文｜胡香赟 编辑｜海若镜 36氪获悉，AI制药全球领军企业深度智耀近期获近5000万美元D轮融资。本轮融资由鼎晖百孚领投，老股东新鼎资本、红杉中国持续加注，指数资本担任独 家财务顾问。募集资金将主要用于"多智能体协作网络"的技术研发迭代及全球交付网络的建设。 深度智耀成立于2017年。相较于早期在单一技术点上的探索，过去三年，深度智耀完成了从"单点AI技术验证"向"AI原生临床研究平台（AI-Native Platform）"的代际跨越。这一进化使其脱离了传统软件供应商的范畴，转型为能够交付临床试验全流程结果的核心业务伙伴。 在公司创始人、CEO李星看来，医药研发的未来不在于单一功能的替代，而在于认知的重构。随着生成式AI的爆发，深度智耀率先将底层的NLP能力升级为 拥有上万个垂直领域智能体的"多智能体协作系统"。 "我们不再是交付单一功能的模块，而是交付一个能够协同完成临床试验全流程的'AI智能体集群'。"李星对36氪表示："我们的核心壁垒在于用'认知原子 论'去重构研发流程，系统将复杂的临床试验拆解为上万个微小的原子化任务，每个任务由专精的Agent负责，它们通过类似脑神经的突触网络连接，实现了 远超通用 ...

Core Insights - The article discusses the strategic evolution of Qunke Technology from a 3D design software provider to a spatial intelligence infrastructure provider, emphasizing its new offerings in the AI and 3D space [1][2]. Group 1: Product Launches - Qunke Technology launched the Aholo Spatial Intelligence Open Platform, which integrates four core capabilities: spatial reconstruction, spatial generation, spatial editing, and spatial understanding, available via API or SDK for developers [5][7]. - The company also introduced LuxReal, a 3D AI content creation tool aimed at addressing the issue of temporal consistency in AI-generated videos, which often suffer from visual discrepancies [10][11]. Group 2: Technological Capabilities - Spatial reconstruction utilizes 3D Gaussian reconstruction technology to convert real-world scenes into navigable digital spaces, allowing users to create high-fidelity 3D environments quickly [5][6]. - Spatial generation automatically creates three-dimensional scenes based on text, images, or video inputs, while spatial understanding enables AI to comprehend spatial relationships and physical structures, crucial for robotics training [7][19]. Group 3: Market Applications and Collaborations - The Aholo platform is currently in beta testing, with applications spanning film, XR, cultural heritage preservation, industrial twins, and robotic simulation, including partnerships with Huace Film and PICO [7][8]. - LuxReal has begun global beta testing and aims to provide stable and realistic content for sectors like e-commerce, film, and advertising [13][12]. Group 4: Business Model and Market Challenges - Qunke Technology is adapting its business model from a traditional seat-based fee structure to a mixed model that includes annual fees, monthly fees, and usage-based charges, reflecting the shift towards AI product utilization [14][15]. - The company faces challenges in AI commercialization, emphasizing the need for products to demonstrate clear value and generate tangible business benefits for clients [14][19]. Group 5: Data and Training - Qunke Technology boasts the largest indoor spatial scene dataset globally, with 500 million 3D structured scenes and 440 million product models, which supports its AI capabilities [7][17]. - The company is also focusing on the development of synthetic data for robotics training, highlighting the significant investment required to create effective training datasets [17][19].

累计服务超过1000家药企， 并通过40000余个项目的实战交付。 "我们不再是交付单一功能的模块，而是交付一个能够协同完成临床试验全流程的'AI智能体集群'。"李星对36氪表示："我们的核心壁垒在于用'认知原子 论'去重构研发流程，系统将复杂的临床试验拆解为上万个微小的原子化任务，每个任务由专精的Agent负责，它们通过类似脑神经的突触网络连接，实现了 远超通用大模型的专业度。" 图源：深度智耀 这种技术架构的演进，在一定程度上推动了商业模式的变革。在行业普遍采用"按人头/工时付费"的传统模式下，深度智耀开始探索基于里程碑的价值付费 （Outcome-based Model）。 文 ｜ 胡香赟 编辑 ｜ 海若镜 封面来源 ｜ IC Photo 36氪获悉，AI制药全球领军企业深度智耀近期获近5000万美元D轮融资。本轮融资由鼎晖百孚领投，老股东 新鼎资本、 红杉中国持续加注，指数资本担任 独家财务顾问。募集资金将主要用于"多智能体协作网络"的技术研发迭代及全球交付网络的建设。 深度智耀成立于2017年。相较于早期在单一技术点上的探索，过去三年，深度智耀完成了从"单点AI技术验证"向"AI原生临床研究平台（A ...

