Hithink RoyalFlush: The Tollbooth of China's Retail Market & The AI Financial Portal

I. Introduction & Episode Roadmap

Somewhere in China right now, at a kitchen table in Chengdu or on a subway platform in Shenzhen, a user unlocks a smartphone and taps a red-and-white icon. Within a second — a response time critical to the platform's user engagement — a candlestick chart resolves on screen, a scrolling ribbon of northbound capital flows appears, and a search bar prompts for a query in plain Mandarin: which ChiNext stocks have return on equity above twenty percent and net institutional inflows this week?

That user is one of roughly 35 million people who opened the same app in the same month.¹ They are a retail investor (散户) — one node in the largest population of individual stock traders on Earth. By the end of 2025, China's registered investor count passed 250 million, with nearly 13.9 million new accounts opened during the year alone.² Those accounts traded a combined ¥420.21 trillion of A-shares in 2025, setting a record with an average daily turnover of about ¥1.73 trillion.³

Almost none of those users pay 浙江核新同花顺网络信息股份有限公司 Hithink RoyalFlush Information Network Co., Ltd. a single yuan. That is the core of the business model: the company built the most-used financial app in China largely by offering it free of charge, then monetizing the institutional access to the audience assembled inside it.

The 2025 financial figures demonstrate how the model performs during favorable market conditions. Revenue reached ¥6.029 billion, up 44.0% year over year. Net profit attributable to shareholders reached ¥3.205 billion, up 75.79%, yielding earnings of ¥5.96 per share.⁴ Gross margin ran at roughly 91.5% and net margin stood above 53% — metrics characteristic of a software licensing business or search platform rather than a conventional financial firm.⁵ The balance sheet reflects a similar structure: of ¥15.83 billion in total assets at year-end, approximately ¥14.04 billion sat in cash and equivalents — roughly 88.65% of the total asset base — against essentially no interest-bearing debt.⁴⁶ Operating cash flow of ¥3.774 billion actually exceeded reported net income.

A key strategic question defines the company's market positioning. In 2015, China's primary peer comparison — 东方财富 East Money (300059.SZ), which originated as a financial news portal — concluded that aggregating an audience was insufficient. It acquired a securities brokerage, embedded its own trading license inside its app, and began directly collecting commissions and margin interest. In 2025, that approach generated ¥16.068 billion in revenue and ¥12.085 billion in net profit, with the securities segment alone contributing ¥12.535 billion.⁷ East Money's net earnings are approximately four times larger.

同花顺 RoyalFlush took the opposite approach. It never acquired a brokerage, remaining by design a neutral platform. Over a hundred securities firms place their account-opening funnels within the app precisely because RoyalFlush does not compete with them for client assets. This neutral positioning avoids the capital demands of maintaining a brokerage balance sheet, though it caps direct financial capture relative to integrated peers.

Evaluating which of these models proves more resilient — and under what market conditions each faces pressure — forms a central theme of this analysis.

Four main threads run through this story. The first is the two-sided platform: a retail trading audience on one side, and brokerages and fund managers bidding for their attention on the other, with RoyalFlush collecting a toll in between. The second is capital allocation: a founder-controlled company with no debt, no major acquisitions, and no goodwill on its balance sheet choosing to return cash to shareholders rather than purchase growth. The third is the technology trajectory — evolving from dial-up desktop terminals in the 1990s to a dominant mobile app, and now to a financial large language model positioned as a competitive moat. The fourth thread is regulatory compliance. Operating an information platform that delivers market analytics to retail investors requires adhering to licensing boundaries established by the 中国证券监督管理委员会 CSRC China Securities Regulatory Commission. Regulatory oversight has previously resulted in fines in 2016 for software facilitating off-exchange leverage,⁸ as well as a three-month suspension on taking new customers at a key subsidiary in late 2024.⁹

Viewing Hithink as an information search engine and attention aggregator clarifies its business model. It aggregates retail investor attention, sells access to financial institutions, and provides speed and convenience to end users rather than direct financial advice or balance-sheet capital. This structure explains both its 90%-plus gross margins and its recurring regulatory oversight, as distributing financial tools to retail investors carries strict compliance requirements.

Understanding why a company with a 91% gross margin constantly navigates regulatory oversight requires examining its origins: a university graduate armed with ¥500,000 and a futures terminal.

---

↑ Back to Top

II. Founder Roots & The PC Software Era (1994–2008)

In 1994, China's Shanghai and Shenzhen stock exchanges were barely three years old. Investing for most Chinese households meant physically walking into a local brokerage branch (营业部) — a room filled with cigarette smoke, folding chairs, and a wall-mounted electronic display board — and shouting orders at a clerk behind glass. Price discovery was a public, in-person spectator sport.

