Core Insights - The 2025 Inclusion Bund Conference in Shanghai focused on the development path of artificial intelligence (AI), discussing its current status, challenges, and future vision [1] AI Development - AI is transitioning from a data-driven paradigm to an experience-driven one, as proposed by Turing Award winner Richard Sutton, indicating a new phase in AI development [2] - The "scale law" continues to dominate AI development, with the emergence of reasoning models shaping a new curve called the "reasoning scale law" [4] - Major U.S. tech companies are expected to spend over $300 billion on AI-related capital expenditures by 2025, indicating a large-scale construction boom in the AI data center industry [4] - The concept of an "intelligent agent economy" is emerging, where numerous intelligent agents interact, execute tasks, and exchange data [4] - Open resources are becoming a key variable in AI competition, with a shift from code openness to resource openness [4][7] AI Challenges - Energy demand is a hard constraint for AI development, with AI currently consuming 1.5% of global electricity, potentially rising to 20% [5] - There is a significant gap in the practical application of AI, with challenges in high-quality data availability and model alignment with robotic control modalities [6] - Ethical and social governance challenges are increasingly prominent, with concerns about decision-making being transferred from humans to algorithms [6] - Organizational management needs to be restructured to adapt to the rapid development of AI technology [6] AI Future - The ultimate goal of AI is linked to energy, with nuclear fusion being highlighted as a breakthrough opportunity [8] - Continuous learning and meta-learning technologies are essential for unlocking the full potential of AI [8] - Collaboration and empathy are crucial for measuring progress in a rapidly evolving technological society [8] - The launch of 12 satellites with an 8B AI model marks a significant opportunity for AI in space [8][9] - The future of AI will require a collaborative approach involving technological breakthroughs, energy support, ethical norms, and social governance [10]

Group 1 - The 2025 Inclusion Bund Conference opened in Shanghai with the theme "Reshaping Innovative Growth," featuring 550 guests from 16 countries and regions, including notable figures like Richard Sutton and Yuval Noah Harari [1] - The conference focused on five main topics: "Financial Technology," "Artificial Intelligence and Industry," "Innovation and Investment Ecosystem," "Global Dialogue and Cooperation," and "Responsible Innovation and Inclusive Future," with a total of 44 forums and a technology exhibition [1] Group 2 - Zhang Hongjiang, a partner at Source Code Capital, discussed the emergence of the "agent swarm" era, where numerous intelligent agents interact and exchange tasks and information, leading to the concept of "agent economy" [2] - He emphasized that models and GPU computing power will become core assets for organizations, urging companies to enhance their computing power and data richness [2] - The integration of "super individuals" and agents is expected to bring significant structural changes to business processes [2]

Core Insights - The "Scaling Law" for large models remains valid, indicating that higher parameter counts lead to better performance, although the industry perceives a gradual slowdown in pre-trained model scaling [3] - The emergence of reasoning models has created a new curve for large-scale development, termed "reasoning scaling," which emphasizes the importance of context and memory in computational demands [3] - The cost of using large language models (LLMs) is decreasing rapidly, with the price per token dropping significantly over the past three years, reinforcing the scaling law [3] - AI is driving massive infrastructure expansion, with significant capital expenditures expected in the AI sector, projected to exceed $300 billion by 2025 for major tech companies in the U.S. [3] - The AI data center industry has experienced a construction boom, which is expected to stimulate the power ecosystem and economic growth, reflecting the core of "AI industrial scaling" [3] Industry Transformation - Humanity is entering the "agent swarm" era, characterized by numerous intelligent agents interacting, executing tasks, and exchanging information, leading to the concept of "agent economy" [4] - Future organizations will consider models and GPU computing power as core assets, necessitating an expansion of computing power to enhance model strength and data richness [4] - The integration of "super individuals" and agents is anticipated to bring about significant structural changes in enterprise processes [4]

