爱芯元智通过港交所聆讯美团腾讯都投了 商报讯(记者叶晓珺)又一家芯片公司即将登陆资本市场，且是国产边缘AI芯片领域的王者。1月25日， 据港交所官网，来自宁波的中国第二大智能汽车SoC供应商爱芯元智半导体股份有限公司(以下简称"爱 芯元智")已通过聆讯，中金、国泰君安国际、交银国际为联席保荐人，若后续进展顺利，有望在农历春 节后登录港股，成为"中国边缘AI芯片第一股"。 灼识咨询的数据显示，爱芯元智是全球第一大中高端视觉端侧AI推理芯片提供商以及中国第三大边缘 AI推理芯片供应商。截至2025年底，其智能汽车芯片累计出货量已近100万片，同比增长近90%，成为 公司业务增长的重要引擎。 业绩方面，爱芯元智近年来营收持续增长，不过整体仍处于亏损阶段。 据招股书，2022年至2024年，其收入分别为0.5亿元、2.3亿元和4.73亿元，年复合增长率206.8%；2025 年1月—9月，实现收入2.69亿元，去年同期为2.54亿元。净利润方面，同期分别为-6.12亿元、-7.43亿 元、-9.04亿元和-8.56亿元，亏损连年扩大。在非国际财务报告准则下，经调整后的净利润分别为-4.44 亿元、-5.42亿元、-6 ...

| [编纂]的[编纂]數目 | [编纂]股H股(視乎[编纂]行使與否而定) | | --- | --- | | 国美版目 | [编纂]股H股(可予[编纂]) | | [瑞豪]数目 .. 定) | [编纂]股H股(可予[编纂]及視乎[编纂]行使與否而 | | 最高[编纂] : | 每股H股[編纂]港元,另加1.0%經紀佣金、 | | | 0.0027%證監會交易微費、0.00015%會財局交 | | | 易微費及0.00565%聯交所交易費(須於申請時 | | | 以港元繳足,多繳股款可予退還) | | 面值 | 每股H股人民幣1.00元 | | 【甜餐】 【湖景】 | | | [聯席保薦人 ·[編纂]] | | 爱芯元智成立于2019年5月，由前紫光展锐CTO仇肖莘创办，主要供应AI推理系统芯片（SoC），专注 于为边缘计算与终端设备AI应用打造高性能感知与计算平台。 该公司在2025年获评国家级专精特新"小巨人"企业，其CEO孙微风曾任职于海思半导体。 芯东西（公众号：aichip001） 作者 | ZeR0 编辑 | 漠影 智东西1月26日报道，1月25日，港交所公开披露浙江宁波AI芯片独角兽企业爱芯元智 ...

1月25日，浙江宁波AI芯片独角兽企业爱芯元智正式递表港交所。 | [编纂]的[编纂]數目 | : | [编纂]股H股(視乎[编纂]行使與否而定) | | --- | --- | --- | | [编纂]數目 | .. | [编纂]股H股(可予[编纂]) | | [编纂]數目 | .. | [编纂]股H股(可予[编纂]及視乎[編纂]行使與否而 | | | | 定) | | 最高[编纂] | | : 每股H股[編纂]港元,另加1.0%經紀佣金、 | | | | 0.0027％證監會交易徵費、0.00015%會財局交 | | | | 易徴費及0.00565％聯交所交易費(須於申請時 | | | | 以港元繳足,多繳股款可予退還) | | 面值 | : | 每股H股人民幣1.00元 | | [编纂] | | : [编纂] | | [聯席保薦人·[编纂]] | | | | 全八言 | 1 國泰培順 | 交銀國際 國泰君安國際 | 爱芯元智成立于2019年5月，由清华大学校友、前紫光展锐CTO仇肖莘创办，主要供应AI推理系统芯片（SoC），专注于为边缘计算与终端设 备AI应用打造高性能感知与计算平台。 该公司在2025 ...

21世纪经济报道记者 赵娜 报道 在技术路线选择上，曦望并未参与训推一体 GPU 的参数竞赛，而是明确聚焦于推理场景，对 GPU 架构进行原生重构，在调 度、缓存、片上 SRAM、互联及编译器路径上进行针对性优化，以降低单位 Token 推理成本。 GPU 芯片领域再现大额融资。 据介绍，曦望选择深耕推理芯片赛道，核心是基于训练与推理场景的本质差异精准布局：训练芯片侧重极致算力与高比特精 度，以支撑模型参数迭代优化；而推理作为 AI 落地的"最后一公里"，更需适配多场景的低能耗、高并发与低延迟需求。不仅如 此，曦望的推理专用GPU兼具通用性和极致性价比，能够为大模型提供更低成本、更低能耗、更高稳定性的推理服务。 1月22日，GPU 芯片公司曦望（Sunrise）宣布一年内已完成近 30 亿元融资，投资方包括三一集团旗下华胥基金、范式智能、杭 州数据集团、正大机器人、协鑫科技等产业投资方，IDG资本、高榕创投、无极资本、中金资本、普华资本、松禾资本、易方 达资本、工银投资、心资本等知名VC/PE机构，以及诚通混改基金等国资背景资本。 在产品进展上，公司截至当前已形成完整的推理 GPU 产品矩阵。其中，S1 云边 ...

