Latest news with #BlackwellB200


Business Insider
6 days ago
- Business
- Business Insider
AMD Stock Rally Is ‘Empty Calories,' Says Citi's Top Analyst
Citi's top analyst, Christopher Danely, calls the recent rally in AMD stock (AMD) 'empty calories' because he believes the chip ban can be reinstated at any time. Yesterday, Advanced Micro Devices rallied 6.4% after President Donald Trump lifted the ban on export of artificial intelligence (AI) chips to China. The news also boosted shares of major semiconductor companies, including Nvidia (NVDA), which rose 4% to reach a new all-time high of $172.40. Elevate Your Investing Strategy: Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence. Make smarter investment decisions with TipRanks' Smart Investor Picks, delivered to your inbox every week. Danely has a 'Hold' rating on AMD stock and a price target of $145, which implies 6.8% downside potential from current levels. Danely ranks #588 out of 9,861 analysts tracked on TipRanks. He boasts a 66% success rate and an average return per rating of 11.40%. Here's Why Citi Is Skeptical About the AI Ban Reversal The easing of export restrictions means AMD can now resume shipments of its MI308 AI chips to China. The White House had previously imposed a ban on the export of advanced AI chips to China due to national security concerns. Notably, AMD was expected to take an $800 million hit from the ban, while Nvidia faced a larger $5.5 billion impact. However, the situation has changed following the easing of trade tensions between the U.S. and China. Danely has a rather skeptical view on the news. He noted that the resumption of MI308 chip sales to China is 'not sustainable' as they could be banned again. The U.S.-China trade relations remain fragile, and even a minor incident could prompt substantial policy shifts. Having said that, Danely believes that the Trump administration's $70 billion investment in Pennsylvania for AI and energy could be highly beneficial for chipmakers, including AMD, NVDA, Broadcom (AVGO), and Micron (MU). AMD Is Positioning as a Serious Rival to Nvidia For years, AMD has been seen as secondary to Nvidia, with the latter's advanced AI chips and GPUs (Graphics Processing Units) commanding global attention. However, AMD is rapidly advancing with its own chip innovations to compete more effectively with Nvidia. Recently, AMD launched its new line of MI350 AI chips, which are expected to be nearly 30% cheaper than Nvidia's comparable chips. Also, AMD's MI355X features high-bandwidth memory (HBM), which is reportedly three times that offered by Nvidia's Blackwell B200. Is AMD a Good Stock to Buy? Despite the current optimism surrounding Advanced Micro Devices, analysts remain divided about its long-term outlook. On TipRanks, AMD stock has a Moderate Buy consensus rating based on 25 Buys and 10 Hold ratings. Also, the average AMD price target of $135.97 implies 12.6% downside potential from current levels. Year-to-date, AMD stock has gained nearly 29%.

Bangkok Post
07-07-2025
- Business
- Bangkok Post
Jitters over ban on AI chip exports
Thailand's strategic investment in cloud and data centres could be disrupted by US President Donald Trump's plan to restrict shipments of artificial intelligence (AI) chips from Nvidia Corp to Malaysia and Thailand, part of an effort to crack down on suspected semiconductor smuggling to China, say tech and cloud data centre executives. According to Bloomberg, a draft rule from the US Commerce Department seeks to prevent China, which faces a US ban on sales of Nvidia's advanced AI processors, from obtaining such components through intermediaries in the two Southeast Asian nations, said people familiar with the matter. The rule has yet to be finalised and could still change, said the sources, who requested anonymity. Pathom Indarodom, managing director of SVOA, said this move may result in the US slapping export restrictions on Thailand for high-tech equipment, especially Nvidia's flagship graphics processing unit (GPU) chips such as A100, H100 and the latest Blackwell B200. These chips power AI models globally, from OpenAI and Google to emerging deep tech startups, government agencies and universities, said Mr Pathom. He said if access to these chips is blocked, Thailand could experience major disruptions, mainly in AI development research and local AI startups, which rely on Nvidia GPUs to train and test models. This would also affect Thailand's competitiveness at the global level, said Mr Pathom. Moreover, major Thai cloud providers that utilise Nvidia chips would be impacted as their AI-as-a-service offerings would be unable to scale, he said. Enterprise clients may shift workloads overseas, while national data sovereignty faces rising risks as data processing will move to offshore, said Mr Pathom. The overall effect would chill Thailand's goal of becoming a regional AI hub, he said. If Western technology is inaccessible, Thailand might have to shift to alternative technology from China. Geopolitical tensions demand Thailand devise backup plans by investing in open-source AI models, said Mr Pathom. Thailand should forge strategic tech alliances with other Asian nations to create a resilient AI ecosystem, he said. According to the Board of Investment, from 2022 to 2024 investment promotion applications for data centre and cloud services totalled 27 projects with a combined investment of 290 billion baht. There were eight new data centre projects in the first five months of this year valued at 183 billion baht. According to a report from Cushman & Wakefield, a leading commercial real estate services firm, Malaysia's strong data centre development will see its population per megawatt ratio consolidate faster than any other Asia-Pacific market through 2030, with Thailand ranked second and Japan third for fastest rates of consolidation. Supparat Singhara na Ayutthaya, general manager for Southeast Asia at Damac Digital, a data centre builder, expressed concern over the planned restriction on shipments of AI chips to Malaysia and Thailand, noting it could undermine investor confidence in Thailand's cloud and data centre industry. He said the Commerce Ministry and Digital Economy and Society Ministry need to collaborate to impose stricter export controls in Thailand, while actively engaging with the US government. If left unaddressed, Thailand may find itself excluded from access to the latest technological advancements, said Mr Supparat. NO DIRECT SANCTIONS Somchai Sittichaisrichart, managing director of SIS Distribution, said the planned US regulation does not directly sanction Thailand, instead introducing stricter oversight to prevent Chinese firms from accessing AI computing services or importing chips via Thai nominees. "Thai companies can still import AI chips," he said. "However, once the regulation takes effect, scrutiny will likely extend beyond the importer to the intended use, particularly to ensure chips are not deployed for Chinese customers." A source in the cloud service sector who requested anonymity said his company may consider expanding operations from Thailand to other countries to mitigate potential risks from the planned US rules. Putt Tratkosit, chief executive of Light Up Total Solution Plc (LTS), said the company does not face any risk or direct impact from the US plan to restrict shipments of AI chips to Malaysia and Thailand. He said the company is not involved in importing chips for resale to China. LTS provides design and installation services for AI infrastructure, as well as commissioning services for government agencies and companies in Thailand. The company works closely with Siam AI Corporation, a high-performance cloud and AI infrastructure provider, which is the only certified Nvidia Cloud Partner in Thailand authorised to purchase Nvidia chips directly from the US. Should US export control measures take effect in the future, importing chips may require additional approval. However, the regulation does not prohibit domestic use, meaning AI and data centre services can continue operating normally, said Mr Putt. Demand for AI solutions in Thailand remains strong, and LTS sees a short-term upside as customers may accelerate imports of AI chips, leading to faster deployment of infrastructure projects, he said.
