Latest news with #Ascend910C
Phone Arena
3 hours ago
- Business
- Phone Arena
Huawei delays production of its next AI accelerator as Nvidia plans a way to enter China
Because of the sanctions placed against the company by the U.S., Huawei has to be content with having its AI accelerators dominate in China rather than compete in America. The U.S. government might have handed this business to Huawei on a silver platter since it would prefer that global leader Nvidia, whose GPUs are used as AI accelerators, not sell its new H20 chips in China. But Nvidia has come up with a way to take on Huawei in China. This month, Nvidia will start production of the B40 AI accelerator. The B40 is made specifically for the Chinese market. Using Nvidia's latest Blackwell architecture, it is designed to perform at a level that will allow the component to be exported to China under U.S. export rules. With AI sizzling hot, high-profile customers in China have created a new rule for AI chip vendors that says, "Sell only what you already have in stock; if you fail to deliver, you're in trouble." Huawei was planning on having its new AI accelerator, the Ascend 910C, built this month. However, a tweet from leaker @Jukanlosreve says that mass production of the chip has been pushed back to the end of the year due to technical reasons. Huawei has had supply chain problems and the tweet notes that Huawei's domestic rivals have no inventory available to sell or they are still in the middle of developing a competing chip. When production does start at the end of this year, SMIC will build the Ascend 910C for Huawei on its second-generation 7nm process node known as N+2. TSMC, without having to deal with U.S. sanctions like SMIC does, builds Nvidia's H20 GPU chip using its 4nm 4N process node. As a result, the Ascend 910C could be slower and less energy efficient than Nvidia's H20. Production of the Huawei Ascend 910C AI accelerator is delayed until the end of 2025. | Image credit-X AI accelerators and GPUs employed as such use parallel processing which can perform the same calculations simultaneously on many different "pieces" of data. The ability to use parallel processing explains why GPUs, like Nvidia's GPU chips, are used as AI accelerators rather than CPUs. The latter handles its workloads sequentially which makes it less qualified for AI use.
Business Insider
2 days ago
- Business
- Business Insider
Nvidia (NVDA) Rival Huawei Is Struggling to Win Over Customers despite Export Rules
Nvidia (NVDA) CEO Jensen Huang has repeatedly warned about the growing competition from China's Huawei in the AI chip market. But so far, Huawei is struggling to convince China's biggest tech companies to adopt its chips, according to The Information. Indeed, although ByteDance, Alibaba (BABA), and Tencent (TCEHY) are testing Huawei's chips, they haven't placed any large orders yet. This is because the firms still prefer Nvidia's products, despite U.S. export restrictions. While Huawei has sold chips to state-owned companies and local governments, it has yet to make major gains with top private tech firms. Confident Investing Starts Here: One major reason why big Chinese tech firms are hesitant to make the switch is that their data centers and engineering teams are built around Nvidia's CUDA software. Moving to Huawei's ecosystem would require rewriting code, retraining staff, and losing access to tools that have been developed over the past 15 years. Testing has also revealed some problems: Huawei's latest AI chip, the Ascend 910C, often overheats, and its Compute Architecture for Neural Networks software lacks many of CUDA's features. Additionally, Chinese tech firms compete with Huawei in cloud computing, which makes them more cautious about adopting a rival's hardware. It also doesn't help that U.S. warnings about penalties for using Huawei's advanced AI chips have added further risk. In fact, after one such warning, a Chinese data center canceled a planned order of Huawei chips. Nevertheless, Huang continues to monitor Huawei's progress closely. What Is a Good Price for NVDA? Turning to Wall Street, analysts have a Strong Buy consensus rating on NVDA stock based on 35 Buys, four Holds, and one Sell assigned in the past three months, as indicated by the graphic below. Furthermore, the average NVDA price target of $172.36 per share implies 21% upside potential.