Into this environment stepped 易峥 Yi Zheng, born in 1970, who graduated from the Department of Electrical Engineering at 浙江大学 Zhejiang University in 1993 after writing market analysis software for local securities firms during his studies.¹⁰ In 1994, Yi registered 杭州核新 Hangzhou Hexin with ¥500,000 in initial capital alongside co-founder 叶琼玖 Ye Qiongjiu.¹⁰¹¹

The company's first commercial product revealed its core engineering orientation. Known as 龙虎榜 Long Hu Bang ("Dragon & Tiger List") — named after exchange disclosures detailing the top buying and selling brokerage branches for volatile stocks — the software sold for ¥120,000 per unit by offering trading desks visibility into rival order flows.¹⁰ Rather than publishing commentary or investment newsletters, Hangzhou Hexin treated market analysis as a data-processing and pipeline problem.

That technical perspective separated Hithink from most Chinese financial media pioneers of the era. While 东方财富 East Money was later founded by a market commentator around an online news portal and discussion forum, Hithink was built by engineers around real-time data architecture. Over the next three decades, both companies would converge on the same retail investor base from opposite directions, retaining distinct strategic priorities shaped by their founding DNA.

Product expansion required adapting to shifting market conditions. A ¥120,000 futures terminal served a narrow addressable market of institutional desks, and domestic futures trading faced tighter regulatory scrutiny mid-decade. Yi pivoted around 1997 as online trading infrastructure became technically viable in China. Hangzhou Hexin developed the 天网 Tianwang online securities analysis and trading system, winning a tender in 1998 to supply Jiangsu Province's securities trading infrastructure.¹⁰¹¹

This contract established Hithink's initial business model as a B2B technology vendor rather than a consumer brand. Late-1990s brokerages needed software capable of streaming quotes over 33.6k dial-up modems and routing client orders directly to exchanges, but few possessed the internal software capabilities to build it. Hithink delivered white-labeled terminals displaying brokerage branding while running on Hithink's engine. Brokerages paid for the software licenses, while retail clients executed trades through the interface.

That B2B foundation created two long-term strategic assets. First, it embedded Hithink directly into brokerage order-routing channels — establishing a sticky technical integration that later enabled the platform to host third-party trading interfaces. Second, it built a real-time data architecture equipped to ingest Level-2 tick feeds from the 上海证券交易所 Shanghai Stock Exchange and 深圳证券交易所 Shenzhen Stock Exchange and stream them to thousands of endpoints simultaneously. In an era dominated by dial-up internet, rendering charts half a second faster than competitors provided a tangible user advantage.

Level-2 data remains one of Hithink's most durable and profitable product lines. Standard quotes provide only basic price and top-level bid-ask data. In contrast, Level-2 data exposes the full order book depth, showing share volumes queued across multiple price levels tick by tick. For short-horizon retail traders who drive a significant share of A-share turnover, this real-time visibility is highly valued. Because exchanges hold the underlying data rights, Hithink's model centers on licensing wholesale feeds, wrapping them in optimized interface software, and distributing access to retail subscribers at a premium.

This early development phase also embedded a focus on execution speed. Because user perception hinged on whether charts rendered instantly or lagged over slow connections, Hithink prioritized low latency over feature breadth. That emphasis established a reputation for speed that helped retain users across technology cycles.

The company's corporate structure evolved alongside its market reach. 上海核新软件技术有限公司 Shanghai Hexin Software Technology was established in August 2001 to supply securities firms directly with quote and trading systems.¹¹ In December 2003, the company adopted 同花顺 (Tonghuashun, literally "straight flush," rendered in English as RoyalFlush) as its financial services brand while launching its consumer-facing web portal.¹¹ That site, operating at 10jqka.com.cn, remains the primary digital gateway for the business.³² The consumer brand name linked engineering-driven data tools with a poker metaphor — reflecting an underlying tension between analytical software delivery and retail speculative interest that regulators would later scrutinize.

During the 2001–2005 bear market, the Shanghai Composite Index declined for four consecutive years, depressing retail trading activity and weakening demand for paid desktop analysis software. Competitors such as 大智慧 Great Wisdom (later 601519.SH) and 指南针 Compass (later 300803.SZ) encountered identical pressures. Hithink maintained financial stability through its B2B revenue base: recurring software maintenance contracts with brokerages provided steady cash flows that offset retail market weakness and supported consumer development through the downturn.

This structure established a key operational dynamic for Hithink: maintaining a recurring institutional revenue floor beneath cyclical consumer market demand. That dual model persisted throughout the company's history, eventually shifting as expanding consumer traffic grew to enhance its institutional offerings.

By 2008, Hithink had established a recognized brand, a real-time data pipeline, deep brokerage technical integrations, and a high-traffic web portal. However, it lacked major growth capital and liquidity options for its founding team — a constraint that upcoming domestic capital market reforms were set to address.

↑ Back to Top

III. The ChiNext IPO & The Mobile Super-App Explosion (2009–2014)

On October 30, 2009, the Shenzhen Stock Exchange launched 创业板 ChiNext — China's NASDAQ-style growth board for enterprises unable to meet main-board profitability thresholds — debuting with 28 initial listings. Hithink joined the second wave, listing in December 2009 during ChiNext's first two months of operation.¹²¹³

The offering terms reflected the market enthusiasm of the launch. Hithink issued 16.8 million shares at ¥52.80 per share, raising ¥887.04 million at an issue price-to-earnings multiple of 92.57. On its December 25 debut, the stock opened at ¥74.¹²¹³ A valuation near 100 times earnings for a mid-sized software vendor underscored market demand for growth-oriented technology listings on the new venue, setting a premium valuation baseline that persisted across much of the company's public history.