Yahoo
08-04-2025
- Business
- Yahoo
NVIDIA Speeds Up Meta's Llama 4 AI Models With Blackwell GPUs, Hitting 40,000 Tokens Per Second
NVIDIA (NVDA, Financials) said it boosted the speed and efficiency of Meta Platforms' (META, Financials) latest Llama 4 artificial intelligence models by running them on its new Blackwell B200 graphics processors, reaching output speeds of over 40,000 tokens per second. Warning! GuruFocus has detected 3 Warning Signs with NVDA. The two models, Llama 4 Scout and Llama 4 Maverick, are available as NVIDIA NIM microservices and are built with a mixture-of-experts design to handle multilingual and multimodal tasks. Llama 4 Scout has 109 billion parameters, with 17 billion active per token, and can handle up to 10 million tokens in a single context. It's built to run efficiently on a single NVIDIA H100 GPU and is suited for summarizing documents, analyzing user activity, and reading large codebases. Llama 4 Maverick is a larger model with 400 billion parameters and 128 experts, accepting up to 1 million context tokens. It's designed for tasks that require understanding both images and text. Both models are tuned for NVIDIA's open-source TensorRT-LLM library, which speeds up inference for large models. NVIDIA said its TensorRT Model Optimizer lets developers restructure and compress models using bfloat16 and FP4 precision without hurting accuracy. On the Blackwell B200 GPU, Llama 4 Scout delivers over 40,000 tokens per second, while Maverick tops 30,000. The company said this represents a 3.4x boost in throughput and 2.6x better cost efficiency compared to its previous H200 GPU, helped by Blackwell's upgraded architecture and support for precision formats like FP8 and FP4. NVIDIA said its work with Meta continues a broader effort to support open-source AI tools. The Llama 4 models can be customized with the NVIDIA NeMo framework, which helps companies prepare datasets, fine-tune models using techniques like LoRA and PEFT, and test performance using custom or standard benchmarks. To help businesses get started, the models will be delivered as NIM microservices that can run on any GPU-accelerated system. These services support standard APIs and are built to scale across cloud, data center, and edge setups with privacy and security features. This article first appeared on GuruFocus. Sign in to access your portfolio
Yahoo
13-02-2025
- Business
- Yahoo
Super Micro's $40 Billion AI Bet Sends Shares Soaring--But One Big Risk Could Change Everything
Super Micro Computer (NASDAQ:SMCI) just dropped a massive forecast bombshell$40 billion in sales by fiscal 2026, crushing Wall Street's $30.7 billion estimate. That's a bold call, fueled by skyrocketing demand for AI-optimized servers, especially with Nvidia's new Blackwell B200 chips in the mix. But here's the twist: the near-term outlook isn't as pretty. The company slashed its fiscal 2025 revenue target to $24.3 billion from $28 billion, and its latest quarterly sales and profit figures missed estimates. Investors, though, are locked in on the long gameSuper Mirco's shares jump 7.4% at 1.04pm today. Warning! GuruFocus has detected 5 Warning Signs with SMCI. Meanwhile, Super Micro is shaking up its balance sheet with a $700 million convertible notes offering, alongside tweaks to its existing $1.725 billion convertible notes due 2029. The goal? Fuel expansion and double down on AI infrastructure. But there's an elephant in the roomregulatory troubles. The company is still racing to meet a February 25 Nasdaq compliance deadline after missing its annual financial filing, and it's under scrutiny from the U.S. Department of Justice and SEC. Despite an independent review finding no evidence of misconduct, questions around governance and transparency linger. Still, the AI gold rush isn't slowing down, and Super Micro is positioning itself as a key player in the high-performance computing arms race. Analysts say the AI wave could keep lifting Super Micro, along with peers like Dell and Hewlett Packard Enterprise. But for now, investors are watching two things: whether SMCI can clear its regulatory hurdles and if its explosive growth projections hold up in reality. This article first appeared on GuruFocus. Sign in to access your portfolio