia explain the security problems associated with the H20 chip, which was designed for the Chinese market to comply with US export restrictions, and submit documentation to support their case.

A Nature survey of more than 1,100 physicists reveals fundamental disagreements about quantum mechanics' relationship to reality, despite the theory's century-long track record as one of science's most successful frameworks. The survey, conducted to mark quantum mechanics' 100th anniversary, found 36% of researchers favor the Copenhagen interpretation while 17% prefer epistemic approaches that treat quantum states as information rather than physical reality.

Another 15% support the many-worlds interpretation. Researchers split evenly on whether a boundary exists between quantum and classical worlds -- 45% said yes, 45% said no. When asked about the wavefunction's nature, 47% called it a mathematical tool while 36% considered it a representation of physical reality. Only 24% of respondents expressed confidence their chosen interpretation was correct, with others viewing their preference as merely adequate or useful in certain circumstances.

The survey contacted over 15,000 researchers whose recent papers involved quantum mechanics, plus attendees of a centenary meeting on Heligoland island. Despite quantum mechanics enabling technologies from computer chips to medical imaging, physicists remain divided on the physical reality underlying the mathematics.

An anonymous reader quotes a report from CNBC: Samsung Electronics has entered into a $16.5 billion contract for supplying semiconductors to Tesla, based on a regulatory filing by the South Korean firm and Tesla CEO Elon Musk's posts on X. The memory chipmaker, which had not named the counterparty, mentioned in its filing that the effective start date of the contract was July 26, 2025 -- receipt of orders -- and its end date was Dec. 31, 2033. However, Musk later confirmed in a reply to a post on social media platform X that Tesla was the counterparty.

He also posted: "Samsung's giant new Texas fab will be dedicated to making Tesla's next-generation AI6 chip. The strategic importance of this is hard to overstate. Samsung currently makes AI4.TSMC will make AI5, which just finished design, initially in Taiwan and then Arizona. Samsung agreed to allow Tesla to assist in maximizing manufacturing efficiency. This is a critical point, as I will walk the line personally to accelerate the pace of progress," Musk said on X, and suggested that the deal with Samsung could likely be even larger than the announced $16.5 billion.

Samsung earlier said that details of the deal, including the name of the counterparty, will not be disclosed until the end of 2033, citing a request from the second party "to protect trade secrets," according to a Google translation of the filing in Korean on Monday. "Since the main contents of the contract have not been disclosed due to the need to maintain business confidentiality, investors are advised to invest carefully considering the possibility of changes or termination of the contract," the company said.

Long-time Slashdot reader hackingbear writes: China's Huawei Technologies showed off an AI computing system on Saturday that can rival Nvidia's most advanced offering, even though the company faces U.S. export restrictions. The CloudMatrix 384 system made its first public debut at the World Artificial Intelligence Conference (WAIC), a three-day event in Shanghai where companies showcase their latest AI innovations, drawing a large crowd to the company's booth. The CloudMatrix 384 incorporates 384 of Huawei's latest 910C chips, optically connected through an all-to-all topology, and outperforms Nvidia's GB200 NVL72 on some metrics, which uses 72 B200 chips, according to SemiAnalysis. A full CloudMatrix system can now deliver 300 PFLOPs of dense BF16 compute, almost double that of the GB200 NVL72. With more than 3.6x aggregate memory capacity and 2.1x more memory bandwidth, Huawei and China "now have AI system capabilities that can beat Nvidia's," according to a report by SemiAnalysis.

The trade-off is that it takes 4.1x the power of a GB200 NVL72, with 2.5x worse power per FLOP, 1.9x worse power per TB/s memory bandwidth, and 1.2x worse power per TB HBM memory capacity, but SemiAnalysis noted that China has no power constraints only chip constraints. Nvidia had announced DGX H100 NVL256 "Ranger" Platform [with 256 GPUs], SemiAnalysis writes, but "decided to not bring it to production due to it being prohibitively expensive, power hungry, and unreliable due to all the optical transceivers required and the two tiers of network. The CloudMatrix Pod requires an incredible 6,912 400G LPO transceivers for networking, the vast majority of which are for the scaleup network."


Also at this event, Chinese e-commerce giant Alibaba released a new flagship open-source reasoning model Qwen3-235B-A22B-Thinking-2507 which has "already topped key industry benchmarks, outperforming powerful proprietary systems from rivals like Google and OpenAI," according to industry reports. On the AIME25 benchmark, a test designed to evaluate sophisticated, multi-step problem-solving skills, Qwen3-Thinking-2507 achieved a remarkable score of 92.3. This places it ahead of some of the most powerful proprietary models, notably surpassing Google's Gemini-2.5 Pro, while Qwen3-Thinking secured a top score of 74.1 at LiveCodeBench, comfortably ahead of both Gemini-2.5 Pro and OpenAI's o4-mini, demonstrating its practical utility for developers and engineering teams.

AMD CEO Lisa Su said that chips produced at TSMC's new Arizona plant will cost 5-20% more than those made in Taiwan, but emphasized that the premium is worth it for supply chain resilience. Bloomberg reports: AMD expects its first chips from TSMC's Arizona facilities by the end of the year, Su said. The extra expense is worth it because the company is diversifying the crucial supply of chips, Su said in an interview with Bloomberg Television following her onstage appearance. That will make the industry less prone to the type of disruptions experienced during the pandemic. "We have to consider resiliency in the supply chain," she said. "We learned that in the pandemic."

TSMC's new Arizona plant is already comparable with those in Taiwan when it comes to the measure of yield -- the amount of good chips a production run produces per batch -- Su told the audience at the forum.

An anonymous reader quotes a report from Tom's Hardware: At the 2025 RISC-V Summit in China, Nvidia announced that its CUDA software platform will be made compatible with the RISC-V instruction set architecture (ISA) on the CPU side of things. The news was confirmed during a presentation during a RISC-V event. This is a major step in enabling the RISC-V ISA-based CPUs in performance demanding applications. The announcement makes it clear that RISC-V can now serve as the main processor for CUDA-based systems, a role traditionally filled by x86 or Arm cores. While nobody even barely expects RISC-V in hyperscale datacenters any time soon, RISC-V can be used on CUDA-enabled edge devices, such as Nvidia's Jetson modules. However, it looks like Nvidia does indeed expect RISC-V to be in the datacenter.

Nvidia's profile on RISC-V seems to be quite high as the keynote at the RISC-V Summit China was delivered by Frans Sijsterman, who appears to be Vice President of Hardware Engineering at Nvidia. The presentation outlined how CUDA components will now run on RISC-V. A diagram shown at the session illustrated a typical configuration: the GPU handles parallel workloads, while a RISC-V CPU executes CUDA system drivers, application logic, and the operating system. This setup enables the CPU to orchestrate GPU computations fully within the CUDA environment. Given Nvidia's current focus, the workloads must be AI-related, yet the company did not confirm this. However, there is more.

Also featured in the diagram was a DPU handling networking tasks, rounding out a system consisting of GPU compute, CPU orchestration, and data movement. This configuration clearly suggests Nvidia's vision to build heterogeneous compute platforms where RISC-V CPU can be central to managing workloads while Nvidia's GPUs, DPUs, and networking chips handle the rest. Yet again, there is more. Even with this low-profile announcement, Nvidia essentially bridges proprietary CUDA stack to an open architecture, one that seems to develop fast in China. Yet, being unable to ship flagship GB200 and GB300 offerings to China, the company has to find ways to keep its CUDA thriving.