-+ 0.00%
-+ 0.00%
-+ 0.00%

Shang Tang (00020) Little Raccoon Financial Edition released, big model ability integrated into deep financial zones

Zhitongcaijing·11/20/2026 07:25:02
Listen to the news

Zhitong Finance App learned that on November 19, Shangtang Technology (00020) officially launched the “Shangtang Little Raccoon · Financial Edition” with the aim of deepening the scenario-based application of AI in the financial industry. Earlier, Shangtang Technology had just announced its business progress for the first three quarters of 2026. Revenue quality continued to improve, and the AI business commercialization process was further accelerated.

The direction of this set of actions by Shangtang Technology is very clear: on the one hand, it uses financial figures to prove growth, and on the other hand, it uses financial vertical products to show the specific path of the industry's implementation. It's easy to see that Shangtang Technology's commercial narrative has expanded from “selling model ability” to “selling and landing results.”

Let's look at the data first. For the nine months ended September 30, 2026, Shangtang Technology's revenue increased by 30.4% year on year to RMB 3.79 billion, a further acceleration from the first half of the year; gross margin increased 4.3 percentage points to 43.4%, and gross profit growth was faster than revenue growth. The growth rate of generative AI continued to be higher than the overall revenue growth rate of Shangtang Technology, which more than doubled year-on-year in the third quarter. On the infrastructure side, a 10,000-card heterogeneous inference cluster was built in the third quarter. Overall token throughput increased by about 5 times, and SenseCore's average daily token service volume increased from over 2.4 trillion yuan in July to 6 trillion yuan in October. This is a typical set of “scaling up, quality improving” financial data.

Let's take a look at the landing again. Relying on the Japan-Japan New Big Model System, Little Raccoon Financial Edition is a vertical version created by Shangtang Little Raccoon for the financial industry. It penetrates the five core scenarios of credit risk control, insurance claims, compliance management, business analysis, and investment and research insights. The customer case shows a quantifiable perspective: in public credit risk control cooperation with Ping An Bank, due diligence reports can be generated within 30 minutes as fast as possible, monitoring coverage exceeds 99%, and fund flow analysis was shortened from 4 hours to 15 minutes; in cooperation with Jianxin Life, the traditional data extraction and analysis process that took 3 to 5 days was shortened to the minute level. The cooperation with Pacific Insurance has completed actual testing in the group insurance project, effectively shortening the processing time of individual cases and identifying the risk of unreasonable compensation.

Putting the two things together, the real change is the delivery pattern. According to financial data, Shangtang Technology's AI infrastructure is expanding. Token throughput has increased 5 times, the average daily service volume has reached 6 trillion, and large devices rank first in the Chinese Neocloud market with a share of nearly 50%. The product side digs deep into the vertical field: Little Raccoon Financial Edition repeatedly emphasizes multi-modal document analysis, traceability of conclusions, easy verification, and enterprise-level security, controllability, and private deployment. The former is the scale of computing power, while the latter is delivered by the industry. The story Shang Tang Technology wants to tell is to integrate computing power advantages all the way into the business processes of different industries such as financial institutions, forming a closed loop from model to revenue, and the sustainability of revenue growth has a fulcrum.