Large Models Are Advancing Swiftly, Yet AI Applications Lack Momentum
2 hour ago / Read about 0 minute
Author:小编   

MiniMax unveiled its inaugural semi-annual report following its public debut, disclosing that its revenue for the first half of the year soared to USD 117 million, marking a staggering 283.1% year-on-year increase, while losses contracted by 11% compared to the same period last year. The impetus for business expansion has transitioned from the consumer (C) segment to the business (B) segment. This shift underscores the growing disparity in commercial revenue between the AI model layer and the application layer: revenue for foundational model firms has surged into the tens of billions of dollars, whereas AI application companies boasting revenue surpassing USD 100 million remain exceedingly scarce. The count of leading products with monthly active users exceeding one million is also on the decline, constraining the growth potential of the middle application layer. Despite the swift expansion of the AI application user base, a number of AI applications that were once highly regarded by investors have exited the market. Leading AI applications are grappling with growth challenges, characterized by ambiguous profit models, and no universally acknowledged national-level products have emerged. The development of the AI application layer is being stifled by foundational model companies, with super apps engulfing (or absorbing) the vertical application market. Most AI applications exhibit a high degree of reliance on dominant large models, lacking autonomy. Users frequently switch tools due to variations in model performance, and the industry's attention is gradually shifting towards the model layer and intelligent agents. Concurrently, capital enthusiasm for the AI application layer is diminishing, with more funds being channeled into AI infrastructure and large model domains, resulting in a reduction in financing scale for the application layer. Some AI applications are encountering difficulties due to funding shortages and challenges in retaining users, and their transformative value has not been fully realized. Additionally, there are concerns regarding user trust in the direction of intelligent agents.