Financial Data and Key Metrics Changes - Adobe reported record revenue of $23.77 billion for FY25, growing 11% year-over-year, with non-GAAP EPS of $20.94, reflecting a 14% increase [4][27] - In Q4, revenue reached $6.19 billion, a 10% year-over-year growth, with GAAP EPS of $4.45 and non-GAAP EPS of $5.50, increasing 17% and 14% year-over-year respectively [29][27] - Total Adobe ending ARR was $25.2 billion, growing 11.5% year-over-year, with digital media ending ARR at $19.2 billion, also growing 11.5% year-over-year [28][27] Business Line Data and Key Metrics Changes - Digital Media achieved revenue of $4.62 billion in Q4 and $17.65 billion for the full year, both growing 11% year-over-year [9][27] - Digital Experience segment revenue was $1.52 billion in Q4 and $5.86 billion for FY25, with subscription revenue growing 11% year-over-year [21][27] - The consumption of generative credits in Creative Cloud, Firefly, and Express increased approximately 3x quarter-over-quarter [32][78] Market Data and Key Metrics Changes - Monthly active users (MAU) for creative offerings surpassed 70 million in Q4, growing over 35% year-over-year [18][20] - The number of customers with over $10 million in ARR grew 25% year-over-year, exceeding 150 [32][27] - Digital Experience subscription revenue grew over 40% year-over-year, indicating strong market demand [22][27] Company Strategy and Development Direction - Adobe's strategy focuses on leveraging AI to enhance product offerings and customer engagement, with significant investments in AI-driven solutions [4][7] - The company aims to integrate AI deeply into its solutions, with a target of over 10% total Adobe ARR growth for FY26 [8][27] - The intent to acquire Semrush is part of Adobe's strategy to enhance brand visibility and address marketers' needs in the evolving digital landscape [23][34] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in Adobe's ability to deliver industry-leading innovation and double-digit ARR growth, driven by strong global demand for AI solutions [36][37] - The company highlighted the importance of customer experience orchestration and the growing role of AI in marketing and content production [21][22] - Management noted that Q4 showed strong indicators of growth, setting a positive tone for FY26 targets [63][66] Other Important Information - Adobe's cash flows from operations reached $10.03 billion, with a strong ending cash and short-term investment position of $6.6 billion [28][27] - The company executed record share repurchases totaling nearly $12 billion, reducing shares outstanding by over 6% during the year [27][28] - The introduction of Firefly Foundry and other AI-driven products is expected to unlock significant economic potential for Adobe [41][46] Q&A Session Summary Question: Interest in Firefly Foundry and its economic potential - Management discussed the early stages of customer engagement with Firefly Foundry, emphasizing its ability to create customized AI models for enterprises, which could significantly enhance content production efficiency [41][42][46] Question: ROI focus on new productivity enhancements - Management highlighted the opportunity to optimize content creation and delivery, emphasizing the unique ability to track the effectiveness of marketing campaigns through integrated analytics [50][51][53] Question: Strategic rationale behind the Semrush acquisition - The acquisition aims to enhance brand visibility for marketers across various channels, including LLMs and search engines, providing a comprehensive solution for customer experience orchestration [55][57][59] Question: Future ARR growth stabilization - Management expressed confidence in the momentum across all customer groups, indicating that Q4 was an inflection point for future growth, with a strong outlook for FY26 [63][66][70]