Group 1 - The core viewpoint of the discussions at the 2025 Inclusion Bund Conference emphasizes the rapid development of AI technology and the opportunities it presents for innovation in various fields [1][2] - Wang Xingxing, CEO of Yushu Technology, expressed regret for not focusing on AI development earlier and highlighted the current state of AI applications as being on the verge of explosive growth, particularly in the realm of data utilization [1] - A significant challenge identified is the difficulty in determining the standards for high-quality data and how to effectively collect it, which is crucial for improving data utilization rates [1] Group 2 - Zhang Hongjiang, a partner at Source Code Capital, introduced the concept of the "agent swarm" era, where numerous intelligent agents interact and execute tasks, leading to the emergence of an "agent economy" [2] - The interaction between humans and these agent swarms is expected to reshape enterprise processes, indicating a potential for substantial structural changes in business operations [2]

Group 1 - The core principle of "Scaling Law" for large models remains valid, indicating that higher parameters lead to better performance [2] - The emergence of reasoning models has created a new curve for large-scale development, termed "reasoning scaling" [2] - The rapid decline in the cost per token for large language models (LLM) over the past three years will further reinforce the scaling law [2] Group 2 - AI is driving large-scale expansion of infrastructure, with the AI data center industry experiencing significant construction activity over the past year [2] - The large-scale construction in the IDC industry will stimulate the power ecosystem and economic development, reflecting the core of "AI industrial scaling" [2] Group 3 - Humanity is entering the "agent swarm" era, characterized by numerous agents interacting, executing tasks, and exchanging information [3] - The interaction between humans and agent swarms will form the basis of the "agent economy" [3] - Models and GPU computing power will become core assets for future organizations, necessitating the expansion of computing power to enhance models and enrich data [3]

Core Insights - The "Artificial Intelligence +" initiative in China is expected to drive significant transformations across various industries, with AI projected to contribute over one trillion yuan to the global economy by 2030, becoming a key growth driver [1] Group 1: AI in Research and Development - The initiative will accelerate the application of large scientific models and promote the intelligent upgrade of fundamental research platforms and major infrastructures, including the sharing of high-quality scientific datasets [3] - AI technology is being utilized in new material development for aerospace applications, significantly reducing the design cycle from over ten years to under three years for new 3D printed steel alloys [3] Group 2: Innovation and Collaboration - The initiative supports the promotion of intelligent R&D tools and platforms, fostering collaborative innovation in fields such as biomanufacturing, quantum technology, and 6G [5] - AI is expected to replace many experimental processes with simulations, potentially leading to breakthroughs in Chinese scientific research with growth rates of hundreds or even thousands of times [5] Group 3: Industrial Development - The initiative aims to advance the intelligent development of all industrial elements, nurturing a new generation of AI-native enterprises [6] - China is positioned to leverage its strengths in manufacturing by solidifying existing processes with intelligent systems, leading to flexible digital workflows and organizational structures [6] Group 4: Consumer Quality Enhancement - The initiative will accelerate the integration of AI with technologies such as the metaverse, low-altitude flight, and brain-computer interfaces, enhancing product innovation [8] - AI is seen as a means to democratize productivity, with applications in logistics, including drone delivery and emergency response, presenting new business models and governance frameworks [8]

