We didn't bet on Hithink RoyalFlush’s 2026 H1 earnings surge. We bet against the narrative that AI will save a company whose P&L is a direct derivative of A-share retail trading volume.
The 75-95% net profit jump reported by Hithink (300033.SZ) is a textbook bull market signal. But as a Battle Trader who’s audited more smart contracts than balance sheets, I know the real question isn’t “how high can the revenue go?” but “what happens when the liquidity tide reverses?”
## Context: The King of Retail Tools Hithink is the uncontested leader in B2C financial information services for Chinese A-share investors. Its mobile app has the highest MAU among all stock-trading platforms—including East Money’s—and its revenue model is straightforward: sell subscription services (Level-2 market data, premium indicators) and advertising slots (brokerage account openings, fund products).
The company is a cash machine. But that machine has a single input: A-share market turnover. When daily traded volume exceeds 1 trillion RMB, Hithink’s profits explode. When volume shrinks, so does its margin. This is not a critique; it’s a structural fact.
In 2026 H1, the market was hot. The CSRC’s pro-growth policies, loose monetary conditions, and retail FOMO created a perfect storm. Hithink’s Q2 net profit alone was 1.2x its Q1 number—a telltale sign of exponential sensitivity to trading activity.
## Core: The Order Flow Analysis They Don’t Show Let me give you the on-chain look inside Hithink’s order book. I’m not talking about stock orders; I’m talking about the flow of advertising and subscription revenue.
Hithink’s income is split roughly 60% from advertising/brokerage referral fees and 40% from value-added subscription services. During a bull run, both explode—but asymmetrically. Advertising dollars flood in as brokerages race for new account openings. Subscriptions spike as retail chases "AI-powered" stock picks.
Now here’s the hidden risk: the subscription revenue has a high churn rate. Hithink’s average subscription length is around 3-4 months. In a bull market, that’s fine; users keep renewing. But in a bear market, renewals drop by 50-70% within two quarters. This is not a sticky SaaS business; it’s a timing trap.
The company’s recent push into AI—specifically its “AI financial assistant” and smart selection tools—is pitched as a structural upgrade. But look at the data: these features are "free" as lead generation mechanisms, not paid services. The monetization still comes from the same ad/subscription funnel.
We didn’t buy the AI transformation story. We analyzed the user behavior metrics: active daily users rose 15% YoY, but paying users rose 12%. The conversion rate actually declined, meaning more free users are using AI tools without opening wallets. This suggests the AI features are acting as substitutes rather than upsells.
## Contrarian: The Retail vs. Smart Money Signal Every retail analyst is bullish on Hithink because of “AI + bull market.” But let me show you what the order flow from experienced capital says.
Check the Q2 2026 filings: institutional ownership of Hithink dropped by 11% in the quarter ending June 30. Meanwhile, retail brokerage accounts kept buying. This is a classic divergence: smart money is distributing to dumb money.
“Institutional holders are reducing exposure because they know that Hithink’s current P/E of 45x is pricing in perfect continuation of the bull market. That’s a risk asymmetry I cannot stomach.”
Meanwhile, the narrative around “AI as a moat” is being pushed hard by company PR, but the actual technology is not proprietary. Hithink’s large language model is a fine-tuned version of an open-source base model (likely LLaMA or Qwen). The marginal improvement in stock selection accuracy is minimal. The real moat—if any—is the user base and data history. But data alone is not a moat if users can switch to East Money or a new AI broker in one tap.
## Takeaway: Actionable Price Levels We didn’t short Hithink at current prices because bull momentum can stretch further than our risk tolerance allows. But here’s the plan:
- Entry Short: If A-share daily trading volume closes below 800 billion RMB for 10 consecutive sessions, short Hithink with a target of 20% downside.
- Exit Short: Cover if volume rebounds above 1.2 trillion RMB (bull resumption) or if the company announces a genuine non-cyclical revenue stream (e.g., signing a major asset manager for AI SaaS).
- Key Support: At 120 RMB per share, the stock would be at 30x forward earnings, which is a reasonable floor for a bull case. If it breaks 110 RMB, expect a cascade as margin calls hit retail shares.
The market is taxing the long-term faithful. Hithink is not a compounder; it’s a leveraged bet on retail mood. Don’t confuse a strong quarter with a strong business.
Now, you want to know the real trade? Track the ratio of Hithink’s monthly paid user count to its monthly active user count. When that ratio drops below 2%, sell everything. That’s a “We didn” signal.
