Latest news with #H800
South China Morning Post
16-05-2025
- Business
- South China Morning Post
DeepSeek paper offers new details on how it used 2,048 Nvidia chips to take on OpenAI
Chinese artificial intelligence (AI) research lab DeepSeek has released a new research paper revealing in detail for the first time how it built one of the world's most powerful open-source AI systems at a fraction of the cost of its competitors. 'Insights into DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architectures', co-authored by DeepSeek founder Liang Wenfeng and released on Wednesday, attributes the start-up's breakthrough in training high-performance, cost-efficient AI systems to a hardware-software co-design approach. 'DeepSeek-V3, trained on 2,048 Nvidia H800 GPUs, demonstrates how hardware-aware model co-design can effectively address these challenges, enabling cost-efficient training and inference at scale,' the researchers wrote. DeepSeek and its hedge fund owner High-Flyer had previously stockpiled the H800, which Nvidia originally designed for the China market to comply with US export restrictions but were banned from export to to the country in 2023. The start-up's training approach stemmed from the team's awareness of hardware constraints and the 'exorbitant costs' of training large language models (LLMs) – the technology behind AI chatbots such as OpenAI's ChatGPT – according to the paper. The paper details technical optimisations that boost memory efficiency, streamline inter-chip communication, and enhance overall AI infrastructure performance – key advancements for reducing operational costs while scaling capabilities. These offer a 'practical blueprint for innovation in next-generation AI systems', the researchers said. Play DeepSeek also highlighted its use of a mixture-of-experts (MoE) model architecture, a machine-learning approach that divides an AI model into separate sub-networks, or experts, each focused on a subset of the input data while working collaboratively.
Epoch Times
15-05-2025
- Business
- Epoch Times
House Lawmakers Introduce Chip Security Act to Address Smuggling of AI Chips to China
Bipartisan House lawmakers on May 15 introduced a bill aimed at preventing the Chinese Communist Party (CCP) from accessing advanced American chips. The United States has put export controls on advanced chips since 2022, intending to cut off the CCP's access to cutting-edge and artificial intelligence (AI) technologies that would advance its military. But they have not had the intended effect Reps. John Moolenaar (R-Mich.) and Raja Krishnamoorthi (D-Ill.), chair and ranking member of the House Select Committee on the CCP, respectively, along with Reps. Rick Crawford (R-Ark.), Bill Foster (D-Ill.), Josh Gottheimer (D-N.J.), Bill Huizenga (R-Mich.), Darin LaHood (R-Ill.), and Ted Lieu (D-Calif.), point to 'mounting evidence' that the CCP has access to restricted technology. The lawmakers say that this access can enable weapons that could be used against the United States in a conflict, advance the Chinese regime's surveillance state, and supplant the U.S. tech industry's dominance in AI and other fields. 'For too long, the Chinese Communist Party has exploited weaknesses in our export control enforcement system—using shell companies and smuggling networks to divert sensitive U.S. technology that helps fuel its military advancement and extend its surveillance capabilities to further its repression,' Moolenaar said in a The Chip Security Act would require location verification for advanced AI chips, enforce mandatory reporting from chipmakers on potential diversion of their products, and task the Department of Commerce with studying additional necessary steps. Related Stories 5/14/2025 5/8/2025 Sen. Tom Cotton (R-Ark.) CCP Workarounds When Chinese AI company DeepSeek launched its free chatbot globally in January, it shook up the tech industry In February, Chinese state media reported that Huawei founder Ren Zhengfei told Chinese regime leader Xi Jinping during a closed-door meeting that China would reach 70 percent semiconductor self-sufficiency by 2028. Ren said that Huawei had made breakthroughs that meant he no longer had concerns about the obstacles U.S. export controls would pose. The House Select Committee on the CCP in April released a The United States has updated its 2022 export controls more than once, limiting more technologies and blacklisting end users, but on an entity-by-entity basis. For example, the Nvidia H800 chip, developed as a watered-down H100 chip for the Chinese market that DeepSeek publicly said it used, got added to the export control list late in 2023. Lawmakers, including Moolenaar, have warned that the CCP and its military already have workarounds to this approach, such as shell company buyers. He cautioned then-Commerce Secretary Gina Raimondo last Companies tend to oppose broad restrictions, according to records of public comments on these rules, because it can put them at risk of not knowing when they have violated these rules.
