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.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
3 hours ago
- Yahoo
Better Artificial Intelligence (AI) Stock: CoreWeave vs. Nvidia
CoreWeave has made a splash in the market as it quickly grows its cloud services business. Nvidia is proving its AI lineup of products is becoming more and more pervasive. CoreWeave is valued at a high multiple and has massive capital spending planned. 10 stocks we like better than Nvidia › There will prove to be many winners as artificial intelligence (AI) infrastructure continues to grow and AI end-uses expand. Nvidia (NASDAQ: NVDA) has been the Wall Street darling surrounding everything AI for the past two years. CoreWeave (NASDAQ: CRWV) has been getting the love most recently, though. Shares of the AI hyperscaler providing cloud services have soared about 185% in just the past month as of this writing. Nvidia stock has increased 24% in that time. CoreWeave just went public in late March, and the shares have jumped about 270% since that initial public offering (IPO). Investors may wonder if Nvidia's shine is fading, and it's time to buy CoreWeave instead. I'd argue that is flawed thinking, however. Investors may be taking a breather after the early exponential gains in Nvidia stock. Growth in the business itself has also slowed, though that was inevitable. Sales of its advanced chips in the data center segment had been growing like a weed. Revenue in that segment has been increasing in each consecutive quarter for the last two years. In the most recent fiscal quarter, that growth rate slowed to 10%, though, as seen below. Despite that trend, it's clear AI demand hasn't yet peaked. Remember, these are still sequential quarterly increases in data center sales. For perspective, that fiscal first-quarter revenue was a 73% jump compared to the prior year period. Management also guided investors to expect further revenue growth in the current quarter. So, while an unsustainable growth rate slows, the company is still solidly in growth mode. That's because it's not just Nvidia's advanced GPU and CPU chips driving sales and expanding AI infrastructure. Its AI ecosystem includes interconnect technologies, the CUDA (compute unified device architecture) software platform, and artificial intelligence processors that are part of many different types of architectures. CEO Jensen Huang recently touted Nintendo's new Switch 2 gaming console, for example. The unit includes Nvidia's AI processors that Huang claims "sharpen, animate, and enhance gameplay in real time." Nvidia has a broad array of customers. As AI factories and data centers are built, it will continue to be a major supplier and one that investors should benefit from owning. Nvidia also invests in the AI sector. It makes sense to look at where the AI leader itself sees future gains. One of the AI companies in which Nvidia holds a stake is CoreWeave. Nvidia should know CoreWeave well, too, as an important customer. CoreWeave leases data center space to companies needing the scalable, on-demand compute power it has control of from the 250,000 Nvidia chips it has purchased. It's a desirable option for enterprises that require significant computational power to process large amounts of data efficiently. There appears to be plenty of demand. But there is plenty of risk for investors, too. It just announced a new lease agreement to further increase capacity. Applied Digital, a builder and operator of purpose-built data centers, has agreed to deliver CoreWeave 250 megawatts (MW) of power load on a 15-year term lease at its recently built North Dakota data center campus. CoreWeave has the option to expand the load by an additional 150 MW in the future. Demand is quickly driving growth for CoreWeave. That's led investors to jump in and drive the stock higher in recent months. Valuation is just one major risk with CoreWeave. Customer concentration is another. Last year, Microsoft accounted for nearly two-thirds of revenue. CoreWeave also disclosed that 77% of 2024 revenue came from just its top two customers. CoreWeave is also spending massive amounts of capital to grow AI cloud capacity. It had about $5.4 billion of liquidity available as of March 31 and raised another $2 billion from a late May debt offering. That's approximately its level of capital expenditure in just the first quarter alone, though. That spending may pay off. But there are risks there as well. Customers could develop their own AI infrastructure or could redesign systems that don't require its services. CoreWeave stock also trades at a high valuation after the stock has soared. It recently had a price-to-sales (P/S) ratio of about 30. That could be cut in half this year with its strong sales growth, but it isn't earning any money yet. At the same time, Nvidia sports a price-to-earnings (P/E) ratio of about 30 based on this year's expected profits. Remember, too, that as CoreWeave grows, so do Nvidia's profits. Applied Digital CEO Wes Cummins said that its leased North Dakota data center campus will be full of Nvidia Blackwell class servers. I think the risk profile, financial picture, and massive potential for Nvidia make it the better AI stock to buy now. Before you buy stock in Nvidia, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Nvidia wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $669,517!