Investment Rating - The report maintains an "Increase" rating for the computer industry [4] Core Insights - The global AI market continues to thrive, with increasing adoption rates in B-end applications and maturing business models. Companies are expected to increase their AI budgets by an average of 75% over the next year, indicating that AI is now essential for business operations rather than experimental [11][13] - Sovereign AI construction is becoming a crucial part of national technology strategies, with significant investments from countries like Saudi Arabia and the UAE in collaboration with NVIDIA to enhance digital sovereignty [28][29] - The demand for computing power is surging due to a dramatic increase in the usage of large models, with IDC projecting that the volume of model calls in China's public cloud will reach 114.2 trillion tokens by 2024 [32][35] Summary by Sections AI Adoption and Business Impact - A survey by a16z indicates that the share of innovation budgets allocated to AI has dropped from 25% to 7%, reflecting a shift towards integrating AI into core IT and business budgets [13][15] - OpenAI remains the leading model provider, with 67% of its users deploying non-cutting-edge models in production, compared to 41% for Google and 27% for Anthropic [15][16] - Software development has emerged as a key use case for AI, with reports showing that nearly 90% of code in a high-growth SaaS company is now generated by AI, up from 10-15% a year ago [21] Sovereign AI Development - The report highlights the growing importance of sovereign AI, with countries investing in AI infrastructure to ensure control over their AI systems. For instance, Saudi Arabia's HUMAIN and the UAE's Khazna data center are collaborating with NVIDIA to build AI factories [28][29][31] - Chinese tech companies are also positioning themselves in the sovereign AI landscape, with initiatives aimed at establishing a "Digital Silk Road" strategy to provide AI infrastructure solutions in Southeast Asia, the Middle East, and Africa [31] Computing Power Demand - The report notes a significant increase in the daily token usage of large models, with the Doubao model reaching over 16.4 trillion tokens, a 137-fold increase from the previous year [38] - Google has reported a 50-fold increase in token processing, with its products and APIs handling over 480 trillion tokens monthly [41] - IDC's findings indicate that 80% of enterprises plan to adopt more generative AI in the next 18 months, with a focus on improving operational efficiency and customer satisfaction [32][35] Investment Opportunities - The report suggests focusing on companies involved in computing power, such as Cambrian, Alibaba, and Huagong, as well as those in the AI agent space like Alibaba and Tencent [7][46] - The automotive sector is also highlighted, with companies like Xpeng and Li Auto being noted for their advancements in autonomous driving technology [47]
全球AI持续景气