The capital raised proved less consequential than the technological shift that followed. Within eighteen months of the IPO, smartphone adoption accelerated across urban China. The central premise of Hithink's consumer business — that investors analyzed market data on desktop Windows terminals during trading hours — rapidly grew outdated.

This shift presented a classic incumbent's dilemma. Hithink's paid desktop software relied on annual subscriptions for advanced charting, custom indicators, and real-time feeds. A free, responsive mobile application threatened to cannibalize that recurring revenue stream directly. For a newly public company evaluated on a high earnings multiple, the immediate incentive was to defend desktop software subscriptions and relegate mobile tools to a companion feature.

Yi Zheng chose a different path, directing core engineering resources toward making the mobile app the primary consumer product. Offered free at the point of use, the application was optimized for launch speed and chart rendering on basic mobile hardware and early 3G networks. Crucially, Hithink integrated third-party trading connections, allowing users to execute trades through their existing brokerage accounts directly within the interface.

That design defined Hithink's platform strategy. Rather than seeking to own the underlying brokerage transaction, Hithink focused on owning the primary interface where investment decisions were made. Once an investor's watchlist, indicators, and daily workflow resided inside one mobile app connected to their existing broker, the software became the default surface, turning individual brokerages into replaceable back-end execution channels.

Market dynamics soon accelerated user adoption. The 2014–2015 A-share bull market generated unprecedented retail trading activity and record turnover. Millions of first-time investors entered the market via smartphones without legacy desktop habits. Hithink's application expanded rapidly to become China's dominant securities app, a market-leading position it retained across the subsequent decade.¹

The monetization framework that continues to drive the business took shape during this expansion, built around three primary revenue channels:

Selling data privileges to retail users. While basic market quotes are provided free of charge, Level-2 order-book depth — showing order queues behind top-of-book prices — requires a paid subscription. Hithink licenses these wholesale feeds from exchanges and resells access, converting proprietary market data into high-margin recurring subscriptions.

Selling audience access to brokerages. As new investors seek to open trading accounts, Hithink monetizes its position at the point of decision by charging securities firms for placement and completed account conversions. This lead-generation model operates with economics resembling search advertising rather than traditional financial services.

Selling premium analytics and signals. Hithink distributes stock screeners, technical indicators, alert systems, and advisory-adjacent tools. This segment represents the platform's highest-margin and fastest-growing revenue line, though it also carries elevated regulatory risk.

From a strategic perspective, Hithink navigated a disruptive technology shift by accepting short-term software subscription cannibalization in exchange for building a broader consumer platform. Rather than relying on major capital expenditures, the transition was achieved primarily through the reallocation of software engineering talent.

However, the shift to mobile fundamentally altered Hithink's revenue characteristics. Desktop software generated relatively steady, product-driven subscription revenue. In contrast, the mobile platform tied revenue directly to retail trading volume. When market activity surges, user engagement, account openings, data subscriptions, and brokerage advertising demand expand simultaneously. Conversely, during market contractions, all four revenue drivers pull back in unison, introducing significant operational reflexivity and cyclicality to the business model.

The competitive landscape shifted as well. 大智慧 Great Wisdom, previously the leading desktop market data vendor, delayed its mobile transition and subsequently faced prolonged regulatory and accounting challenges. Meanwhile, 东方财富 East Money successfully transitioned into mobile by leveraging its online investor community. Secondary desktop vendors lost market relevance as interface habits reset around mobile screens. Hithink recognized the platform shift early and adapted its user interface — a transition it now faces again as artificial intelligence reshapes financial software.

Yet the same retail market boom that drove Hithink's user expansion was fueled in part by financial tools distributed on its platform. In the summer of 2015, those leverage dynamics encountered severe regulatory scrutiny.

↑ Back to Top

IV. The 2015 Crash, HOMS / Shadow Leverage Crackdown, & Strategic Pivot

Understanding the 2015 market crash requires examining a piece of financial plumbing with a deceptively mundane name: the sub-account system.

Chinese regulators cap how much leverage a retail investor can take on through official margin financing and require brokerages to verify client identities. 场外配资 (off-exchange margin financing) emerged as a workaround. A financing firm would open a single legitimate brokerage account, then use specialized software to divide that account into thousands of virtual sub-accounts rented out to individual retail traders at leverage ratios of 4-to-1, 5-to-1, or up to 10-to-1. To the exchange, it appeared as a single account; to regulators, individual borrowers were invisible.

The software that scaled this practice was HOMS, developed by Hangzhou Hengsheng Network Technology, a unit of 恒生电子 Hundsun Technologies. Hithink developed a comparable system, as did Shanghai Mingchuang. Technically sold as asset-management and order-routing software, these systems functionally enabled unlicensed entities to operate shadow securities trading operations.

This structure operated much like subletting a single apartment into dozens of uninspected micro-rooms—lucrative until a fire broke out.