Group 1 - OpenAI's new image models Chestnut and Hazelnut are set to debut alongside GPT-5.2, but initial tests show they lag behind Google's Nano Banana Pro in generating high-quality images, particularly in facial rendering [1] - Mistral AI has released its next-generation code models, Devstral 2 and Devstral Small 2, achieving 72.2% and 68.0% on SWE-bench Verified, respectively, with a cost efficiency seven times higher than Claude Sonnet [2] - Zhiyu has launched the GLM-ASR-2512 cloud model and GLM-ASR-Nano-2512 edge model, achieving a CER of 0.0717, marking a significant advancement in speech recognition technology [3] Group 2 - Alibaba's Tongyi Lab introduced the Qwen-Image-i2L open-source tool, allowing personalized style transfer with just one sample, and offers various model variants optimized for different applications [4] - The Echo-N1 emotional model, with 32 billion parameters, outperformed a 200 billion parameter commercial model in multi-turn emotional support tasks, showcasing advancements in AI emotional intelligence [6] - The formation of the Agentic AI Foundation by major tech companies aims to establish interoperability standards for AI agents, with OpenAI contributing foundational standards already adopted by over 60,000 open-source projects [7] Group 3 - AI tools have been successfully utilized to design antibody-like molecules, with companies like Nabla Bio and Chai Discovery producing drug-like antibodies that target various diseases [8] - Anthropic's 14,000-word "AI Constitution" aims to guide AI behavior towards positive values, with a small team monitoring its real-world applications and potential risks [9]

Core Insights - The "EVOLVE 2025" summit showcased the roadmap for enterprise-level AI agents and introduced a "3+2+2" product matrix to facilitate rapid development of AI agents for businesses [1] - The summit emphasized the collaboration among major cloud service providers to create a sustainable AI ecosystem through the "Super Connection" global partner program [1] Group 1: AI Development Trends - Piero Scaruffi highlighted a clear trend of technological integration in generative AI by 2025, with innovations like diffusion Transformers and multi-modal capabilities becoming standard [3] - The emergence of new technologies such as thinking chains and expert mixtures is reshaping the landscape of AI applications [3] Group 2: Evolution of AI Agents - The distinction between traditional AI products and advanced AI agents was made, with the latter being likened to autonomous driving, capable of executing complex workflows independently [4] - The operational mechanism of these AI agents is summarized as a cycle of perception, decision-making, action, and learning, allowing them to adapt to various environmental changes [4] Group 3: Multi-Agent Systems - The transition from applications to multi-agent systems introduces challenges in orchestration, necessitating a new technology stack that includes hardware, cloud services, and orchestration layers [5] - The concept of "context engineering" is emphasized, requiring AI agents to understand organizational structures and goals beyond executing single tasks [5] Group 4: Industry Applications - Various sectors are witnessing innovative applications of AI, particularly in customer support, where intelligent systems can understand context and emotions, enhancing user experience [6] - Companies like Johnson Controls have developed integrated AI systems that significantly improve efficiency in maintenance and troubleshooting [6] Group 5: Trust in AI - The "Waymo effect" illustrates the growing trust in AI as autonomous vehicles become more prevalent, laying a foundation for broader AI agent applications [7] - Scaruffi envisions a future where multiple AI agents collaborate dynamically, akin to human social interactions, to achieve common goals [7]

根据一项至关重要的市场调研数据，约三分之二的华尔街金融服务公司在全公司范围采用人工智能应用之后， 最初可能将在实际经营模式中看到员工人数显著增加，这也让投资者们对于这一新技术是否会迅速带来大规模 成本节省产生了疑问。这项最新调查凸显出，人工智能或者机器人还不会大规模抢占工作岗位，至少现在不 会。 根据Bloomberg Intelligence面向华尔街各大机构的一项调查，超过70%的华尔街受访者表示，他们预计人工智能 驱动的应用软件将在未来三年内导致更高的经营成本。然而，大多数受访者仍然预计企业生产力将更快地提 高。该项针对美国高端金融服务行业151名高级员工的调查报告于周三正式发布。 "调查结果强化了一个观点，即人工智能在金融服务领域的最早期阶段，可能更多是关于能力普及与建设，而非 成本大幅削减。"Bloomberg Intelligence的资深分析师迪克莎.格拉(Diksha Gera)和托马斯.诺茨尔(Tomasz Noetzel) 在周三发布的报告中写道。"随着AI驱动自动化技术的规模化，我们认为成本比率在2027至2028年后可能会趋于 正常化，从而将大幅推动经营效率驱动的利润扩张。" 华尔街金 ...

