ARR quarterly growth rate surges by 300%, XUNCE (03317) Token launches pay-per-use AI commercialization paradigm.
A true transformation about the commercial path of AI is quietly happening on the "Token First Stock" Xunze (03317).
On May 12th, Morgan Stanley released a report stating that China's artificial intelligence (AI) is entering a new stage, shifting from catching up in technological capabilities to capturing value. The focus is shifting from training to reasoning, from technology to application, and from potential to actual profitability. This assessment aligns with macro policy guidance. Recently, the People's Bank of China explicitly pointed out in the "2026 First Quarter Report on Chinese Monetary Policy Implementation" that artificial intelligence will inject new momentum into high-quality economic development through accelerated technological iteration, deepened industrial integration, and expanded international market.
A real transformation of the commercialization path of AI is quietly happening with XUNCE (03317), known as the "First Stock of Tokens".
Business model upgrade: Token revenue share expected to increase from 5% to 30%
Financial data shows that in 2025, XUNCE Technology achieved operating revenue of 1.285 billion RMB, a year-on-year increase of 103.28%. The revenue in the first half of the year was only 198 million RMB, but it surged to 1.087 billion RMB in the second half, a staggering increase of 449% from the previous half, signaling a clear explosion of demand. The ARPU value of customers in 2025 increased from 2.72 million RMB in 2024 to 5.59 million RMB, a year-on-year growth of 105.4%; the average revenue per customer increased by 135.25% to 2.87 million RMB; and the gross profit margin for the whole year remained at a high level of 61.66%, significantly higher than that of power computation manufacturers and model manufacturers.
In April 2026, XUNCE's Token-called ARR quarterly soared by 300% in a single quarter, and the proportion of Token revenue exceeded 5%, with the annual target set to increase to the range of 20% to 30%. The significance of this set of data lies not in the absolute scale, but in the signal of a qualitative change in the business model: the company is completing a systemic transition from traditional subscription to Token payment and revenue sharing models.
XUNCE's management also stated in a recent interview: "The core logic behind this goal is that the way customers use our services has changed. In the past, customers mostly paid for modules, solutions, or projects; but when AI truly enters business processes, customers are more concerned about how much has been actually utilized and the results and business value generated."
According to XUNCE's pricing logic, Token payment is not simply based on "pay per word", but is a combination of three dimensions: the price of a single call (depending on data scarcity, real-time requirements, and industry complexity), the number of calls (the actual usage by customers in real business scenarios), and the depth of modules (the more modules accessed and embedded into the process, the higher the overall value). Essentially, this mechanism links the method of payment directly to the actual value created by customers - "the deeper the use, the more calls, and the higher the demand for results, the greater the value of the data capabilities we provide."
The company's management expects that with the explosion of AI reasoning demand, the Token-based payment model will gradually become a significant part of the company's revenue structure in the next two to three years.
During the AI reasoning phase, vertical Token premiums
The requirements for data quality in the AI reasoning phase are fundamentally different from the training phase: it requires real-time, accurate, and measurable data supply, rather than massive but coarse-grained historical data sets. XUNCE's core competitiveness lies exactly in this aspect.
Having been deeply involved in the real-time data infrastructure field for 10 years, XUNCE Technology has built a full-chain technology system from data acquisition, cleaning, standardization, real-time calculation to model optimization around the AI Data Agent, covering ten high-growth tracks such as finance, Siasun Robot & Automation, and commercial aerospace. With its accumulated private data operations in high-threshold industries, XUNCE has become a "high-quality data fuel" that can be consumed at the application layer, model layer, and even operating system level in AI, enhancing the business value of each Token unit and forming a solid competitive barrier.
More importantly, the daily Token call volume in China has exceeded 140 trillion times, increasing over a thousand times in two years; JPMorgan Chase predicts that AI reasoning Token consumption will grow 370 times by 2030; IDC expects global AI Agent growth to increase by 139% in five years. Each intelligent entity running 24/7 is a new engine for vertical Tokens.
Currently, the pricing of vertical Tokens at XUNCE has reached $10 to $100 per million Tokens, more than ten times of similar products by Anthropic. This pricing capability is built upon data barriers accumulated from nearly ten high-value industry scenarios in China.
The company's management stated, "We will further expand the AI training data platform for Siasun Robot & Automation, the commercial satellite computing platform, and the aviation industry computing platform - each of them represents new billion-dollar tracks, and the application scenarios in these areas in the Chinese market are becoming richer."
The core characteristic of the Chinese AI "value capture" stage described in the Morgan Stanley report is that AI capabilities are no longer just about showing off technically, but are truly penetrating into core business processes of enterprises and forming sustainable commercialization. XUNCE's upgrade in Token payment model is a microcosmic reflection of this characteristic.
As the management at XUNCE summarized the company's value proposition: "Big models determine what AI can do, but XUNCE determines whether AI can continuously, reliably, and scalably create value in real enterprise scenarios." From "selling data" to "selling Token calls", XUNCE's performance and stock price are expected to experience the "Davis Double" and aim for a market value of 200 billion RMB!
Related Articles

Shanghai Shibei Hi-Tech (600604.SH) plans to invest 40 million yuan in the Bilin Star Fund.

Investment banking advisor company Lincoln International (LCLN.US) plans to price its IPO at $18-20 per share, with a target fundraising of $400 million.

South Korean AI disaster prevention software company Roze AI (ROZE.US) files to list on NASDAQ for direct listing.
Shanghai Shibei Hi-Tech (600604.SH) plans to invest 40 million yuan in the Bilin Star Fund.

Investment banking advisor company Lincoln International (LCLN.US) plans to price its IPO at $18-20 per share, with a target fundraising of $400 million.

South Korean AI disaster prevention software company Roze AI (ROZE.US) files to list on NASDAQ for direct listing.

RECOMMEND

Two Mainland Accounting Firms Approved for H‑Share Audits, Lowering Listing Costs and Deepening Mainland–Hong Kong Market Integration**The Ministry of Finance, the CSRC, and Hong Kong’s Accounting and Financial Reporting Council have approved two additional mainland accounting firms—RSM China and ShineWing—to conduct H‑share audit work, marking the first expansion of the list since 2010.
11/05/2026

HKEX Tightens Rules on Auditor Dismissals as Sudden “Audit Firm Switches” Raise Governance Concerns
11/05/2026

The Chip Stock Frenzy Is Still Accelerating
11/05/2026