The market downturn arrived in June 2015. The Shanghai Composite Index fell roughly 40% in under three months. Cascading margin calls triggered forced liquidations of leveraged sub-accounts, accelerating the market decline. The 中国证券监督管理委员会 CSRC responded by targeting the underlying technical infrastructure. In August 2015, the regulator announced formal investigations into both Hundsun's network unit and Hithink for suspected violations of securities and futures regulations.¹⁴

The regulatory enforcement actions handed down in late 2016 were notably asymmetric. Hengsheng Network faced ¥132.85 million in confiscated illegal gains plus a ¥398.56 million fine, totaling over ¥531 million. Mingchuang was ordered to forfeit ¥15.99 million in gains alongside ¥47.96 million in fines. In contrast, Hithink's penalty comprised ¥2.177 million in confiscated gains and a ¥6.531 million fine—totaling under ¥8.71 million out of roughly ¥600 million in combined industry penalties. In addition, two Hithink employees received individual administrative sanctions: vice general manager 朱志峰 Zhu Zhifeng received a warning and a ¥100,000 fine, while a product manager received a warning and a ¥50,000 fine.⁸

This disparity provides key analytical context. Hithink accounted for approximately 1.5% of total industry fines and 1.4% of confiscated gains, indicating it was a minor participant rather than a primary provider in off-exchange leverage software. Despite operating during an eighteen-month period of record retail trading activity, the company generated minimal revenue from this segment.

Two interpretations explain this result. The first is that Hithink's primary focus remained its core user platform, leaving leverage modules as an unscaled auxiliary offering. The second is that Hithink entered the market late and was caught in the regulatory net before scaling operations. Because Hithink has not disclosed detailed accounts of this division, the enforcement outcome remains documented while the strategic intent remains unconfirmed.

The episode also highlights a corporate governance detail. Following the 2016 CSRC sanction, Zhu Zhifeng remained in senior management at the company, with no additional internal disciplinary actions publicly disclosed. While opinions vary on how long administrative warnings should affect an executive's standing, this history establishes context for evaluating how Hithink responds to regulatory findings—a question re-emerging in light of subsequent regulatory actions in 2024 and 2026.

Following the crackdown, Hithink shut down its margin-management software. Management opted not to re-enter the leverage market or seek alternative licensing pathways for credit-based products. Instead, the strategic focus shifted toward three core activities: aggregating and distributing user traffic, selling analytical tools directly to retail investors, and providing technical services to financial institutions. This framework avoided holding client funds or extending credit.

The subsequent market environment presented operational challenges. Depressed retail trading turnover from 2016 through 2018 directly impacted Hithink's top line. Without a brokerage balance sheet, the company lacked net interest income to offset lower trading activity.

However, the downturn prompted operational adjustments. The company disciplined its cost structure, emerging from 2018 with leaner operating expenses than in 2015. Additionally, the contraction reinforced Hithink's core business assets: its retail audience, real-time data infrastructure, and brokerage integration network. Management focused on developing additional paid services around these existing capabilities, laying the foundation for future offerings such as premium subscription tiers, account-opening referral services, institutional software terminals, fund distribution platforms, and AI applications.

The regulatory investigation also shaped Hithink's corporate strategy. Having faced CSRC scrutiny over shadow financing software, seeking its own securities brokerage license carried elevated regulatory hurdles. Consequently, maintaining a neutral platform strategy aligned with both operational priorities and regulatory requirements.

This period established the strategic positioning of the company, setting up a sharp contrast with competitors that chose an integrated brokerage model during the same era.

↑ Back to Top

V. The Great Strategic Split: RoyalFlush vs. East Money

In April 2015, as the A-share market neared its peak, 东方财富 East Money announced a plan to acquire 100% of Tibet Tongxin Securities. The deal valued the target at ¥4.405 billion through newly issued shares priced at ¥28.53 each, alongside a ¥4 billion follow-on placement to recapitalize the broker.¹⁵ The transfer closed in December 2015. By March 2016, the acquired firm was renamed East Money Securities and relaunched with commission rates of roughly 0.025% — a price cut sharp enough to draw public complaints from traditional competitors.¹⁶¹⁷

The strategic logic was straightforward: East Money already controlled the investor traffic, and stock brokerage was the highest-value transaction those users conducted. Rather than routing clients to third-party brokers, East Money captured the transaction directly.

The commercial result was immediate. In 2015, Tibet Tongxin had operated as a small regional firm, generating roughly ¥114 million in revenue and ¥57 million in net profit. Once fed by East Money's user traffic, its 2016 revenue climbed to ¥1.1 billion and net profit rose to ¥374 million.¹⁵ By 2025, East Money's securities arm processed ¥38.46 trillion in equity and fund trading volume — up 58.66% year over year — yielding ¥7.98 billion in net fee and commission income and ¥3.033 billion in net interest income.⁷

Hithink watched this transformation and chose not to follow, maintaining its stance without acquiring a securities license.

The Case for Switzerland

The case for remaining a neutral platform rests on a fundamental structural premise: an entity that does not compete with securities brokerages for client assets can serve all of them, whereas a platform operating its own broker eventually hosts only itself.