Core Viewpoint - The development of humanoid robots is a significant technological direction aimed at taking over repetitive physical labor in society, with the potential for widespread adoption similar to automobiles, benefiting humanity as a whole [1][2]. Industry Overview - The humanoid robot industry in China is accelerating through a dual-cycle model of "application validation - technological breakthroughs," driven by national policy guidance and attracting diverse manufacturers to collaboratively build an industrial ecosystem [1]. - The commercial sales volume of humanoid robots in China is expected to reach approximately 2,000 units in 2024, with projections of 60,000 units by 2030, reflecting a compound annual growth rate of 95.3% [1]. Innovation and Technological Breakthroughs - Innovation is identified as the core engine driving the transformation of humanoid robots, creating a pivotal "ChatGPT singularity moment" that will ignite market recognition and industrial enthusiasm [2]. - Continuous technological breakthroughs in humanoid robots include advancements in dexterous limb movement control, enhanced environmental perception, embodied intelligence, improved energy efficiency, and rapid reductions in hardware costs [1]. Company Innovations - **Yushun Technology**: Launched the Unitree R1 humanoid robot at a starting price of 39,900 yuan, featuring 26 joints and a weight of approximately 25 kg, aimed at developers and educational institutions [3][5]. - **Zhejiang Humanoid**: Showcased a full-size humanoid robot capable of precise operations across various terrains, demonstrating robust operational capabilities [5][7]. - **Zhiyuan Robotics**: Released the Lingxi X2 humanoid robot with enhanced interaction and control capabilities, priced between 100,000 to 300,000 yuan [11][13]. - **Zhuji Power**: Introduced the TRON 1 dual-mode humanoid robot, expanding its capabilities for complex terrain navigation and research tasks [17][19]. - **UBTECH**: Achieved multiple innovations in hardware and software, including a hot-swappable battery system for continuous operation and a multi-modal reasoning model for enhanced decision-making [20][22]. Market Dynamics - The humanoid robot industry is expected to experience a non-linear growth trajectory, leading to a "burst singularity" where capital, talent, supply chains, and application scenarios converge at unprecedented speeds [2].

Core Insights - The 13th ISC.AI Internet Security Conference highlighted AI as a central theme, with a focus on intelligent agents as the future of the industry [1] - Zhou Hongyi, founder of 360 Group, predicts that 2025 will be the year of intelligent agents, which will transform organizational operations and economic structures [1] Group 1: Intelligent Agent Levels - Intelligent agents are categorized into five levels (L1-L5), with L1 being chat assistants like ChatGPT, which have limited functionality for enterprise needs [2] - L2 agents are workflow agents that automate standard operating procedures but lack flexibility, primarily used in repetitive tasks [4] - L3 agents are reasoning agents driven by large models, capable of planning tasks autonomously but face challenges in executing complex instructions [3] Group 2: Evolution and Economic Impact - The evolution of intelligent agents is driven by three technological breakthroughs: DeepSeek for reasoning model proliferation, MCP for tool interface standardization, and Manus for L3 agent development [5] - The emergence of "intelligent agent economy" is anticipated, where human roles shift from executors to managers of intelligent agents, creating new value [5][6] Group 3: Human-Agent Interaction - Future human-computer interaction will fundamentally change, with users becoming leaders of intelligent agents rather than mere operators of software [6] - Intelligent agents will take on roles such as shopping and negotiation, leading to a shift in how tasks are managed [7] Group 4: Security Implications - Intelligent agents can address talent shortages in cybersecurity by acting as virtual experts, providing 24/7 threat assessment [7] - However, there are risks of adversarial use of intelligent agents for cyberattacks, necessitating a dynamic defense system [7]

