技术迭代+商业落地，AI多模态及AI应用将走向何方？

Core Insights - The rapid iteration of large models is driving an upgrade across the entire chain of computing power, data, and applications, reshaping future industries and investment landscapes [1] - The focus is shifting towards companies with a "moat and growth potential," particularly in infrastructure and vertical application sectors [2] Industry Trends - The demand for computing power is continuously rising due to the increasing model parameters and inference complexity, leading to a long-term upward trend in the demand for AI infrastructure such as high-performance GPUs and cooling systems [2] - The acceleration of application deployment is evident as stronger large models reduce the costs of business automation and intelligence, fostering new SaaS and vertical AI applications [2] - Competition in the ecosystem is intensifying, with leading model vendors forming platform capabilities and extending upstream and downstream, potentially increasing industry concentration [2] Company Strategies - Wanjing Technology has adopted a global strategy and focuses on empowering creators through creative software, aiming to package complex user needs into products that lower usage barriers [3] - The company has launched the Wanjing Tianmu multimedia large model, which has been upgraded to version 2.0, achieving industry-leading performance in quality, inference speed, and controllability through collaboration with Huawei [5] - Wanjing Technology's product matrix includes mobile apps, desktop solutions, and API services, catering to different user needs from casual creators to enterprises [6] Market Opportunities - The "productization" of large models is significantly impacting both B2B and B2C sectors, enhancing the value and sustainability of revenue streams [7] - For B2B, the evolution from tools to platforms allows for smarter data analysis and automation, increasing customer lifetime value and opening new commercial scenarios [7] - For B2C, large models enable scalable distribution of AI-generated content, enhancing user engagement and willingness to pay [7] Future Outlook - The emergence of multi-modal large models presents vast development potential, especially in video generation, which requires substantial computing power and presents higher technical barriers [9][10] - The AI landscape is expected to evolve towards a market-driven model, where companies with strong market presence and application scenarios will achieve commercial success [20] - The future of AI companies in China is promising, with a rich digital consumption market, a complete industrial chain, and strong policy support [21][22]