This posture addresses a direct conflict of interest. When a third-party brokerage buys placement inside East Money's app, it purchases audience reach from a company whose in-house broker sits one tab away, while sharing user conversion data with a direct competitor. That arrangement may be tolerable when the host's brokerage operation is small, but it grows uncomfortable once that broker becomes an industry leader. Top-tier institutions such as 中信证券 CITIC Securities, 华泰证券 Huatai Securities, and 国泰海通 Guotai Haitong face no such conflict on Hithink's platform, where no competing brokerage sits on the other side of the ledger.

The financial benefits of this positioning appear clearly in Hithink's revenue breakdown. In 2025, advertising and internet business promotion services generated ¥3.462 billion, or 57.43% of total revenue, representing 70.98% year-over-year growth at a gross margin of approximately 96.1%.⁶¹⁸ This segment functions as Hithink's primary tollbooth, where brokerages pay for marketing placement and account conversion during periods of heavy retail participation, such as 2025, when Chinese investors opened nearly 13.9 million new trading accounts.

Value-added telecom services — the regulatory category covering consumer subscriptions for Level-2 order-book data, screeners, and premium analytical tools — contributed ¥1.950 billion, or 32% of total revenue, growing 20.71% with a gross margin of roughly 85.5%.¹⁸ The growth spread between these two divisions illustrates the mechanics of the 2025 expansion: advertising expanded more than three times as fast as subscription sales. The overall business became more dependent on brokerages competing for new client accounts than on existing retail users deepening their software spend. Because advertising rose from 48.36% of revenue in 2024 to 57.43% in 2025, the business mix shifted toward its most volatile growth engine, meaning that operational leverage cuts both ways during market downturns.⁶

The remaining revenue streams reflect narrower focus areas. Software sales and maintenance, which includes the iFinD institutional terminal, generated roughly ¥400 million, or 7% of revenue, up 12.12%. Fund distribution and other transaction fees via the 爱基金 iJijin platform added about ¥220 million, or 4% of total revenue, up 14.24%.¹⁸

That fund distribution figure highlights a significant strategic choice. East Money's wealth management subsidiary, 天天基金 Tiantian Fund, generated ¥3.182 billion in 2025 — more than fourteen times Hithink's total fund distribution revenue.⁷ Despite holding a fund distribution license for years, Hithink has chosen not to build a comparable wealth management platform atop its retail user base.

The Case Against Switzerland

This divergence raises a critical question for investors: whether neutrality reflects a deliberate long-term strategy or simply an unwillingness to expand into capital-intensive financial services.

The operational arguments supporting neutrality remain clear. Hithink carries zero credit risk, no regulatory capital requirements for margin lending, no market-making inventory, and no goodwill on its balance sheet, while converting roughly 63% of revenue into operating cash flow. The company generates about one-fifth of East Money's net profit while operating with minimal capital deployment and virtually no balance-sheet leverage.

However, the drawbacks are equally distinct. Hithink monetizes an audience of comparable size far less intensively. Commission income and margin interest generated across East Money's app pass over screens that, for millions of users, look identical to Hithink's interface. By remaining neutral, Hithink forfeits both transaction revenue and the underlying trading data that could enrich proprietary AI advisory tools. Consequently, the company relies heavily on institutional marketing budgets, which are often the first expense items cut during retail trading slumps.

In short, Hithink's model yields higher profit margins per dollar of revenue but results in a smaller total earning base subject to sharper cyclical swings. Whether this approach proves superior depends on the trajectory of Chinese brokerage economics. If trading commissions continue their multi-decade decline, avoiding a brokerage balance sheet may prove prudent. Conversely, if regulatory policy favors consolidation and scale in margin lending, integrated brokers stand to capture the majority of industry economics.

An additional distinction lies in regulatory exposure. East Money's brokerage functions as a fully regulated financial institution subject to capital adequacy requirements, risk ratings, and comprehensive regulatory supervision. Hithink operates under a lighter supervisory framework as an information service provider, licensed fund distributor, and investment consulting subsidiary. During market expansions, this difference yields high capital efficiency; during contractions, it provides operational flexibility, allowing Hithink to adjust costs and exit product lines faster than a licensed broker.

This structural neutrality also explains Hithink's substantial cash balance. Generating ¥3.774 billion in annual operating cash flow without deploying capital into a licensed financial business leaves substantial liquidity accumulating on the balance sheet.

↑ Back to Top

VI. Capital Allocation & Management Deep-Dive

Where does all that accumulated liquidity go? Primarily back to shareholders. The scale of that capital distribution serves as management's clearest statement regarding internal growth opportunities.

For fiscal 2025, Hithink's board proposed a cash dividend of ¥51 per 10 shares, alongside a bonus issue of 4 shares per 10 from capital reserves. Across approximately 537.6 million shares outstanding, the cash payout totals roughly ¥2.742 billion — representing an 87% payout ratio relative to net income.¹⁹¹⁸ By comparison, Hithink raised ¥887 million during its 2009 initial public offering. In a single fiscal year, the company distributed more than three times its lifetime public equity capital raised.

