From "doing AI" to "investing in AI": BAIRONG AI-W (06608) participates in establishing a RMB 500 million industrial fund, as AICC's growth logic extends along the industry chain.

On October 15, BAIRONG AI-W (06608) disclosed a discloseable transaction announcement. Through its consolidated affiliated entity Beijing Shuke Youxiang, the company entered into a partnership agreement with CICC Capital and two limited partners to jointly initiate and establish an industrial fund with an initial scale of RMB 500 million. Among them, Shuke Youxiang has subscribed capital contributions of RMB 195 million, accounting for 39%; CICC Capital serves as the general partner, executive partner, and fund manager. The fund has an initial term of 7 years, with a 4-year investment period, and its investment direction covers fields including AI, underlying computing power, advanced manufacturing, and intelligent hardware. The strategic coordinates of this investment are clearly stated in the announcement: centered on AI and anchored on AICC, leveraging the manager's professional capabilities and partner network to expand opportunities to engage with AI startups and seek potential business synergies and long-term returns. In light of Bairong's current business progress, this arrangement is better understood as an extension of its core AI business: AICC is validating commercial value through real customers, real tasks, and continuous capacity expansion, while the industrial fund adds a capitalization pathway for the company to connect with related technologies, products, and application scenarios. Extending from "doing AI" to "doing AI + investing in AI," Bairong is exploring a growth model supported jointly by AICC's organic growth and AI industry chain investment. I. The Extensional Growth of the AI Strategy Judging the nature of an industrial investment has never been about how much money is put in, but about how deeply it interlocks with the core business. The strategic nature of this transaction is hidden in three core details. First, the investment direction is fully anchored to the core business's industry chain. The fund's investment direction disclosed in the announcement is leading and high-quality growth-stage enterprises in industries related to the digital economy and new quality productive forces, covering AI, underlying computing power, advanced manufacturing, and intelligent hardware. This fully overlaps with the company's strategic positioning of "centered on AI and anchored on the AICC business." It is not blind diversification, but proactive extension along the industry chain toward upstream and downstream. Second, the contribution ratio binds industrial discourse power. A 39% contribution share is by no means a purely financial allocation. Corresponding to the governance structure, Shuke Youxiang has the right to appoint at least one member of the investment decision committee. The industrial party's judgments and demands are written into the fund terms, rather than remaining at the level of verbal cooperation. More importantly, the investment boundaries are clear. The fund is explicitly prohibited from investing in listed companies (including private placements, strategic placements, and M&A restructuring equity, except for shares held after listing that have not been transferred), and from participating in convertible instruments, deposit arrangements, or quasi-equity investments; in principle, it invests only in non-listed enterprises. This means all funds are anchored to early-stage equity and hard-tech assets, with no room for flows into the secondary market or credit-like assets. In other words, this fund is essentially Bairong Intelligent's "industrial tentacle"using capital to connect upstream technologies and assets and feed back competitiveness in downstream application scenarios. It is not a second business outside the core business, but the extensional growth of the AI strategy. II. The Growth Engine Has Completed Production Validation The confidence for the dual-wheel drive has never come from investment, but from the core business itself standing firm. The 2026 interim results are a typical answer sheet for Bairong during its transformation periodthe new growth curve has entered the volume ramp-up stage. The anchor of the ramp-up lies in the AICC business. In the first half of the year, the Group's AICC business revenue was RMB 166 million, up 51.7% year on year. At the same time, R&D investment is also tilting in tandem. First-half R&D expenses increased 30.7% year on year to RMB 394 million, with resources clearly tilted toward the pan-industry AI transformation. Revenue expansion and increased investment are proceeding in parallel, and the new growth curve is entering a stage of sustained expansion. But more significant than the growth rate is that the methodology for cross-industry replication has completed production-level validationthe logistics scenario was the first to run through the entire process. In July, benchmark customer Silicon-based Customer Service's daily call handling volume climbed from fewer than 1,000 calls in the initial launch period to more than 15,000 calls, with the longest call duration exceeding 20 minutes. The real significance of this case lies in AI beginning to enter complex, open, and continuously occurring real service processes: in logistics inbound calls, user expressions, issue content, and emotional states vary greatly, and dialogue paths are difficult to fully predefine. Being able to stably handle tasks in such an environment is a substantive production-side test of the Silicon-based Customer Service Agent. The brokerage scenario rapidly validates the customer expansion logic. As of August, the number of brokerage customers signed and in the contract process reached 13; the deployed seats of Silicon-based Customer Service at a leading brokerage expanded from about 30 seats at the beginning of the year to about 210 seats. For an enterprise-level AI company, existing customers proactively expanding