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $868,615!* Now, it's worth noting Stock Advisor's total average return is 792% — a market-crushing outperformance compared to 173% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 2, 2025 Howard Smith has positions in Microsoft and Nvidia. The Motley Fool has positions in and recommends Microsoft and Nvidia. The Motley Fool recommends Nintendo and recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy. Better Artificial Intelligence (AI) Stock: CoreWeave vs. Nvidia was originally published by The Motley Fool Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
3 hours ago
- Yahoo
Veteran fund manager reboots Palantir stock price target
Veteran fund manager reboots Palantir stock price target originally appeared on TheStreet. There's been a lot of debate surrounding artificial intelligence stocks this year. A boom in AI spending, particularly by hyperscalers ramping infrastructure to meet surging research and development of chatbots and agentic AI, led to eye-popping returns for companies like Palantir Technologies, which markets data analytics platforms. However, concern that spending could decelerate has picked up in 2025 because of worry over a tariffs-driven recession, causing many AI stocks like chip-maker Nvidia to the eventual impact of tariffs on recession remains a question mark, there's been little to suggest demand for Palantir's services is slipping. Solid first-quarter earnings results and optimism that trade deals could make tariffs manageable have helped Palantir shares rally 63% this year after a 340% surge in 2024. Palantir's resiliency isn't lost on long-time money manager Chris Versace. Versace, who first picked up shares last year, recently updated his price target as Palantir's stock challenges all-time highs. Investors' interest in Palantir stock swelled after OpenAI's ChatGPT became the fastest app to reach one million users when it was launched in December 2022. ChatGPT's success has spawned the development of rival large language models, including Google's Gemini, and a wave of interest in agentic AI programs that can augment, and in some cases, replace traditional activity is widespread across most industries. Banks are using AI to hedge risks, evaluate loans, and price products. Drugmakers are researching AI's ability to predict drug targets and improve clinical trial outcomes. Manufacturers are using it to boost production and quality. Retailers are using it to forecast demand, manage inventories, and curb theft. The U.S. military is even seeing if AI can be effective on the battlefield. The seemingly boundless use cases — and the ability to profit from them — have many companies and governments turning to Palantir's deep expertise in managing and protecting data to train and run new AI apps. Palantir got its start helping the U.S. government build counterterrorism systems. Its Gotham platform still assists governments in those efforts today. It also markets its Foundry platform to manage, interpret, and report data to large companies across enterprise and cloud networks. And its AI platform (AIP) is sold as a tool for developing AI chatbots and apps. Demand for that platform has been big. In the fourth quarter, Palantir closed a "record-setting number of deals," according to CEO Alex Karp. The momentum continued into the first quarter. Revenue rose 39% year-over year to $884 million. Meanwhile, Palantir's profit has continued to improve as sales have grown. In Q1, its net income was $214 million, translating into adjusted earnings per share of 13 cents. "Our revenue soared 55% year-over-year, while our U.S. commercial revenue expanded 71% year-over-year in the first quarter to surpass a one-billion-dollar annual run rate,' said Karp in Palantir's first-quarter earnings release. 