An 87% dividend payout carries two distinct strategic interpretations. From one perspective, returning the vast majority of net income indicates disciplined capital management, reflecting a candid assessment that the business lacks ¥3 billion in internal projects capable of matching its 91% gross margin profile. From a more critical viewpoint, an inability to deploy capital productively within a core market suggests that Hithink's long-term addressable market may be more constrained than top-line growth figures imply.

Ownership and Control

As of mid-2025, founder 易峥 Yi Zheng held 36.13% of the company, co-founder 叶琼玖 Ye Qiongjiu held 11.11%, and affiliated entity Hangzhou Kaishishun Technology held 8.96%.²⁰ Together, founder and affiliate holdings exceed 50% of outstanding shares, granting the core team firm operational control without activist or takeover pressure.

That ownership concentration provides strategic stability, though it also limits external governance checks — dynamics highlighted by a share disclosure in September 2025.

On September 6, 2025, Yi disclosed a plan to trim up to 684,000 shares, while Kaishishun planned to sell 699,100 shares. Combined, the 1.3831 million shares represented 0.26% of total equity, valued at approximately ¥500 million based on the ¥361.50 closing price, with roughly ¥248 million accruing to Yi individually. The holdings had been acquired in April 2022 for ¥50.09 million, representing an unrealized gain of ¥198 million, or a 395% return.²⁰

Public friction arose primarily from the disclosure's stated rationale. In addition to noting that "the staged accumulation objective has been satisfactorily completed," the filing stated that the sale aimed to "cede market participation opportunity, release liquidity, and invigorate market vitality."²⁰ Retail investors interpreted the statement with skepticism, viewing the explanation as an awkward attempt to frame insider profit-taking as a market benefit.

Facing immediate public commentary, Yi canceled the share reduction plan just one trading day later.²¹

While the proposed transaction was small in relative size — representing 0.26% of shares from a controlling holder retaining over 36% ownership — the episode provided insight into executive communications. Framing a personal divestment as an act of market stewardship reflected weak investor messaging, while the rapid cancellation indicated heightened sensitivity to retail sentiment. For shareholders, the event highlighted the dual nature of Hithink's corporate governance: long-term founder control offers operational focus, paired at times with clumsy public relations.

The R&D Line, and What Actually Happened in 2025

A central thesis among market bulls is that Hithink maintains an insurmountable moat by out-investing peers in financial technology. However, financial results from 2025 present a more nuanced picture.

In 2025, total research and development expenditure stood at ¥1.145 billion — down 4.04% from 2024, even as total revenue expanded 44%. Consequently, R&D intensity dropped from 28.50% of revenue in 2024 to 18.99% in 2025.⁶ Executive management attributed this absolute decline to operational efficiency gains enabled by internal artificial intelligence development tools.⁶

While AI-assisted coding tools may indeed have improved developer output, the explanation warrants objective scrutiny. Attributing lower spending to AI-driven efficiency gains is a convenient narrative for software management, but one that external analysts cannot directly verify. Observational data reveals that while absolute engineering expenditure fell, sales and marketing expenses grew 28% to ¥760 million, driven by a 56% surge in advertising and promotional spend to ¥320 million.¹⁸ During a year of surging market profits, incremental capital allocation favored customer acquisition over technical R&D.

Given the surge in retail account openings across China in 2025, prioritizing user acquisition offered compelling short-term returns. Nevertheless, the shift demonstrates that industry-leading R&D intensity characterized 2024 operations more accurately than 2025. Going forward, investors will need to observe whether engineering spending reaccelerates or if 2025 established a lower baseline for R&D spending relative to total revenue. Technical talent metrics remain strong: R&D personnel accounted for 61.75% of total company headcount, including 128 PhD holders and 882 employees with master's degrees.⁶

The Acquisition That Never Happened

Perhaps the most notable characteristic of Hithink's balance sheet is the absence of goodwill. Across seventeen years as a public company spanning multiple market cycles, Hithink has avoided large-scale M&A.

During the 2014–2016 market expansion, numerous domestic tech firms acquired gaming studios, media properties, or financial services licenses at elevated valuations — acquisitions that frequently resulted in massive impairment write-downs between 2018 and 2020. Hithink avoided these deals, preserving a balance sheet free of goodwill write-offs, integration friction, or complex intangible asset amortization schedules. The company's underlying audit profile remains clean relative to sector peers, backed by comprehensive historical filings accessible via the exchange's official information platform.³³

Consequently, financial analysis centers on a small number of core accounting line items. First is contract liabilities, which reflects deferred revenue from multi-year consumer software subscriptions collected upfront. Second is the capitalization of R&D expenses: with annual technology spending exceeding ¥1 billion, accounting choices regarding expense capitalization directly impact reported net income. To date, neither area has raised audit concerns, and reported earnings in 2025 continued to track operating cash flows closely.

However, this conservative dealmaking strategy carries strategic trade-offs. The primary acquisition target Hithink declined — a securities brokerage license during the 2015 market reset — allowed competitors to capture higher-value transaction flows. Executive compensation for Yi Zheng has remained modest relative to peers at comparable market caps, with personal financial returns tied primarily to the 36% equity holding and annual cash dividends.²⁰ While this aligns executive interest with common shareholders, a controlling founder reliant on dividend income may exhibit a lower appetite for aggressive, long-term capital reinvestment than leaders seeking multi-decade enterprise expansion.