deployment after use is the hardest signal of product valueinitial access is only POC validation, while continuous expansion represents customers' willingness to hand more core business processes to AI. What better demonstrates growth potential is the new-scenario data: during the period, AICC new-scenario revenue was RMB 11.26 million, up 195% year on year, and has covered six major industries including logistics, aviation, brokerages, banking, telecom operators, and auto insurance. "Can AICC capabilities be replicated across industries?" This question has already been validated by two benchmark samples, logistics and brokerages. This is precisely the realistic foundation for industry chain investment to land at this moment: only when downstream scenarios have run through real production environments can upstream investment in technology, computing power, and hardware have a real landing point, and can synergies be more than empty talk. III. Why Did Bairong Run Through AICC? There Are Three Layers of Moats Here Enterprise-level AI customer service is widely recognized as one of the three high-certainty AI implementation scenarios at present, alongside programming and enterprise search. Its core features are documented work content, manually judgeable results, and quick returns on investment. There are many players in the track, but Bairong's ability to produce benchmark cases relies on three layers of hard-to-replicate barriers. The first layer is a self-developed vertical large model system. The company has formed a complete matrix of BR-Voice multimodal speech foundation model, BR-LLM-Proactive intent recognition and proactive guidance foundation model, and BR-Vision-Doc document understanding foundation model. Among them, BR-Voice achieved 94.84% test accuracy in the "Multilingual Conversational Speech Understanding" task of the INTERSPEECH 2026 MLC SLM Challenge in July 2026, ranking third globally and first among Chinese industry players; in the four tasks of the IEEE SLT 2026 related challenge, it entered the global top five in all, with two spoken language understanding tasks ranking second globally. Unlike the general large model route, Bairong's model parameter scale is relatively restrained, directly bringing lower customer invocation costs and higher commercial implementation efficiency. The second layer is telephony engineering and delivery capability. AI customer service has never been a contest of single-point model capability, but a systems engineering effort involving concurrency stability, interruption handling, emotion recognition, and ticket closure loops. Traditional telephone lines use 8kHz narrowband voice, retaining far less acoustic information than internet broadband audio, which places extremely high demands on recognition, noise reduction, interruption handling, and turn-taking judgment; production environments also need to simultaneously satisfy low latency, stable concurrency, reliable transfer to human agents, and complete quality inspection. Bairong has built a complete engineering system around telecom lines, softswitches, real-time audio transmission, and business interfaces, and through forward-deployed engineers going deep into customer sites, it translates model capabilities into runnable business processes. The highly reliable delivery system honed through years of serving financial institutions is an entry threshold that latecomers find difficult to replicate in the short term. The third layer is compliant data assets. The company has continuously served financial industry scenarios for more than a decade and accumulated compliant private-domain data; at the same time, the financial industry itself has very high requirements for data accuracy and compliance, and such data has been accumulated under stringent standards. Combined with the industry know-how of top customers and deep integration with business systems, this together forms extremely high switching costs. In terms of business model, the company adopts the RaaS (results-as-a-service) model: charging by successfully completed call tickets that require no transfer to human agents, charging by the number of deployed silicon-based employees, and sharing profits based on the scale of facilitated business. From the scenarios already implemented, the charge per successful call is significantly lower than the customer's corresponding labor costcustomers can directly calculate the input-output account, which is precisely the core reason why new scenarios can complete the journey from POC to volume ramp-up within half a year. After visiting a number of Chinese local AI application companies, analysts at Guoyuan International judged in an industry research report titled "AI Theme Research SummaryThe Era of Application Profit-Taking Is Coming" released on September 25 that enterprise-level AI applications have moved from "proving AI can do it" to "proving AI can do it quickly and cheaply," and that delivery cycles, labor replacement rates, and the proportion of mature products will determine who can capture the next round of customer budgets. Bairong is standing at the key switching point from demonstration to delivery. IV. Industry Chain Investment: Growing Along the Boundaries of the Core Business, Technology and Capital Synergy If AICC is the downstream application layer close to customers and data, then each layer upward in the industry chain corresponds to real bottlenecks and also to the fund's investment targets. For example, in the AI models and applications segment, this layer relates to the upper limit of multi-turn dialogue and intent understanding capabilities. If the fund invests here, what is fed back is the iteration of vertical models and the supplementation of agent capabilities; another example is underlying computing power and infrastructure, which determine inference costs and concurrency elasticity and affect implementation cost-effectiveness and delivery gross margins. The corresponding investment feedback lands on lower unit computing power costs