Core Viewpoint - Nvidia and Groq have announced a $20 billion non-exclusive licensing agreement, which analysts suggest is a strategic move to navigate antitrust scrutiny while maintaining a competitive facade in the AI market [1] Group 1: Strategic Implications - The non-exclusive licensing agreement is seen as a common strategy among tech giants to avoid regulatory challenges, particularly antitrust risks [1] - Analysts from Bernstein highlight that structuring the deal as a non-exclusive license helps to maintain the appearance of competition in the market [1] - Cantor analysts describe the agreement as both an offensive and defensive strategy, aiding Nvidia in expanding its complete system technology stack and solidifying its leadership in the AI market [1] Group 2: Market Significance - Bank of America characterizes the deal as "surprising, expensive but strategically significant," emphasizing Nvidia's focus on the future growth of AI inference chips [1]

Core Viewpoint - Nvidia has made its largest acquisition to date, entering a $20 billion technology collaboration agreement with AI chip startup Groq, which includes the integration of Groq's core team into Nvidia [2][11]. Group 1: Acquisition Details - The agreement is a non-exclusive licensing deal allowing Nvidia to utilize Groq's inference technology, while Groq will continue to operate as an independent entity under CEO Simon Edwards [4][11]. - This acquisition model, termed "Acqui-hire 2.0," focuses on technology asset transfer and talent acquisition without equity purchase, thereby avoiding antitrust scrutiny [4][11]. - Nvidia's CEO Jensen Huang stated that this agreement will enhance Nvidia's capabilities, particularly in AI inference and real-time workloads [11]. Group 2: Groq's Technology and Market Position - Groq's LPU chip, designed specifically for large language models, boasts inference performance improvements of 10 to 100 times compared to conventional GPUs and TPUs, positioning it as a significant competitor to Nvidia's GPU dominance [5][12]. - The global AI inference chip market is projected to reach three times the size of the training market by 2025, indicating a shift in focus within the AI industry [12]. Group 3: Competitive Landscape - Nvidia currently holds a 90% market share in AI training chips but faces increasing competition in the inference space from companies like Google, Meta, and various startups [12]. - Other tech giants are also pursuing similar acquisition strategies, with companies like Meta and AMD acquiring AI chip startups to bolster their capabilities [13]. Group 4: Financial and Investment Context - Groq has attracted significant investment since its inception, including a recent $750 million round led by firms such as Sequoia Capital and BlackRock, valuing the company at $6.9 billion [6][7]. - Nvidia's cash reserves of $60.6 billion provide a strong financial foundation for such strategic acquisitions, allowing it to eliminate potential competitors while enhancing its technological edge [11].

Core Viewpoint - The article discusses a new paradigm for technology giants to acquire technology and talent through "licensing + talent acquisition" instead of direct acquisitions, in response to increasing regulatory scrutiny [2][15]. Group 1: Nvidia and Groq Partnership - Nvidia has entered into a significant partnership with AI chip startup Groq, involving a $20 billion non-exclusive licensing agreement to utilize Groq's inference technology [2][15]. - Groq will continue to operate as an independent entity, with its CEO Simon Edwards leading the company, while key personnel including founder Jonathan Ross will join Nvidia [3][5]. - This partnership is part of a trend in Silicon Valley known as "Acqui-hire 2.0," which focuses on technology asset transfer and talent onboarding without equity acquisition, thereby avoiding antitrust scrutiny [3][15]. Group 2: Groq's Technology and Market Position - Groq's LPU chip, designed for large language models, boasts inference performance that is 10 to 100 times better than conventional GPUs and TPUs, posing a significant threat to Nvidia's GPU dominance [5][13]. - The integration of Groq's technology is expected to enhance Nvidia's capabilities in AI inference, particularly in applications like autonomous driving and high-frequency trading [13][16]. Group 3: Market Trends and Competitive Landscape - The global AI inference chip market is projected to reach three times the size of the training market by 2025, indicating a shift in focus within the AI industry from training to inference [16]. - Nvidia currently holds a 90% market share in AI training chips but faces increasing competition in the inference space from companies like Google, Meta, and various startups [16]. - Other tech giants, including Meta and AMD, are also pursuing similar strategies to acquire AI chip startups, reflecting a broader trend in the industry [15][17]. Group 4: Chinese Counterparts - Chinese companies are mirroring the trends seen in the U.S. AI chip market, with firms like Zhonghao Xinying and others developing technologies comparable to those of their American counterparts [19][20]. - Zhonghao Xinying recently completed a significant acquisition of Tianpu Co., with a total investment exceeding 2.1 billion yuan, highlighting the growing interest in AI chip technology within China [23][24].

