Bairong (06608)'s AI “Long Moon Contract”: Digging Deeper into H1 Yingjiao's Teaser

If an investor sees a short-term downturn in a listed company, he is excited rather than anxious, proving that he is investing rather than speculating because he is waiting to get chips at a cheaper price. Different investors read the same announcement and see different things, because others in our eyes are who we are.

The difference in measuring the value of the capital market is the ruler. “Quarter” is engraved on the ruler for short-term customers, and the “cycle” is engraved on the ruler used by the old rivers and lakes. Old rivers and lakes read announcements. Favorite thing to do - first take a look at the current month, then take a look at the “far moon.” why? Because in the Yuanyue column, the story that will be played in the next game of cards is written.

^{On the same card, unraveling the H1 performance forecast of Bairong Intelligence (06608), in addition to the H1 net loss of [3.2-3.6] billion, “the revenue of the business (“AICC business” [1]), which is not affected by the above regulations and based on the AICC (Intelligent Communications) technology exhibition industry (“AICC business” [1]), showed significant growth in the new business. Among them, the new business serving customers in new scenarios showed significant growth, demonstrating the Group's ability to implement AICC technology in new scenarios.”}

This line is written on Bairong's “Yuen Moon” on the AI circuit. Which chapter of the cycle did Bairong Smart write in this H1 performance preview? It is clearly written in the announcement - what Bairong H1 focuses on is the introduction of new AICC scenarios and new customers.

Why is Bairong focusing on this? Because the AI battlefield window is coming to an end.

Let's take a look at what happened on the AI battlefield from 2024 to 2026 —

First, the general AI big model circuit has already entered the “commercialization” stage. The price war has begun, and gross margins are falling. The GM Big Model business is changing from “high technology” to “infrastructure” — just like cloud computing, servers, and new energy batteries back then. Once it becomes infrastructure, gross margins will converge, and excess profits will not be earned.

Second, the window period for vertical AI applications is still 2-3 years. At the end of the day, AI cannot just be a tool in a dialog box; in the end, it must also be a “silicon-based employee” that can help solve problems on the B-side. So the generic big model will be commercialized, but the “vertical model + industrial deployment + customer success” won't — this is the last window for differentiation. In this window, whoever first establishes “industrial delivery” capabilities will win future customer relationships. If you miss this window, you'll only be a “follower” if you go back in.

At the same time, capital began to be concentrated in companies “capable of commercialization.” First-tier capital has seen clearly — whoever can make “industrial delivery” within 2-3 years is the next betting target.

From 2025 to 2026, B-side customers will gradually use actual purchase orders to vote — AI that can be used is left behind and cannot be replaced. Customers don't wait 5 years to see if a company can successfully commercialize; they only spend 12 months judging whether a company can solve its problems now.

Competition on the AI battlefield is a competition with a three-tier structure of “base+medium+application” at the same time. If you don't have one floor, you'll be left behind; if you just make one floor, the ceiling will be low. Bairong just has these three layers:

Agentic AI infrastructure: base layer - “hydroelectric coal” in the AI era, the base for all upper layer applications;

Self-developed model matrix: middle stage - a key step in switching from “generic ammunition” to “special ammunition”;

Enterprise-level AI Voice + AICC application scenarios: application layer+trial order layer—AI products sold directly to customers, and RaaS (Results as a Service) results-oriented business model.

Short-term customers may choose to skip cars when they see a net loss — they think this is the end. Insiders, especially AI practitioners, saw that H1 invested in a base+medium+application — they knew this was a starter.

And if you look at Bairong's AI company H1, look at three things —

First, look at the direction of investment. AI talents, AI models, AI foundations — are building a solid foundation.

Second, look at your feet. The position is stable. The announcement revealed that after the reporting period, the Group's customer expansion across scenarios continued to make progress. Recently, it has achieved substantial implementation as a benchmark customer in new scenarios such as logistics, so early investment and accumulation will not go to waste.

Third, check the account. The accounts are sufficient. The announcement revealed that the Group's current cash reserves are relatively stable, and there are no interest-bearing liabilities. Bairong's current cash reserves are sufficient to support ongoing operations and business development needs.

Looking back at Bairong's 2026 H1 performance forecast in 3-5 years, what will we see?

The base layer, Agentic AI infrastructure becomes the “water, electricity, and coal” of the AI era — the base for all upper level applications, all enterprise-level agents, all silicon-based employees, and all vertical models; the middle platform layer, self-developed model matrices become “dedicated ammunition depots” — vertical field SOTA models; application layer, enterprise-level AI VoiceGPT + AICC becomes an “enterprise-grade silicon-based labor platform” — focusing on carbon-based customer service scenarios that can be transformed into AI in 3-5 years.

Taken together, it corresponds to the trillion-level AICC track chosen by Bairong - intelligent communication and the enterprise-grade silicon-based labor market.

[1] The Group will reclassify according to business attributes in the 2026 interim results report. The “AICC business” described above mainly refers to the use of self-developed AI VoiceGPT assistance: (1) intelligent marketing and intelligent customer service for customers in new scenarios such as brokerage firms, banks, operators, logistics, etc.; (2) some financial institutions carry out intelligent marketing and intelligent operation of credit products (products of this part of the business service are not affected by recent regulatory policies).