Xunce Technology Delivers Exceptional Interim Results with 389% Surge, Validating the Value of “Genuine AI”

The Hong Kong-listed AI application players have long struggled to achieve meaningful profitability. Xunce (03317.HK, the “Company”) defies the industry trend with a blowout earnings performance, proving that AI can deliver sustained, powerful and high-margin profitability. Its first-half revenue nearly quadrupled year-over-year, marking its first-ever interim profit while maintaining robust margins. Fueled by its proprietary Token ecosystem and strategic roadmap, the Company has not only unlocked explosive growth but also delivered definitive proof of a viable, high-potential AI commercialization model.
 

Profitability inflection point: core metrics outpace peers

On August 21, Xunce Technology released its 2026 interim results. First half revenue hit RMB 967million, rocketing 389% year over year to an all time high for the period. Profit attributable to shareholders stood at RMB 72.51million, reversing a loss of roughly RMB 89.40million in the prior year comparable period. Adjusted net profit reached RMB 67.00million, versus an adjusted net loss of approximately RMB 105million one year earlier. Outstripping peers on all key financial metrics, the Company has definitively reached its profitability inflection point, signaling the Company’s decisive pivot from capital intensive investment toward sustained profit generation.

During H1 2026, as AI application priorities shifted from model training to inference‑driven real‑world deployment, enterprise demand for Token surged. Profit quality of the Company improved accordingly. Total gross profit reached roughly RMB 580 million, up 340.5% year‑over‑year, with an overall gross margin holding firm at 60.1%. This underscores the resilience of its platform‑based delivery model amid business scaling-up.

The sharp performance growth stems from four converging drivers: accelerated deployment of enterprise AI real‑time data infrastructure, expanding footprint across multiple verticals, rollout of the Token‑based business model, plus international business expansion and ecosystem development.

Three powerful forces underpin this sharp profit reversal. First, faster rollout of end‑to‑end TokenOS and TokenCloud solutions has lifted the revenue share of high‑margin Token offerings, boosting overall profitability. Second, scale benefits from its platform‑based, modular product architecture have kicked in. More than 400 standardized functional modules have been built, raising solution replication efficiency and streamlining R&D, sales and G&A expense structures. Third, improved capital management efficiency generated incremental gains that bolstered profits in the period.

Operational efficiency also kept climbing. In the first half, average revenue per user (ARPU) jumped from RMB 1.64 million to RMB 5.56 million, a 240% year‑over‑year increase. Revenue per employee rose 379%, placing its workforce productivity among top-tier AI tech companies. The three major expense categories continued to improve with materially lower expense ratios. Customer retention has consistently stayed above 90%, demonstrating strong lock‑in and high stickiness based on deep integration within customers’ workflows.
 

One‑stop AI infrastructure via full‑stack tech; Token revenue surpasses 10%

As a leading Chinese provider of AI‑powered real‑time data infrastructure and analytics, Xunce Technology has built closed‑loop capabilities spanning nine end‑to‑end data processing stages and five data Tokenization workflows. With over 400 standardized and reusable functional modules, it boasts three core technical strengths: millisecond‑level real‑time responsiveness, 100% accurate data processing, and data Tokenization capabilities.

Sustained business growth is underpinned by robust product offerings. Its solutions are deeply deployed across customers’ self‑managed cloud and on‑premises environments. The Company has assembled a full product portfolio covering the entire data lifecycle and built a full‑stack technical architecture for enterprise AI adoption.

In the AI era, enterprises are rapidly shifting from raw data processing to real‑time AI‑powered decision‑making. The quality of an enterprise’s proprietary data stands as a critical determinant of successful AI deployment. Enterprise AI adoption is still in its early stages. High demand customers mainly fall into two groups: sector incumbents with a well‑defined AI transformation strategy, and organizations with clear business expectations for AI‑driven outcomes.

Xunce has actively expanded its AI‑to‑B business. On the data infrastructure front, it has built a data development system anchored on the AIDP platform. Acting as a “data hub”, it embeds AI deep into customers’ core workflows, delivering intelligent data insights and forecasts to support high quality business decision making.