欧盟方面表示，监管机构担心谷歌可能通过对出版商和内容创作者施加不公平条款，或为自身提供对相 关内容的特权访问，从而在训练大型模型时获取竞争者难以复制的数据优势。 外界认为，欧盟正试图在全球科技竞争中巩固对平台行为的规则引导权。 欧盟委员会认为，谷歌可能在创作者无法真正选择的情况下，使用上传至YouTube的视频训练自家的 Gemini与Veo3模型，而创作者在上传内容时被要求授予谷歌广泛的数据使用许可，使得"同意"带有默认 性质，缺乏现实的选择空间。 同时，谷歌禁止第三方公司使用YouTube视频训练模型，除非版权持有人明确授权，这使谷歌可能在训 练数据层面形成天然壁垒，进一步激化外界对其市场支配力的担忧。 对此，谷歌回应称，相关投诉可能抑制本已竞争激烈的市场创新，并强调其已与新闻和创意产业保持合 作，帮助他们适应AI带来的行业变化。 尽管谷歌公司否认有任何滥用市场地位的行为，但欧盟此次行动仍被视为欧洲近年来针对美国科技企业 监管升级的又一次体现。 欧盟委员会近日宣布将对谷歌展开正式调查，重点评估其在训练Gemini等人工智能（AI）模型时，使 用在线出版商内容以及YouTube创作者视频的方式是否违反了欧洲 ...

Core Insights - DeepSeek, a Chinese AI company founded in July 2023, has rapidly ascended to become the sixth largest unicorn globally, with a valuation of 1.05 trillion yuan, second only to ByteDance in China [1][2]. Company Performance - DeepSeek's explosive growth began in early 2025, with its app reaching 180 million monthly active users within a month of launch, and further increasing to 194 million by March [3]. - However, by May 2025, the monthly active users dropped to 169 million, and by September, it was surpassed by ByteDance's Doubao, which had 172 million users [3]. - The company released its DeepSeek-V3.2 model on December 1, 2025, achieving reasoning capabilities comparable to GPT-5 and close to Google's Gemini-3.0-Pro [3]. Competitive Landscape - The AI sector is witnessing intense competition, with major players like ByteDance and Alibaba investing heavily in AI infrastructure, with ByteDance spending 80 billion yuan in 2024 and Alibaba committing 380 billion yuan over three years [3]. - DeepSeek has adopted an open-source strategy, offering competitive API pricing, with input costs for DeepSeek-V3 as low as 0.5 yuan per million tokens, significantly cheaper than GPT-4 Turbo [6]. Technological Developments - The generative AI landscape is evolving with three main technological directions: text generation, image generation, and video generation [4][5]. - Major international players, including Google, are making significant advancements in generative AI, with Google launching multimodal models that enhance image and video quality [6]. Industry Transformation - AI is reshaping various industries, enhancing productivity in programming, transforming artistic creation, and revolutionizing the film industry [7]. - The emergence of new job roles such as AI trainers and prompt engineers reflects the changing job landscape due to AI integration [7]. Infrastructure and Energy - The competition in AI is increasingly tied to computational power and energy resources, with a shift from chip supply issues to energy shortages [8]. - China, possessing the largest power infrastructure and rapidly growing renewable energy capacity, is positioned to leverage its energy advantages for AI development [8]. Conclusion - DeepSeek's rise as a global AI unicorn highlights China's potential in the AI sector, driven by a unique approach to technology and market strategy [9]. - The global generative AI competition encompasses various dimensions, including technological breakthroughs and infrastructure development, with China developing a differentiated competitive edge [9].

Core Insights - The emergence of AI applications like Google's Gemini 3.0 and Ant Group's Lingguang APP highlights a significant shift in application development capabilities, allowing users to create applications using natural language [1][2][4] - Lingguang APP has seen rapid adoption, with users creating 3.3 million "flash applications" within just two weeks of its launch, showcasing the potential for mass creativity when development tools are accessible to the general public [1][4][6] Group 1: Application Development Trends - The simplicity and accessibility of "flash applications" appeal to users, similar to the rise of blogs and short videos, which thrived despite lacking the complexity of traditional media [2][6] - The traditional app market, with approximately 3 million apps available globally, contrasts sharply with the rapid creation of flash applications, indicating a new era of user-driven development [4][6] Group 2: Addressing Fragmented Needs - Many personal needs, such as customized weight loss plans or pet care applications, are difficult to meet with standard commercial apps, highlighting the demand for tailored solutions [3][5] - Lingguang's flash applications can address these fragmented and long-tail needs, providing a platform for users to create highly personalized tools [5][6] Group 3: Future Implications - The potential for a "flash application market" and integration with third-party applications through APIs suggests a transformative future for AI-driven application development [9] - The ongoing evolution of AI applications is expected to significantly enhance productivity and meet the diverse needs of users and businesses alike [6][9]