Radeon AI PRO R9700

Search documents
中国科技公司争先恐后地抛弃英伟达
Xin Lang Cai Jing· 2025-05-31 05:35
Core Viewpoint - Chinese tech companies are rapidly seeking alternatives to Nvidia due to tightened export controls from Washington, particularly following the H20 chip ban [2][4]. Group 1: Market Dynamics - Major Chinese tech giants like Alibaba, Tencent, and Baidu are testing alternatives to Nvidia's AI chips, such as Huawei's Ascend 920 AI chip, which was launched shortly after the H20 ban [2]. - Industry executives indicate that there is sufficient inventory of Nvidia AI chips to support AI development until early 2026, but many are actively searching for alternatives due to uncertainties surrounding the performance of H20's replacement B20 and AMD's Radeon AI PRO R9700 [2]. Group 2: Technological Advancements - Although Chinese AI chips still lag behind Nvidia's top products, significant progress has been made in R&D, with some companies now able to produce chips that are only one generation behind [4]. - The export controls have prompted innovation among local chip manufacturers, as they recognize the risks of relying solely on US-made chips for AI processing capabilities [4]. Group 3: Challenges and Future Outlook - The primary challenge in deploying alternative AI chips lies in the need to migrate systems from Nvidia hardware, which will incur substantial costs and require extensive support from software and hardware engineers [4]. - Once the migration is completed, subsequent deployments are expected to be easier, and the demand for local AI chip development will continue to grow, potentially threatening Nvidia's global dominance in AI technology [4].
中国芯片,未雨绸缪
半导体芯闻· 2025-05-30 10:08
如果您希望可以时常见面,欢迎标星收藏哦~ 来源:内容来自 观察者网 ,作者:王一,谢谢 。 英国《金融时报》5月30日报道称,随着美国不断收紧对华先进芯片的出口,中国大型科技企业已 开始为没有美国芯片制造商英伟达芯片的未来做准备,积极转向采用国产芯片以发展人工智能 (AI)。 在前任拜登政府一系列芯片出口管制措施的基础上,美国特朗普政府4月再进一步收紧了对华的出 口,要求英伟达在华销售其为中国市场定制的H20芯片时,需先取得美国政府的许可证。 《金融时报》称,新芯片订单通常需要3至6个月才能发货,现有的英伟达处理器库存又只能维持 到明年年初前后,这促使中国大型科技企业加紧将他们的应急计划提上日程。 百度集团执行副总裁、百度智能云事业群总裁沈抖上周表示,该公司可以从一系列芯片选项中选 择,取代英伟达芯片,尤其是推理芯片。 "我们相信,随着时间的推移,自主研发的芯片以及日益高效的自主软件堆栈将共同为中国AI生态 系统的长期创新奠定坚实的基础。"沈抖在百度2025年一季度业绩电话会上说道。 腾讯公司总裁刘炽平在14日举行的2025年第一季度业绩电话会议上透露,腾讯在努力提高芯片使 用效率的同时,也在考虑替代产品。他 ...
AMD Introduces New Radeon Graphics Cards and Ryzen Threadripper Processors at COMPUTEX 2025
Globenewswire· 2025-05-21 03:00
Core Insights - AMD has launched the Radeon RX 9060 XT and Radeon AI PRO R9700 graphics cards, along with the Ryzen Threadripper 9000 Series processors, aimed at enhancing performance in gaming, content creation, and AI development [1][2][4] Product Launches - The Radeon RX 9060 XT features AMD's RDNA 4 architecture, offering smooth 1440p gaming with FSR 4 ML-enhanced upscaling and improved ray tracing capabilities [2][5] - The Radeon AI PRO R9700 is designed for AI workloads, equipped with 32GB of graphics memory and PCIe Gen 5 support, facilitating local AI inference and complex creative tasks [3][9] - The Ryzen Threadripper 9000 Series processors provide high core counts and memory bandwidth, suitable for demanding multi-threaded workloads, with the top model featuring 96 cores and 192 threads [4][14] Performance Specifications - The Radeon RX 9060 XT comes in two variants: 8GB priced at $299 and 16GB priced at $349, with a boost clock of up to 3.13 GHz [7][8] - The Radeon AI PRO R9700 has 64 compute units, 32GB of VRAM, and a boost clock of up to 2.92 GHz, with a total board power of 300W [11][12] - The Ryzen Threadripper PRO 9995WX offers a base frequency of 2.5 GHz and a boost frequency of 5.4 GHz, with a TDP of 350W [16][14] Market Availability - The Radeon RX 9060 XT is expected to be available from various board partners later this year, while the Radeon AI PRO R9700 will be available starting July 2025 [8][12] - The Ryzen Threadripper 9000 Series processors will be available from major manufacturers and retailers starting July 2025 [20] Strategic Vision - AMD emphasizes its commitment to innovation, aiming to empower creators, gamers, and professionals with advanced performance and efficiency [2][21] - The collaboration with ASUS and Lenovo highlights the focus on AI-powered computing solutions for enterprise environments [21][22]