Core Viewpoint - Google Cloud's donation of the A2A (Agent-to-Agent) protocol to the Linux Foundation has sparked significant interest in the AI industry, indicating a strategic response to competitors like Anthropic's MCP protocol and OpenAI's functions, while highlighting the industry's consensus on the need for foundational rules in the agent economy [1][4]. Summary by Sections A2A Protocol and Industry Response - The A2A protocol includes agent interaction protocols, SDKs, and developer tools, backed by major tech companies like Amazon, Microsoft, and Cisco [1]. - The decision to donate A2A is seen as a strategic move against competing protocols, emphasizing the necessity for collaborative foundational rules in the AI sector [1][4]. MCP Protocol Insights - MCP focuses on enabling AI models to safely and efficiently access real-world tools and services, contrasting with A2A's emphasis on agent communication [4]. - Key aspects of developing an MCP Server include adapting existing API systems and ensuring detailed descriptions of tools for effective service provision [7][8]. Development Scenarios for MCP - Two primary scenarios for implementing MCP services are identified: adapting existing API systems and building from scratch, with the latter requiring more time for business logic development [8][9]. - The importance of clear tool descriptions in the MCP development process is highlighted, as they directly impact the accuracy of model calls [13]. Compatibility and Integration Challenges - Compatibility issues arise when integrating MCP servers with various AI models, necessitating multiple tests to ensure effective operation [10][11]. - The need for clear descriptions and error monitoring mechanisms is emphasized to identify and resolve issues during the operation of MCP systems [14]. Future Directions and Innovations - The MCP protocol is expected to evolve, with predictions that around 80% of core software will implement their own MCPs, leading to a more diverse development landscape [40]. - The introduction of the Streamable HTTP protocol aims to enhance real-time data handling and communication between agents, indicating a shift towards more dynamic interactions [15][40]. A2A vs MCP - MCP primarily addresses tool-level issues, while A2A focuses on building an ecosystem for agent collaboration, facilitating communication and discovery among different agents [32][33]. - The potential for A2A to create a more extensive ecosystem is acknowledged, with plans for integration into existing products and services [34][35]. Security and Privacy Considerations - The importance of safeguarding sensitive data in MCP services is stressed, with recommendations against exposing private information through these protocols [28]. - Existing identity verification mechanisms are suggested to manage user access and ensure data security within MCP services [28]. Conclusion - The ongoing development of both MCP and A2A protocols reflects the industry's commitment to enhancing AI capabilities and fostering collaboration among various agents, with a focus on security, efficiency, and adaptability to evolving technologies [40][43].

Summary of Key Points from Conference Call Records Industry Overview - The conference primarily discusses the **computer industry** and **AI technology** developments, particularly focusing on the transition from training to application in AI investments, with a significant emphasis on the **inference demand** expected to exceed 70% of overall computing power needs by 2025[1][2]. Core Insights and Arguments - **AI Investment Shift**: The investment logic in AI is shifting from training to application, with inference demand projected to grow significantly, indicating a widening supply-demand gap in computing power[1][2]. - **Market Performance**: The computer industry experienced a "rise and fall" trend in the first half of 2025, with initial optimism driven by the release of DeepSeek, which later faced a market correction due to underperformance expectations for 2024[4][5]. - **Financial Metrics**: The computer industry showed year-on-year revenue improvement, but the net profit growth rate outpaced revenue growth due to significant cost optimization. However, the overall asset-liability ratio is rising, and ROE is declining, indicating the industry is still in a bottom-seeking phase[6][7][8]. - **AI Agent Technology**: AI Agent technology has made unexpected advancements in environmental perception, planning, tool usage, and memory capabilities, but the actual product deployment and user adoption remain below expectations due to the absence of a "killer app"[10][12]. - **DeepSeek R2 Release**: The anticipated release of DeepSeek R2 is expected to catalyze AI development in the second half of 2025, with potential improvements in computing power efficiency and performance[13][14]. Additional Important Insights - **Global Supply-Demand Gap**: The global supply-demand gap for inference computing power is expected to continue expanding, with significant demand for H200 GPUs projected at approximately 3.8 million units in 2025 and over 13 million units in 2026[3][16][17]. - **Investment Opportunities**: Current investment opportunities in the AI industry are concentrated in areas such as NVIDIA's computing power chain, domestic AI application ecosystems, and AI Agent application tracks[18][19]. - **Solid-State Battery Market**: The solid-state battery market is entering a production phase in 2025, but its penetration rate remains low due to the dominance of traditional liquid electrolyte batteries. The transition to solid-state technology is expected to accelerate in specific applications, particularly in electric vehicles[20][23]. - **Technological Innovations**: Innovations in solid-state battery manufacturing processes, such as dry electrode technology, are identified as key investment areas, alongside the evolving roles of separators and electrode materials in battery performance[24][25][26][27][28]. Conclusion - The conference highlights a transformative period for the computer and AI industries, with significant shifts in investment focus, technological advancements, and emerging market opportunities. The anticipated developments in AI applications and solid-state battery technologies are expected to shape future investment landscapes.