On the resource orchestration front, Xunce launched TokenOS, the world’s first operating system focused on data capabilities. It refines enterprise proprietary data into high-density scenario‑specific Tokens and connects the full pipeline from data ingestion to model invocation. Building on this foundation, the TokenCloud platform centers on model capabilities and “data + model” competencies. Together, they deliver integrated end‑to‑end products and services spanning “computing power‑data‑Token‑model‑application”, equipping customers with one‑stop AI infrastructure. Customers can benefit from lower deployment costs and improved matching efficiency across computing power, data and AI models.
 

End‑to‑end product matrix: “computing power – data – Token – model – application”

Additionally, leveraging ten‑year expertise in real‑time data infrastructure, Xunce Technology centers data and Tokens as core hubs to connect upstream and downstream stakeholders, building hard‑to‑replicate differentiated strengths. These capabilities reach end‑user customers via its long‑standing FDE model: on‑site engineers work side‑by‑side with customer teams. Through on‑site collaboration and rapid iteration, large models and other AI technologies are woven into core business operations. Products and use cases then evolve with customers, enabling fast adoption of the Company’s full product suite.

Today, Xunce’s Token‑based business model has been deployed across high‑value use cases in biotech & healthcare, energy dispatch, industrial quality inspection and many more. Token‑related revenue exceeded 10% of total H1 revenue. Annual recurring revenue (ARR) for Token services surged, posting a 410% quarter‑over‑quarter jump in June and emerging as a new growth engine.

Unlike general-purpose Tokens built for everyday use cases, scenario‑specific Tokens target specialized vertical workflows. They draw on industry‑specific or enterprise proprietary datasets for model training and inference – exactly the category of scenario‑focused Tokens developed by Xunce.

Industry observers note the fundamental divide between general-purpose Tokens and scenario‑specific Tokens lies in specialized workflow requirements. For instance, energy‑specialized models prioritize precise regional consumption forecasting to drive energy savings or trading arbitrage. For AI‑collaboration models deployed on high-end manufacturing lines, customers care whether Token‑powered AI inference can tangibly boost yields, increase production capacity and capture critical market opportunities.
 

Clear four‑phase roadmap unlocks exponential growth potential

Looking ahead, Xunce has laid out a clear four‑phase strategic roadmap for its AI data infrastructure business. The Tokenization phase delivered milestone progress in H1. Building on this momentum, the Company intends to roll out a global Token exchange platform in the second half of the year to break down barriers for Token circulation across enterprises, industries and use cases. This would potentially elevate its business model from “pay‑per‑Token‑usage” toward “platform and ecosystem‑driven monetization”, expanding growth beyond individual customer Token consumption to ecosystem‑wide transactions. Over the longer term, sector-specific datasets will be further refined to train continuously evolvable small models for enterprise deployment, offering flexible commercial models including on‑demand subscription, perpetual license, ongoing fine‑tuning and performance‑based pricing.

Xunce stated it will implement strategic initiatives across five priorities: advance business model evolution, accelerate cross‑industry replication, pioneer cutting‑edge applications, build strategic partner ecosystems, and steadily expand its businesses globally. Overseas revenue soared 264% year‑over‑year, unlocking new international market upside.

Alongside its explosive business growth, Xunce Technology closed an equity placement and convertible bond financing worth roughly HKD 2.35 billion in early July, materially bolstering its cash reserves. The financing represents strong capital market validation for its business model and supplies ample capital for future computing power build‑outs, deeper sector vertical penetration and global expansion. In the near term, it will smooth cashflow volatility amid investment cycles. Over the medium‑to‑long run, it will accelerate the delivery of the Token‑exchange ecosystem and enterprise small model offerings, lending greater certainty to revenue mix optimization and profit margin expansion.

As a rare profitable AI‑to‑B benchmark among Hong‑Kong‑listed names, Xunce Technology is cruising onto a high‑speed value growth track powered by twin engines: proven earnings delivery and strengthened capital support.