Reading Management Through Its Own Words

Hithink maintains a minimal investor communications footprint. Unlike many Western software firms, the company does not host quarterly earnings calls, communicating instead through mandated annual reports, periodic briefings, and the exchange's interactive investor Q&A portal. This limited disclosure cadence provides fewer qualitative touchpoints for evaluating strategic execution.

Where corporate disclosures can be evaluated over time, strategic messaging displays high consistency. Across multiple fiscal years, annual reporting has reiterated a uniform core model: financial information services as the primary business, technology R&D as the key differentiator, and expanding application use cases as the primary growth driver. The 2026 interim guidance maintained this consistent framing, explicitly targeting AI agents, financial data terminals, and specialized user services as current expansion vectors.²⁶ Operating history shows no sudden strategic pivots, abandoned core business lines, or restructuring write-offs.

However, public disclosures surrounding operational setbacks remain sparse. Following regulatory action against a key subsidiary in late 2024, corporate releases limited commentary to quantifying immediate revenue impacts as modest. When market participants inquired via the interactive investor platform in February 2025 whether regulatory restrictions had been fully lifted, management provided procedural responses without detailing internal compliance reforms.³¹ This minimal disclosure approach leaves open questions regarding internal operational controls, increasing scrutiny on subsequent regulatory reviews in fund distribution and investment advisory services.

Against this backdrop, Hithink's artificial intelligence initiatives represent management's primary forward-looking capital deployment — serving as a key test of whether the company intends to build next-generation platforms or primarily harvest cash flows from its legacy position.

↑ Back to Top

VII. The AI Pivot: 问财 Wencai, LLMs, & The iFind Terminal Challenge

Type this query into most financial terminals and nothing happens: "ChiNext stocks with ROE above 20%, revenue growth above 30% for three consecutive years, and net northbound inflows this week."

The prompt requires breaking down a market filter, four numerical conditions, a time window, and a data-source join—work that takes a junior analyst twenty minutes in a spreadsheet or a quant three minutes in SQL. 同花顺问财 Hithink Wencai was designed to resolve the request in a single line of Mandarin, delivering answers through progressively updated iterations long before the advent of modern large language models.

Understanding the significance of this tool requires a brief technical examination. A conventional stock screener functions like a rigid form: users select pre-set check boxes, and the software filters the dataset accordingly. While functional, it remains limited to queries anticipated by its developers. In contrast, Wencai relies on semantic parsing—converting unstructured human language into a structured database query. The primary challenge lies not in basic linguistics, but in domain-specific logic. The system must recognize that "净流入" (net inflows) refers to a specific transaction metric, that "创业板" (ChiNext) denotes a board classification rather than a business sector, and that "连续三年" (three consecutive years) requires a precise multi-period database join. Generic language models often struggle with these tasks due to specialized financial terminology and proprietary data schemas.

Hithink's primary asset is not the model itself, but the three decades of structured Chinese financial data underlying it—including tick histories, fundamental time series, regulatory archives, and capital-flow records—all normalized into a schema the parser can query. Management has described pre-training its system on a financial corpus at a trillion-token scale, supplemented by monthly additions of fresh training and fine-tuning data.²²

In January 2024, Hithink initiated beta testing and formally launched its core model, HithinkGPT, a decoder-only transformer available in 7-billion, 13-billion, 30-billion, 70-billion, and 130-billion parameter configurations with context windows up to 32,000 tokens, supporting API access, web embedding, and private deployment for institutional clients.²³²⁴ Two primary claims accompanied the release. First, HithinkGPT was among the initial financial conversational models to complete registration with China's cyberspace administration—a mandatory regulatory hurdle for public generative AI services in China. Second, the 70-billion-parameter version reportedly passed 17 financial industry credentialing tests, including securities practitioner and certified public accountant examinations, achieving an average score of 75.9.²⁴

Standardized exam scores, however, offer a limited proxy for commercial utility. Demonstrating memory recall on a CPA exam does not guarantee that retail subscribers will increase their spending or that institutional clients will replace existing software workflows.

The choice to offer five model sizes—ranging from 7 billion to 130 billion parameters—reflects a dual market strategy. Smaller models feature lower operational costs and can run on client hardware, whereas larger models offer expanded capabilities but require centralized infrastructure. By maintaining this full range, Hithink targets two distinct customer segments: its consumer application, where per-query inference costs across tens of millions of users represent a primary constraint, and institutional clients whose compliance policies prohibit sending internal research data to external servers. The private-deployment capability highlights a B2B enterprise software model designed to compete for the securities firms and asset managers traditionally served by 万得 Wind.

The central commercial question is whether AI capabilities will meaningfully transform Hithink's two primary revenue engines.

On the consumer side, artificial intelligence currently serves as a defensive mechanism rather than an offensive growth catalyst. It lowers customer support overhead, enhances retention for premium subscriptions, and widens the product gap relative to smaller rivals. However, it has not generated a clear step-change in subscription monetization. Value-added telecom services grew 20.71% in 2025—a steady performance, yet well behind the 70.98% growth in market-driven advertising during one of the strongest retail trading environments of the past decade. Had AI significantly boosted average revenue per user, subscription growth would have outpaced cyclical advertising revenue rather than trailing it.