West Australian
12-05-2025
- Business
- West Australian
THE ECONOMIST: How China is still getting its hands on Nvidia's gear
Last month Jensen Huang, the boss of Nvidia, landed in Beijing with a clear message: the maker of the world's leading artificial-intelligence chips planned to 'unswervingly serve the Chinese market'. America would rather it didn't. A few days earlier the Trump administration had introduced new controls that, in effect, banned the company from selling its H20 microprocessor to China. Over the past few years America has sought to hobble its main rival in the AI race by controlling access to its advanced semiconductors. The performance of an AI processor depends mostly on two factors: computing power (how fast a chip processes data) and memory bandwidth (how quickly it moves data between processor and memory). In October 2022 the Biden administration barred sales to China of American chips that exceed a threshold on both fronts. Nvidia responded with the H800, a made-for-China model engineered to stay just under the limits. A year later, America tightened the regulations again, banning any chip with too much computing power, regardless of memory bandwidth. Nvidia's answer was the H20. The trouble for America is that restricted Nvidia chips continue to make their way into the hands of Chinese AI developers. A shadowy supply chain has emerged, designed to work around sanctions. Some customers lease access to offshore data centres; others buy chips through murky intermediaries. A new round of restrictions, due to come into effect on May 15th, aims to tighten the net further still. The new rules target not just hardware, but also the cloud providers offering access to it. Yet they will probably fail, too. To see why it is so difficult to restrict access to AI chips, consider Johor, a part of southern Malaysia once better known for its palm-oil plantations. Located just across the border from Singapore, the region has become a hub for data centres. Land and electricity are cheap, and permits easier to obtain than in the city-state. All the big American cloud providers—Amazon, Google, Microsoft and Oracle—have set up shop. According to Knight Frank, a property consultancy, Johor's total data-centre capacity (built, under construction or planned) grew from 10 megawatts in early 2021 to more than 1,500 megawatts by 2024. Johor also provides a convenient back door into China. Big Chinese firms such as ByteDance, the owner of TikTok, have rented capacity there. Leasing cloud capacity in Malaysia allows companies like it to gain access to chips that cannot be imported into China. SemiAnalysis, a consultancy, estimates that nearly half of Johor's projected data-centre capacity in 2027 will incorporate AI processors such as Nvidia's. Malaysian data-centre operators insist they comply with American export regulations and do not provide capacity to blacklisted entities. Yet workarounds are straightforward. A lawyer advising firms in the region says it is relatively easy for Chinese companies to get hold of restricted AI chips by setting up local subsidiaries. Figures on trade flows support this. Nvidia's high-end chips are produced by TSMC, the world's biggest chipmaker, in its Taiwanese factories. In the first quarter of this year Taiwan exported $3.6 billion ($5.6b AUD)-worth of graphics-processing units—the kind used to train AI models—to Malaysia, nearly matching the total for all of 2024. In March alone shipments more than tripled from the previous month to reach almost $2bn. Then there are the smugglers who traffic chips directly into China. These are typically diverted through third countries not covered by American restrictions. A source familiar with the practice says goods often pass through several jurisdictions and front companies to obscure their origin. Export papers are doctored; restricted products are mislabelled to slip past customs. Erich Grunewald of the Institute for AI Policy and Strategy, a think-tank based in San Francisco, estimates that last year smuggled American chips made up between one-tenth and one-half of China's