and improved delivery gross margins. The core value of this structure lies in: investment is not a second business running parallel to the core business, but the natural growth of the core business's capability map. Investee companies gain real business scenarios and validation channels, while Bairong gains technology supply and scenario entry points, forming a positive cycle of two-way empowerment. It is worth noting that the fund's investment direction is not chosen out of thin air, but is highly synergistic with the company's internal technology R&D roadmap. Bairong's interim results announcement has explicitly mentioned R&D of full-duplex speech models, as well as technical validation oriented toward embodied intelligence, wearable devices, and other directions. This means that investment directions such as intelligent hardware and advanced manufacturing have natural connection points with the cutting-edge technical capabilities the company is exploringif technical complementarity can be formed in suitable projects in the future, the company's voice interaction and agent capabilities will have the opportunity to achieve application implementation through more terminals and scenarios. The ideal synergy path is clear: the application side proposes real business needs, industry chain partners provide corresponding technologies and products, and both parties then validate commercial value through customer sites. For example, customized computing power solutions adapted to business needs may help improve inference efficiency and reduce unit costs; mature terminal and system partners may significantly shorten the implementation cycle for entering new scenarios. The fund provides an organized channel for finding these partners, and the actual effects will gradually emerge with project screening, investment, and cooperation. The fund's manager is CICC Capital, whose quality is self-evident. As of December 31, 2025, CICC Capital's cumulative subscribed management scale exceeded RMB 680 billion, making it a leading comprehensive private equity investment platform in China. It also serves as the manager of the Beijing-Tianjin-Hebei Venture Capital Guidance Fund with a target scale of RMB 50 billion, with investment directions explicitly covering frontier fields such as general artificial intelligence and embodied intelligence. The backgrounds of the other two limited partners are equally key: one is a government investment fund indirectly wholly owned by the Chaoyang District State-owned Assets Supervision and Administration Commission, and the other is a government guidance fund ultimately held with more than 30% equity by the Beijing Municipal State-owned Assets Supervision and Administration Commission system. This brings two core values to the fund. First, professional guaranteethe GP's mature due diligence, valuation, and post-investment management systems are capabilities that a single industrial party finds difficult to build on its own; second, project sources and policy synergyBeijing has established a RMB 10 billion municipal-level AI industry investment fund, and Chaoyang District is also advancing the "AI+" initiative, explicitly proposing to explore CVC cooperation with chain-leading enterprises to establish vertical segmented industry sub-funds. The seven-year initial term and four-year investment period are not arbitrarily set time windows, but deeply match the growth patterns of the AI industry. Correspondingly, the contribution pace has also been designed to be smooth. In principle, capital calls are made in 5 installments, corresponding to the 4-year investment period. Funds are invested gradually rather than as a one-time expenditure, greatly smoothing the impact on current cash flow. V. Value Outlook: Short-Term Strategic Signal, Long-Term Ecosystem Synergy Returning to the initial question: why launch this fund now? The answer lies in the resonance of three time windows. First, the core business has reached the node for outward extension. AICC has completed production-level validation of cross-industry replication, and the two samples of logistics and brokerages prove that the methodology is transferable; at this time, connecting outward to computing power, hardware, and model capabilities gives synergy a real landing point. Second, the industry window is in sync. Domestic enterprise AI budgets are moving from pilot demonstration to large-scale implementation. Early-stage asset valuations in the primary market, the RMB 50 billion Beijing-Tianjin-Hebei guidance fund, the Beijing municipal AI industry fund, and the Chaoyang District CVC mechanismpolicy and market dividends are being released concentratedly in the same time window. Using RMB 195 million of limited capital to access this system buys not only financial returns, but also project sources and an industrial ecosystem position. Third, the contribution pace matches the industry cycle. The fund contributes in 5 installments with a 7-year term, and funds are invested gradually according to investment progress, matching the relatively long growth cycle of the AI industry and leaving room for subsequent project screening and synergy implementation. From an industry perspective, Guoyuan International judged in its AI application industry research that the entire track is moving from "proving AI can do it" to "proving AI can do it quickly and cheaply," and that 2027-2028 will be the validation window for large-scale profitability of leading enterprises. In this window, a single self-developed capability is no longer enough, and building a dual-wheel ecosystem of "self-development + investment" is becoming the common choice of leading players. Bairong Intelligent's RMB 500 million industrial fund is essentially using the core business's scenario barriers to exchange for upstream technology supply, and using precise industrial judgment to exchange for early-stage growth dividendsthe former consolidates current delivery competitiveness, while the latter opens the long-term growth ceiling.