'We are delivering the operating system for the modern enterprise in the era of AI." AI's rapid rise has opened Palantir's products to an increasingly new range of industries, allowing it to diversify its customer base. For example, Bolt Financial, an online checkout platform, recently partnered with Palantir to use AI tools to analyze customer behavior better. More Palantir: Palantir gets great news from the Pentagon Wall Street veteran doubles down on Palantir Palantir bull sends message after CEO joins Trump for Saudi visit The potential to ink more deals like this has caught portfolio manager Chris Versace's attention. "The result [of the Bolt deal] will be technology that can offer shoppers a customized checkout experience, embedded within retailers' sites and apps, and it is one that will extend to agentic checkout as well," wrote Versace on TheStreet Pro. "We see this as the latest expansion by Palantir into the commercial space, and we are likely to see more of this as AI flows through payment processing and digital shopping applications." Alongside Palantir's deeply embedded government contracts, growing relationships with enterprises should provide Palantir with cross-selling opportunities, further driving sales and profit growth, allowing for increased financial guidance. Palantir is guiding for full-year sales growth of 36%, and U.S. commercial revenue growth of 68%. The chances for Palantir growth to continue accelerating has Versace increasingly optimistic about its shares. As a result, he's increased his price target to $140 per share from $ fund manager reboots Palantir stock price target first appeared on TheStreet on Jun 8, 2025 This story was originally reported by TheStreet on Jun 8, 2025, where it first appeared. Sign in to access your portfolio
New York Post
4 hours ago
- New York Post
US-China trade talks to open in London as new disputes emerge
US-China trade talks in London this week are expected to take up a series of fresh disputes that have buffeted relations, threatening a fragile truce over tariffs. Both sides agreed in Geneva last month to a 90-day suspension of most of the 100%-plus tariffs they had imposed on each other in an escalating trade war that had sparked fears of recession. Since then, the US and China have exchanged angry words over advanced semiconductors that power artificial intelligence, 'rare earths' that are vital to carmakers and other industries, and visas for Chinese students at American universities. Advertisement 3 President Trump spoke at length with Chinese leader Xi Jinping by phone last Thursday in an attempt to put relations back on track. REUTERS President Trump spoke at length with Chinese leader Xi Jinping by phone last Thursday in an attempt to put relations back on track. Trump announced on social media the next day that trade talks would be held on Monday in London. The latest frictions began just a day after the May 12 announcement of the Geneva agreement to 'pause' tariffs for 90 days. Advertisement The US Commerce Department issued guidance saying the use of Ascend AI chips from Huawei, a leading Chinese tech company, could violate US export controls. That's because the chips were likely developed with American technology despite restrictions on its export to China, the guidance said. The Chinese government wasn't pleased. One of its biggest beefs in recent years has been over US moves to limit the access of Chinese companies to technology, and in particular to equipment and processes needed to produce the most advanced semiconductors. 'The Chinese side urges the US side to immediately correct its erroneous practices,' a Commerce Ministry spokesperson said. US Commerce Secretary Howard Lutnick wasn't in Geneva but will join the talks in London. Analysts say that suggests at least a willingness on the US side to hear out China's concerns on export controls. Advertisement 3 US Commerce Secretary Howard Lutnick will take part in the talks in London. One area where China holds the upper hand is in the mining and processing of rare earths. They are crucial for not only autos but also a range of other products from robots to military equipment. The Chinese government started requiring producers to obtain a license to export seven rare earth elements in April. Resulting shortages sent automakers worldwide into a tizzy. As stockpiles ran down, some worried they would have to halt production. Trump, without mentioning rare earths specifically, took to social media to attack China. Advertisement 'The bad news is that China, perhaps not surprisingly to some, HAS TOTALLY VIOLATED ITS AGREEMENT WITH US,' Trump posted on May 30. 3 China dominates the mining and processing of rare earth minerals. REUTERS The Chinese government indicated Saturday that it is addressing the concerns, which have come from European companies as well. A Commerce Ministry statement said it had granted some approvals and 'will continue to strengthen the approval of applications that comply with regulations.' The scramble to resolve the rare earth issue shows that China has a strong card to play if it wants to strike back against tariffs or other measures. Student visas don't normally figure in trade talks, but a US announcement that it would begin revoking the visas of some Chinese students has emerged as another thorn in the relationship. China's Commerce Ministry raised the issue when asked last week about the accusation that it had violated the consensus reached in Geneva. It replied that the US had undermined the agreement by issuing export control guidelines for AI chips, stopping the sale of chip design software to China and saying it would revoke Chinese student visas.