Fast Company
15-05-2025
- Business
- Fast Company
Trump's Middle East tour is all about AI diplomacy
Welcome to AI Decoded, Fast Company 's weekly newsletter that breaks down the most important news in the world of AI. You can sign up to receive this newsletter every week here. Trump's Middle East tour is all about AI diplomacy The U.S. enjoys its superpower status mainly because of two things: its military and its financial influence. What we're seeing in Trump's tour of the Middle East this week is the rise of another lever of geopolitical power: AI. And the competition between the U.S. and China in this realm is heating up. The U.S. is becoming more focused on exporting the best U.S. AI technology to other countries. Trump's lavish reception by heads of state in the Middle East this week can be explained in part by a major policy change: The Trump Commerce Department announced plans Tuesday to rescind Biden's 'AI diffusion rule,' which had restricted the export of the most powerful AI chips to other countries, including those in the Middle East. The removal of the chip restrictions opens big new markets for American AI chipmakers (to wit, Nvidia's stock rose 6% Tuesday) and could cause an increase in global investment in new AI data centers in the Middle East. Trump announced a series of U.S.-Saudi investment deals, including a partnership between Nvidia and Humain, a newly formed Saudi AI firm backed by the kingdom's sovereign wealth fund. The plan: to build AI data centers 'powered by several hundred thousand of Nvidia's most advanced GPUs.' The change in posture couldn't be starker. During Biden's presidency, the U.S. took a more cautious approach to AI. Biden-era chip export controls were seen as necessary to protect national security and preserve America's edge in the AI race. Many in the tech industry supported them, at least when it came to chips. Restricting access to the best hardware, as one source put it, was 'one lever that the U.S. can pull' to maintain its lead. The result: U.S. firms like OpenAI and Anthropic had access to elite silicon, while Chinese competitors like DeepSeek were left scrambling. But the game has changed since Biden was in office. The U.S. is no longer home to the only company (Nvidia) that can supply chips powerful enough to train state-of-the-art AI models. The Chinese multinational company Huawei is now shipping the Ascend 910C, a chip that rivals Nvidia's best, along with a high-end server rack, the CloudMatrix 384, that competes with Nvidia's GB200 NVL72. These systems are powering research inside China and are being pushed into global markets. That has raised alarms in D.C. The Commerce Department recently warned that organizations using Huawei's Ascend chips could be violating U.S. export rules, since the chips were likely manufactured with U.S.-origin technology. But enforcement will be difficult as more countries seek alternatives or try to hedge their bets between the U.S. and China. AI models and chips offer a new way for state actors to project power on the world stage. That's what's unfolding in the Middle East this week. The Trump administration isn't so much trying to open new markets for Nvidia as it is trying to advance American AI as the prevailing standard around the world. GOP bill would freeze state AI laws for a decade A sweeping AI regulatory ban that would prevent states from overseeing the technology for a decade has been quietly inserted into a powerful Republican tax and spending bill currently under review by the House Energy and Commerce Committee. If passed in its current form, the bill would mark a major victory for the U.S.'s largest tech companies, which argue that state-level regulations threaten innovation. It would impose a 10-year freeze on 'any law or regulation regulating artificial intelligence models, artificial intelligence systems, or automated decision systems.' For companies like Meta, Microsoft, OpenAI, and Alphabet's Google, the provision offers a way to sidestep pending or active state laws that are imposing stricter oversight than the federal government. Their pitch in recent months has been that any slowdown in AI development could allow Chinese competitors to outpace the U.S., a message that's resonating with many Republicans. Currently, these companies face a wave of state-level scrutiny. In this year alone, states have introduced at least 550 AI-related bills—covering issues from deepfakes to algorithmic discrimination— according to a tracker by the National Conference of State Legislatures. And it's only May. The House committee's draft bill could effectively nullify these efforts, a move that has alarmed AI safety advocates and critics of Big Tech, including leading Democrats. 