Recently, JPMorgan released a report initiating coverage on Xunce Technology (3317.HK) with an "Overweight" rating and a target price of HK$160, implying approximately 58% upside from the closing price of HK$101.10 on September 23.

The report is bullish on Xunce's position in the enterprise AI implementation layer: as models become cheaper and more capable, enterprise adoption increasingly depends on reliably connecting proprietary data, real-time business context, production systems, and AI pipelines. Xunce has built these capabilities through a decade of real-time data deployment and is now expanding them from project delivery to reusable modules and usage-based token monetization.

This opens up a growth path that relies more on deepening consumption among already-deployed customers, while product reuse can improve delivery economics as the business scales. On the financial side, the first half of 2026 not only achieved rapid cross-industry growth but also recorded large-scale token revenue and positive adjusted profit.

In the first half of 2026, Xunce's revenue grew 389% year-on-year to RMB 967 million, with net profit attributable to shareholders of RMB 72.5 million and adjusted net profit of RMB 67 million, achieving its first profitable half-year. Customer ARPU increased 240% year-on-year to RMB 5.56 million, customer retention remained above 90%, and revenue per employee rose 379% year-on-year, reflecting significantly improved operational efficiency.

JPMorgan noted that the main upside for Xunce lies in the change in revenue quality. The revenue model is evolving from two factors (number of customers × ARPU) to four factors (number of customers × number of modules × token unit price × usage volume), with the new factors directly tied to actual business outcomes, allowing AI workflow penetration among existing customers to continuously contribute marginal revenue.

Token usage is becoming a new revenue driver, while an expanding library of reusable modules can reduce the incremental work required for each new deployment. If this approach works, the business can gradually shift from customer growth requiring additional engineering to incremental revenue generated from existing deployments, thereby improving margins and enhancing operating leverage.

Meanwhile, JPMorgan published an in-depth report on China's AI industry chain, pointing out that the scarcity in the AI value chain is migrating from upstream computing power and model capabilities toward enterprise deployment and usage-based monetization; it is estimated that China's inference token consumption in 2030 will reach about 60 times the 2026 level, with the enterprise segment accounting for approximately 69%.

With its positioning in the enterprise AI deployment layer, Xunce Technology is well-positioned to benefit from this migration of scarcity. JPMorgan estimates that Xunce Technology's revenue CAGR from 2026 to 2028 will be approximately 91%, with adjusted net profit of approximately RMB 2.3 billion in 2028. If token expansion, module reuse, and improving customer economics translate into sustained profit and cash flow growth, the company's share price could have further upside.