ai-model-training capacity. Before the first round of export controls in 2022, China accounted for about 22 per cent of Nvidia's revenue. That share has since fallen to 13 per cent. At the same time, sales to Singapore—a city with few end-users—have more than doubled, and now make up nearly 18 per cent of the total, making it Nvidia's second-largest market after America. The company says the shift is routine: many clients invoice through Singapore but ship to permitted destinations. Fewer than 2 per cent of chips sold there are delivered locally. In February, however, Singaporean police arrested three men over the sale of $390m-worth of servers that incorporated Nvidia chips. Prosecutors allege these were first sent to Singaporean firms, then re-exported to Malaysia. Whether that was their final stop remains unknown. What is clearer is the incentive: demand has turned the grey market into a gold mine. According to one industry executive, banned Nvidia chips now sell at a 30-50 per cent markup through intermediaries. China is not the only destination. In October America placed several Indian companies under sanctions for re-exporting restricted chips to Russia. Among them was Shreya Life Sciences, a pharmaceutical firm based in Mumbai. According to figures from The Trade Vision, a data provider, the company exported $322m-worth of tech goods to Russia in 2024, much of it Dell servers containing Nvidia chips. All this puts Nvidia in a difficult position. The company insists it complies with American export rules. But its operations are vast: it expects to sell more than 6m AI chips this year and it sits several steps removed from the end user. Nvidia supplies processors to cloud giants such as Google and Microsoft, and to equipment-makers like Dell and Supermicro, which integrate them into servers. From there, responsibility for compliance is diffuse. Cloud providers and hardware firms are expected to vet their customers. Nvidia itself conducts periodic audits. But oversight is uneven, and servers often change hands quietly after passing initial checks. One executive at a server manufacturer says properly verifying all end users is 'practically impossible'. America's government has responded by introducing ever tighter regulations. In January the Biden administration unveiled its 'AI Diffusion Framework', which splits the world into three tiers: 18 trusted countries including Britain and Japan face no restrictions; a second tier of 120 countries including Singapore and India are subject to quotas; and a final tier including China and Russia are barred outright. The rules also prohibit American cloud providers from offering services powered by restricted chips to Chinese customers. The Trump administration is reportedly planning to tinker with the regulations before they come into effect—perhaps tying access to chips into wider trade negotiations—but it is yet to announce any changes. Any new controls will encounter familiar problems. The Bureau of Industry and Security (BIS), the agency tasked with enforcing tech export controls, is severely understaffed. Mr Grunewald notes that it has just one export-control officer responsible for all of South-East Asia and Australasia—a region central to the shadow trade in AI chips. And just as the regulatory web grows more intricate, BIS is being stretched even more: it faces a cut to its budget of 12 per cent this year. Some pundits have proposed technical solutions, such as disabling chips used in prohibited locations. Nvidia argues that such hardware-level controls would introduce dangerous vulnerabilities and are unworkable. Instead, it suggests that software-based tools could transmit limited telemetry—including information on location and system configuration—back to the company to confirm that the chips are where they are supposed to be. Even better enforcement has its limits, however. Nvidia cannot trace every chip. BIS cannot inspect every server. Smugglers will continue to find loopholes. If America wants to keep ahead of China in the AI race, it will need to innovate faster, rather than clamp down harder.
