UNISOUND (09678) is scheduled to release its 2025 annual results on March 26. A performance forecast issued in January this year provided an early indication, showing a more than tenfold surge in revenue from its large model business, which now accounts for nearly half of total revenue. The upcoming official results will serve as a genuine stress test for its business model viability.
In the current AI sector, numerous companies are labeled as "large model concept stocks," with share prices fluctuating alongside technical parameter announcements. However, as market sentiment cools, there is growing clarity that leading technological metrics, if not translated into a commercial cycle, ultimately represent little more than a "specification race" on presentation slides.
How can genuine "strong performers" be identified? The answer lies in three critical questions: Can the business model achieve sustainable operations? Can a business cycle be established? Can a financial cycle be realized?
From Forecast to Fulfillment: Large Model Business Becomes the "Engine" According to the performance forecast, UNISOUND's large model-related business revenue for 2025 is projected to reach between 600 million and 620 million yuan, representing a year-on-year increase of 1057% to 1095%. Its proportion of total revenue is expected to rise to between 48% and 53%. The large model business has evolved beyond a mere "technical showcase" to genuinely supporting nearly half of the company's revenue.
Key Focus Areas for the Results Announcement: Monitor Three "Cycle Closure" Metrics If the forecast demonstrates the "speed of growth," the official results must focus on the "quality of growth."
First, examine the "business cycle closure": the transition from project-based work to productization. With the maturation of the "Shan Hai" large model and the unified intelligent agent platform, has the company successfully shifted from "project delivery" to "product subscriptions and service output"? Are customer retention rates and Average Revenue Per User (ARPU) consistently improving in vertical scenarios such as smart healthcare and smart transportation? These are core indicators for testing whether a true business cycle has been formed.
Second, assess the "financial cycle closure": moving from "one-time monetization" to "maximizing customer value." True AI monetization should not stop at one-off project revenue. Instead, it should maximize value from the same customer through layered monetization—from model API calls, to scenario subscriptions, and then to value-added services. For this results announcement, investors should closely monitor whether net operating cash flow has significantly improved and if adjusted net losses have substantially narrowed. These factors directly determine the company's sustainable "self-funding capability" and are key evidence of a realized financial cycle.
Third, evaluate the "business model evolution": transitioning from linear growth to exponential growth. UNISOUND's implementations in critical sectors like healthcare, medical insurance, and transportation are building an "exponential" growth engine. As AI becomes integral infrastructure for core workflows—such as hospital medical record quality control, medical insurance cost containment, and smart cockpits—the frequency of API calls and depth of scenario penetration will replace traditional customer number accumulation, acting as a "compound interest engine" for revenue growth.
Conclusion: The Second Half of the AI Era is About "Longevity" Capital markets are no longer providing unlimited funding based on "technological idealism" but are increasingly focused on the practical strength required for "survival through implementation." UNISOUND's results announcement on March 26 is essentially a comprehensive examination of its substance: Has technology been genuinely transformed into products? Are these products truly being paid for by the market? Is the revenue genuinely generating cash flow?
Comments