'This is an outrageous abdication of congressional responsibility and a gift-wrapped favor to Big Tech that leaves consumers vulnerable to exploitation and abuse,' said J.B. Branch, Big Tech accountability advocate at Public Citizen. 'This isn't leadership; it is surrendering to corporate overreach and abuse under the guise of 'protecting American innovation.' ' Sen. Ed Markey of Massachusetts warned that the proposal 'will lead to a dark age for the environment, our children, and marginalized communities.' Illinois Rep. Jan Schakowsky said the ban would allow 'AI companies to ignore consumer privacy protections, let deepfakes spread, and allow companies to profile and deceive consumers using AI.' The bill is advancing through Congress via the budget reconciliation process, which allows certain legislation to bypass the Senate filibuster and pass with a simple majority. However, as Bloomberg reported, the provision may not survive this route, since Senate rules require that such measures be primarily fiscal in nature. Still, the proposal is offering insight into the GOP's broader stance on AI regulation. Vice President JD Vance has already cautioned that overregulation could 'kill' the AI industry—a sentiment that appears to be gaining traction among lawmakers. New Heartland/Rasmussen survey shows 60% of voters say AI companies should pay for lost jobs A new survey from the The Heartland Institute and Rasmussen Reports finds that voters support the idea of AI companies paying reparations for the jobs their technology eliminates. A majority of those surveyed (62%) said that if AI advancements were to cause the elimination of millions of jobs, they would support 'a government program that taxes big technology companies and then uses the funds to provide every American with an income large enough to pay for basic necessities like housing, clothes, and food.'
Yahoo
02-05-2025
- Business
- Yahoo
Nvidia Warns Export Curbs Could Cost Billions
Nvidia (NASDAQ:NVDA) warned U.S. lawmakers that tightening export restrictions could inadvertently supercharge Huawei's AI chip ambitions, potentially redrawing the global semiconductor map. Warning! GuruFocus has detected 3 Warning Signs with NVDA. At a closed-door session Thursday, Nvidia CEO Jensen Huang and other executives told the House Foreign Affairs Committee that fresh Commerce Department restrictions on exporting the company's H20 AI server chips to Chinanow subject to license requirements under a March directivecould cost Nvidia as much as $5.5 billion in potential sales this year, even as Huawei Technologies prepares to begin mass shipments of its own Ascend 910C processor next month and has quietly approached leading Chinese tech firms to field-test its forthcoming Ascend 910D flagship chip. Those stakes are underscored by reported orders of at least $16 billion for Nvidia's H20 chips placed by Alibaba, Tencent and ByteDance in the first quarter of 2025, signaling robust demand for high-performance AI accelerators even as geopolitical tensions flare. If DeepSeek R1 had been trained on Huawei chips or a future open-source Chinese model had been optimized exclusively for Huawei's architecture, that would risk creating a global market demand for Huawei chips, said a senior congressional committee staff member speaking to Reuters, capturing Capitol Hill's deep concern that U.S. export curbs could inadvertently fuel China's homegrown chip ecosystem at a critical juncture in the AI race. Against that backdrop, lawmakers probed Huang on the long-term implications of diminished U.S. market share and pressed for insight into policy tweaks that might preserve America's edge in next-generation semiconductors without ceding ground to Beijing. Nvidia spokesperson John Rizzo told Reuters that Jensen met with the House Foreign Affairs Committee to discuss the strategic importance of AI as national infrastructure and the need to invest in U.S. manufacturing, reiterating the company's full-throated support for government efforts to bolster American semiconductor leadership through both policy and capital investment. In a March interview with the Financial Times, Huang had described Huawei as the single most formidable technology company in China, lamenting that U.S.-led restrictions have been done poorly even as the Chinese rival has clawed back smartphone market share and advanced its AI chip roadmap, and he committed to spending hundreds of billions of dollars on U.S.-made chips and electronics over the next four years as part of a broader push for onshore capacity under the CHIPS Act. Investors should note that U.S. license requirements designed to hobble Huawei could backfire by accelerating demand for Chinese chips, potentially reshaping revenue projections for Nvidia and its peers. The outcome may hinge on how swiftly Washington can fine-tune export rules without undermining domestic champions in the global AI hardware war. Investors will now watch Nvidia's early-June investor day and the Commerce Department's export license rulings for the next AI hardware catalyst. This article first appeared on GuruFocus. Sign in to access your portfolio