Yahoo
07-05-2025
- Business
- Yahoo
Nvidia stock dives as chipmaker sees $5.5 billion hit from 'surprise' China chip controls
Nvidia (NVDA) stock sank nearly 7% Wednesday after the AI chipmaker disclosed that it would take a $5.5 billion hit from the US government's surprise new controls on its semiconductor exports to China. Nvidia said in a regulatory filing late Tuesday night that the US government informed the company it would require a special license for exports of its H20 chips made specifically for the Chinese market to comply with US trade rules. Notably, no licenses for GPU shipments into China have ever been granted, given the US government's concern that the chips could be used to build AI supercomputers in the country, Jefferies analyst Blayne Curtis wrote in an analysis following the news that the latest rule is effectively a ban. Other Wall Street analysts noted the move was a "surprise," given a recent report from NPR that the Trump administration had backed off its plans to restrict Nvidia's H20 chips following a dinner with CEO Jensen Huang at Mar-a-Lago. Nvidia said it will incur $5.5 billion in charges in its first quarter from the latest curb. Read more: How does Nvidia make money? Jefferies' Curtis projected the company will take an even bigger hit to revenue — $10 billion in lost sales — over the coming quarters. Curtis said that's because the majority of the write-down is related to finished or partially finished goods rather than future supply agreements, meaning Nvidia will essentially have to toss out billions worth of now-unsellable chips rather than simply canceling future orders. "Banning the H20 makes little sense to us," wrote Bernstein analyst Stacy Rasgon in a note to investors early Wednesday. "H20 performance is low, well below already-available Chinese alternatives; a ban essentially simply hands the Chinese AI market over to Huawei." Nvidia CEO Jensen Huang speaks about new products as he delivers the keynote address at the GTC event in San Jose, Calif., on March 18, 2025. (Josh Edelson/AFP via Getty Images) · JOSH EDELSON via Getty Images Raymond James analyst Ed Mills wrote in his own note: "The restrictions on H20 chips comes as a surprise, given explicit approval of the product by the Biden administration and recent media reports that the U.S. government was walking back from banning the product." Nvidia rival Advanced Micro Devices (AMD) also dropped Wednesday, declining more than 7% as the company said in a filing that its chip exports to China would also be restricted, costing the chipmaker up to $800 million. Fellow chip stocks Broadcom (AVGO) and Qualcomm (QCOM) sank more than 2%, while Intel (INTC) dropped over 3%. The stocks' declines weighed on the tech-heavy Nasdaq (^IXIC), which fell 3%. Nvidia has made multiple specialized chips for China since 2022 — the A800, H800, L20, L2, and the H800's successor, H20 — to comply with ever-changing trade rules as the US looks to restrict China's access to hardware necessary to innovate AI. China accounted for $17 billion, or 13%, of Nvidia's revenue in its fiscal year 2025, Rasgon noted.
Mint
07-05-2025
- Business
- Mint
How China is still getting its hands on Nvidia's gear
Last month Jensen Huang, the boss of Nvidia, landed in Beijing with a clear message: the maker of the world's leading artificial-intelligence (AI) chips planned to 'unswervingly serve the Chinese market". America would rather it didn't. A few days earlier the Trump administration had introduced new controls that, in effect, banned the company from selling its H20 microprocessor to China. Over the past few years America has sought to hobble its main rival in the AI race by controlling access to its advanced semiconductors. The performance of an AI processor depends mostly on two factors: computing power (how fast a chip processes data) and memory bandwidth (how quickly it moves data between processor and memory). In October 2022 the Biden administration barred sales to China of American chips that exceed a threshold on both fronts. Nvidia responded with the H800, a made-for-China model engineered to stay just under the limits. A year later, America tightened the regulations again, banning any chip with too much computing power, regardless of memory bandwidth. Nvidia's answer was the H20. The trouble for America is that restricted Nvidia chips continue to make their way into the hands of Chinese AI developers. A shadowy supply chain has emerged, designed to work around sanctions. Some customers lease access to offshore data centres; others buy chips through murky intermediaries. A new round of restrictions, due to come into effect on May 15th, aims to tighten the net further still. The new rules target not just hardware, but also the cloud providers offering access to it. Yet they will probably fail, too. To see why it is so difficult to restrict access to AI chips, consider Johor, a part of southern Malaysia once better known for its palm-oil plantations. Located just across the border from Singapore, the region has become a hub for data centres. Land and electricity are cheap, and permits easier to obtain than in the city-state. All the big American cloud providers—Amazon, Google, Microsoft and Oracle—have set up shop. According to Knight Frank, a property consultancy, Johor's total data-centre capacity (built, under construction or planned) grew from 10 megawatts in early 2021 to more than 1,500 megawatts by 2024. Johor also provides a convenient back door into China. Big Chinese firms such as ByteDance, the owner of TikTok, have