Recently, International Data Corporation (IDC) released its latest analysis report on the landscape of China's enterprise-grade MaaS market for the first half of 2026. The report shows that in H1 2026, the China MaaS market reached RMB 10.63 billion in size, up nearly 9 times year-over-year; in the large model privatization market, SenseTime SenseCore once again firmly holds second place with a 9.4% market share, continuing to lead the industry's top tier.
The report focuses on the changes in the landscape of China's enterprise-grade MaaS market in the first half of this year, concluding that three major trends—model upgrades, scenario scale-up, and modality expansion—have driven explosive growth in the size of the China MaaS market. For cloud vendors and model vendors alike, MaaS has become a primary growth battleground that cannot be lost.
Among this, as an important component of the MaaS market, the large model privatization market has also grown rapidly, with market size reaching RMB 2.88 billion in the first half, up 108.4% year-over-year.
IDC emphasizes that customers in the large model privatization market are shifting their focus from single-point model capabilities to the full capabilities of platforms. Data security, business customization, independent controllability, and Xinchuang (IT innovation) adaptation are all indispensable, while engineering delivery and long-term operations have become key to project success. At the same time, customers place greater value on AI's actual support for business, and have begun measuring return on investment by business metrics rather than model leaderboards.
As a representative vendor in the large model privatization market, SenseTime SenseCore continues to build end-to-end multimodal solutions, and through the FDE (Forward Deployed Engineer) model goes deep into the front lines of industry, working with customers to define scenarios, co-create models, and share goals, driving AI capabilities to truly enter enterprises and be transformed into deployable, measurable business value. It now covers numerous industry scenarios including urban governance, AIGC, embodied intelligence, and AI for Science.
Taking the "Yunyu Xingkong (Cloud-Universe Starry Sky) Foundation Model" as an example, it is the first foundation model in the national planning and resources field. In actual project delivery, SenseTime SenseCore starts from customers' real business needs, carrying out accompanied co-creation through expert services and the FDE model, working together with customers to sort out business processes, define problem standards, and verify application results, forming a closed loop of "problem definition—data and knowledge preparation—agent development—business verification—feedback iteration." Ultimately, it achieved a 60% reduction in business processing time and a 70% increase in per-capita output efficiency, effectively unleashing the creativity of professionals.
To accelerate the intelligent transformation of central and state-owned enterprises as well as local industries, SenseTime SenseCore has further distilled complex large model projects into replicable, scalable methodologies, forming the AI "123" delivery rule—clarify business value in 1 day, complete Demo verification in 2 days, and achieve stable version launch in 30 days.
This model has been validated in multiple benchmark scenarios. In addition to the "Yunyu Xingkong Foundation Model," in the advanced manufacturing field it has served 5 large state-owned enterprises in Shanghai, deployed in dozens of industry scenarios, and won four awards at the "CICC 2026 State-Owned Assets Scenario Openness and Industry Application Challenge," fully validating the industry value of this model.
Facing the future, SenseTime SenseCore will continue to deepen AI infrastructure construction, bridge the last mile of enterprise AI implementation, and accelerate the large-scale deployment of AI capabilities in real business scenarios.
This article was provided by SenseTime and reprinted by QbitAI with authorization; the views belong to the original author.