Yahoo
01-05-2025
- Business
- Yahoo
Huawei Ascend AI 910D processor designed to take on Nvidia's Blackwell and Rubin GPUs
When you buy through links on our articles, Future and its syndication partners may earn a commission. Huawei's next-generation HiSilicon Ascend 910D AI processor is expected to offer better performance than Nvidia's H100, reports Reuters. The new processor will be slower on a chip vs chip basis compared to Nvidia's Blackwell B200 and Blackwell Ultra B300 GPUs, never mind the next-generation Rubin GPUs slated to launch next year. However, Huawei's approach of building pods with hundreds of processors should allow the Ascend 910D to compete against pods based on Nvidia's current Blackwell and upcoming Rubin GPUs. Huawei is preparing to start tests of its most advanced artificial intelligence processor, the Ascend 910D, with the performance goal of surpassing Nvidia's H100 and offering a domestic alternative amid U.S. export restrictions. According to sources, Huawei has approached several local companies to assess whether the new Ascend 910D chip meets performance and deployment requirements. Initial samples are expected by late May. Separately, Huawei plans to start large-scale shipments of its dual-chiplet Ascend 910C AI processors to Chinese customers (and probably full systems based on the chips) as early as next month. The majority of of these processors were reportedly produced by TSMC for a third-party company. It remains to be seen whether the Ascend 910D will be made by China-based SMIC, or whether — nearly five years after the U.S. government restricted Huawei's access to leading-edge semiconductor production capabilities — Huawei will once again find a way to circumvent U.S. sanctions. Reaching Nvidia H100 performance levels won't be easy for Huawei. The company's latest dual-chiplet Ascend 910C offers around 780 BF16 TFLOPS of performance, whereas Nvidia's H100 can deliver around 2,000 BF16 TFLOPS. In order to achieve H100 performance levels, Huawei will have to redesign the internal architecture of the Ascend 910D and possibly increase the number of compute chiplets. To stay competitive in the AI industry next year, Huawei will have to achieve performance comparable to that of AI clusters developed in the U.S. This year, the company introduced its CloudMatrix 384 system with 384 Ascend 910C processors. It can reportedly beat Nvidia's GB200 NVL72 in certain workloads, but at the cost of significantly higher power consumption due to dramatically lower performance-per-watt. It also has over five times as many 'AI processors' as an NVL72 rack. Whether the interconnect can scale well to the required number of processors remains to be seen. Without access to leading-edge process technologies, it will become significantly more difficult for Huawei to maintain competitive positions next year. Nvidia is on-track to introduce its codenamed Rubin GPUs for AI and HPC in 2026. Rubin GPUs are set to be made on TSMC's N3 (or a more advanced) fabrication process, and they should offer even higher performance-per-watt than the current-generation Blackwell GPUs. Rubin GPUs are slated to offer around 8,300 TFLOPS of FP8 training performance, and presumably half that for BF16 — about twice the performance of the B200. Huawei's Ascend 910D and next-generation CloudMatrix systems with 384 of such processors could theoretically offer competitive AI performance on the rack level. However, it remains to be seen what performance benefits Huawei's Ascend 910D and Nvidia's Rubin GPUs will offer compared to existing offerings. Also, it should be noted that Nvidia will barely be able to sell its high-performance Rubin GPUs in China, so for that market Huawei won't really have a direct competitor. Regardless of performance or efficiency, Huawei's Ascend 910D processors will likely become China's workhorses when it comes to AI training in the coming years. Given the strategic importance of AI, the power consumption of the Ascend 910D (or any other domestic AI processor) will not be a limiting factor, as the number of deployed units could offset the efficiency of Nvidia's (or AMD, Intel, Broadcom, etc.) AI processors. The main limiting factor for China will be its ability to produce enough processors — either domestically, or overseas using proxy companies.