以下文章来源于芯东西 ，作者ZeR0 芯东西 . 芯东西专注报道芯片、半导体产业创新，尤其是以芯片设计创新引领的计算新革命和国产替代浪潮；我们是一群追"芯"人，带你一起遨游"芯"辰大海。 来源 | 芯东西 芯东西12月25日报道，今日，美国AI芯片独角兽Groq宣布，Groq的联合创始人Jonathan Ross、总裁Sunny Madra以及Groq团队的其他 成员将加入英伟达。 Groq今日发文宣布，其与英伟达达成一项 非独家授权协议 ，授权英伟达使用Groq的推理技术。该协议体现了双方共同致力于扩大高性能、 低成本推理技术的应用范围。作为该协议的一部分，Groq团队成员将加入英伟达，以帮助推进和扩大授权技术的规模。 Groq将继续作为一家独立公司运营，Simon Edwards将担任CEO。GroqCloud将继续正常运营，不会中断。 另据外媒报道，Groq投资方之一Disruptive的CEO Alex Davis透露，英伟达已同意以 200亿美元（约合1402亿元） 现金收购Groq的资 产。 需注意的是，英伟达并没有收购Groq，只是付费获得使用Groq技术的授权。 而为了协助英伟达，Groq公司 ...

当地时间周三，美国股市因传统节假日临近提前结束交易，市场整体交投清淡。反映市场对标普500指 数未来波动预期的"恐慌指数"VIX跌至一年来新低，表明投资者对未来短期出现风险事件的担忧情绪得 到缓解。同时，市场仍预计美联储明年至少降息两次，降息预期拉动房地产、金融等周期股普遍上涨， 美国三大股指周三集体上涨。截至收盘，道指涨0.60%，标普500指数涨0.32%，纳指涨0.22%。其中， 道指与标普500指数均创收盘历史新高。 24日国际金价微跌 白银期价再创新高 贵金属方面，国际金价在周二站稳每盎司4500美元关口并创下历史新高后，部分投资者选择获利了结， 周三金价微跌。截至收盘，纽约商品交易所明年2月交割的黄金期价收于每盎司4502.8美元，跌幅为 0.06%。 国际金价在周二站稳每盎司4500美元关口并创下历史新高后，部分投资者选择获利了结，周三金价微 跌。白银期价周三延续强劲涨势，连续第四个交易日创下历史新高。有分析人士指出，银价快速攀升过 程中包含大量投机头寸，投资者需警惕银价短线大幅回调的风险。 24日美国三大股指集体收涨 道指与标普500指数创收盘历史新高 为拓展推理芯片业务 英伟达挖走一AI初创 ...

Core Insights - The recent winning bid of 122 million yuan for the Longgang AI Phase I project highlights the strategic importance of integrating urban governance intelligent systems and robotic command platforms, which can create a sustainable operational framework for high-frequency, real-time edge reasoning needs [1][2] - The project signifies a shift from pilot demonstrations to a scalable, replicable model, allowing the company to address common urban governance challenges through standardized protocols and operational frameworks [2][3] - The focus on cost efficiency and controllable delivery in real-world systems positions the company favorably for long-term success in the competitive AI chip market [3] Company Strengths - The company emphasizes its self-developed NPU reasoning chips and toolchains, aiming to enhance efficiency and control in edge computing [2] - The project allows for the tangible manifestation of demand for AI chips, as it directly relates to operational metrics such as power consumption, latency, stability, and maintenance costs [2] - By creating a reusable product framework, the company can transform its chip and toolchain value into a default option for expanding applications across multiple cities [2] Industry Context - The significance of the Longgang project lies not in its district-level scale but in its ability to demonstrate the company's commitment to solving real cost and efficiency issues in operational systems [3] - As the industry matures, the competition will increasingly focus on unit costs and controllable delivery, which are critical for long-term viability [3]