rented capacity cloud capacity in Malaysia allows companies like it to gain access to chips that cannot be imported into China. SemiAnalysis, a consultancy, estimates that nearly half of Johor's projected data-centre capacity in 2027 will incorporate AI processors such as Nvidia's. Malaysian data-centre operators insist they comply with American export regulations and do not provide capacity to blacklisted entities. Yet workarounds are straightforward. A lawyer advising firms in the region says it is relatively easy for Chinese companies to get hold of restricted AI chips by setting up local subsidiaries. Figures on trade flows support this. Nvidia's high-end chips are produced by TSMC, the world's biggest chipmaker, in its Taiwanese factories. In the first quarter of this year Taiwan exported $3.6bn-worth of graphics-processing units—the kind used to train AI models—to Malaysia, nearly matching the total for all of 2024 (see chart 1). In March alone shipments more than tripled from the previous month to reach almost $2bn. Then there are the smugglers who traffic chips directly into China. These are typically diverted through third countries not covered by American restrictions. A source familiar with the practice says goods often pass through several jurisdictions and front companies to obscure their origin. Export papers are doctored; restricted products are mislabelled to slip past customs. Erich Grunewald of the Institute for AI Policy and Strategy, a think-tank based in San Francisco, estimates that last year smuggled American chips made up between one-tenth and one-half of China's ai-model-training capacity. Before the first round of export controls in 2022, China accounted for about 22% of Nvidia's revenue (see chart 2). That share has since fallen to 13%. At the same time, sales to Singapore—a city with few end-users—have more than doubled, and now make up nearly 18% of the total, making it Nvidia's second-largest market after America. The company says the shift is routine: many clients invoice through Singapore but ship to permitted destinations. Fewer than 2% of chips sold there are delivered locally. In February, however, Singaporean police arrested three men over the sale of $390m-worth of servers that incorporated Nvidia chips. Prosecutors allege these were first sent to Singaporean firms, then re-exported to Malaysia. Whether that was their final stop remains unknown. What is clearer is the incentive: demand has turned the grey market into a gold mine. According to one industry executive, banned Nvidia chips now sell at a 30-50% markup through intermediaries. China is not the only destination. In October America placed several Indian companies under sanctions for re-exporting restricted chips to Russia. Among them was Shreya Life Sciences, a pharmaceutical firm based in Mumbai. According to figures from The Trade Vision, a data provider, the company exported $322m-worth of tech goods to Russia in 2024, much of it Dell servers containing Nvidia chips. All this puts Nvidia in a difficult position. The company insists it complies with American export rules. But its operations are vast: it expects to sell more than 6m AI chips this year and it sits several steps removed from the end user. Nvidia supplies processors to cloud giants such as Google and Microsoft, and to equipment-makers like Dell and Supermicro, which integrate them into servers. From there, responsibility for compliance is diffuse. Cloud providers and hardware firms are expected to vet their customers. Nvidia itself conducts periodic audits. But oversight is uneven, and servers often change hands quietly after passing initial checks. One executive at a server manufacturer says properly verifying all end users is 'practically impossible". America's government has responded by introducing ever tighter regulations. In January the Biden administration unveiled its 'AI Diffusion Framework", which splits the world into three tiers: 18 trusted countries including Britain and Japan face no restrictions; a second tier of 120 countries including Singapore and India are subject to quotas; and a final tier including China and Russia are barred outright. The rules also prohibit American cloud providers from offering services powered by restricted chips to Chinese customers. The Trump administration is reportedly planning to tinker with the regulations before they come into effect—perhaps tying access to chips into wider trade negotiations—but it is yet to announce any changes. The chip has sailed Any new controls will encounter familiar problems. The Bureau of Industry and Security (BIS), the agency tasked with enforcing tech export controls, is severely understaffed. Mr Grunewald notes that it has just one export-control officer responsible for all of South-East Asia and Australasia—a region central to the shadow trade in AI chips. Some pundits have proposed technical solutions, such as disabling chips used in prohibited locations. Nvidia argues that such hardware-level controls would introduce dangerous vulnerabilities and are unworkable. Instead, it suggests that software-based tools could transmit limited telemetry—including information on location and system configuration—back to the company to confirm that the chips are where they are supposed to be. Even better enforcement has its limits, however. Nvidia cannot trace every chip. BIS cannot inspect every server. Smugglers will continue to find loopholes. If America wants to keep ahead of China in the AI race, it will need to innovate faster, rather than clamp down harder.