On the institutional side, the expansion opportunity presents higher potential but faces entrenched headwinds. Hithink's iFinD terminal competes against market leader 万得 Wind, which holds a dominant position among Chinese buy-side institutions. Wind's advantage stems from deep network effects: analysts are trained on its interface, compliance departments have pre-approved its integrations, institutional spreadsheets connect to its plug-ins, and industry peers use its data syntax. To capture market share, iFinD has relied primarily on aggressive pricing. Industry surveys indicate Wind charges between ¥27,000 and ¥40,000 per seat annually, whereas iFinD has historically ranged from ¥9,000 to ¥20,000. Recent standard single-seat pricing places iFinD near ¥14,000 annually, compared to roughly ¥18,000 for 东方财富 East Money's Choice terminal, supplemented by volume and multi-year contract discounts.²⁵ Despite this pricing differential, iFinD's annual revenue remains estimated at approximately 10% of Wind's total market revenue.²⁵

The persistent market share gap demonstrates that pricing discounts exceeding 50% have been insufficient to overcome high institutional switching costs. The strategic thesis behind Hithink's AI deployment is that natural-language querying shifts the competitive focus from offering discounted feature parity to introducing an alternative analytical workflow. By eliminating the steep learning curve associated with legacy terminal syntax, natural-language interfaces could diminish Wind's primary competitive advantage: accumulated user familiarity and workflow habits.

Management has repositioned iFinD from a conventional data terminal to an "asset-management AI assistant," announcing AI product partnerships with major brokerages including Guotai Haitong and GF Securities.⁶²⁵ In its July 2026 interim guidance announcement, the company highlighted its first-half focus on embedding AI capabilities across financial intelligent agents, data terminals, and client services.²⁶

Evaluating this transition provides a clear benchmark for management's claims. If AI capabilities generate a durable competitive moat, revenue from software sales and maintenance—which contains the iFinD product line—must accelerate meaningfully beyond the 12.12% growth recorded in 2025. Concurrently, value-added telecom services should begin decoupling from underlying equity market turnover. Conversely, if segment revenue continues to track daily trading volume despite product enhancements, AI will have proven to be an internal efficiency tool and marketing asset rather than a structural growth driver.

An additional risk factor involves the underlying cost structure of large language models. Inference compute across a vast consumer base represents an ongoing operational expense, whereas Hithink's core application remains free at the point of use and monetized indirectly via advertising and product upsells. Servicing high volumes of AI queries for non-paying users risks eroding gross margins. While the company's blended gross margin expanded from 89.4% in 2024 to approximately 91.5% in 2025—indicating that AI compute costs have not yet impaired profitability¹⁸—widespread adoption of consumer AI agents could manifest initially as margin compression rather than immediate revenue expansion.

Beyond commercial costs, AI deployment carries heightened regulatory exposure. An automated system providing stock recommendations in natural Mandarin approaches the regulatory definition of investment advisory services. Under CSRC rules, delivering personalized investment advice requires specific licenses, qualified personnel, and strict compliance controls—areas where Hithink's subsidiaries have previously faced regulatory scrutiny.

↑ Back to Top

VIII. Regulatory Stress Test, Risks, & Trust-Damaging Patterns

On November 13, 2024, the CSRC's Zhejiang Bureau issued an administrative decision against 浙江同花顺云软件有限公司 Zhejiang Tonghuashun Cloud Software, a key Hithink subsidiary. The regulator's findings exposed the compliance vulnerabilities of selling investment guidance atop a high-traffic media platform.

According to the regulatory order, marketing personnel had displayed customer testimonials, screenshots of client praise, and historical stock returns without adequate risk warnings. Staff had made misleading claims, effectively promising investment returns, while several individuals delivering stock recommendations lacked required securities advisory credentials. In addition, live-stream promotional channels lacked compliance oversight and contained implicit stock tips. The bureau ordered corrective measures and suspended the subsidiary from onboarding new clients for three months.²⁷

Hithink's public disclosures sought to contextualize the financial exposure, noting that the cloud software unit accounted for ¥199 million in securities investment consulting revenue — roughly 5.58% of consolidated top-line results at the time. Management emphasized that the suspension applied solely to new client acquisition, leaving existing customer service intact, and anticipated minimal impact on full-year 2024 earnings.⁹ While technically accurate, the operational friction became apparent in subsequent reporting: sell-side research analyzing the first quarter of 2026 explicitly attributed a portion of that quarter's 112.58% net profit surge to the artificially depressed earnings base of early 2025.¹⁸²⁸ A three-month freeze on new customer onboarding at a high-margin division created a tangible financial headwind, despite initial management framing.

Eighteen months later, regulatory enforcement reached a second subsidiary operating under a distinct financial license.

On April 10, 2026, the Zhejiang Bureau issued corrective orders against 浙江同花顺基金销售有限公司 Zhejiang Tonghuashun Fund Sales, recording the violation in the national securities and futures market integrity database. The decision detailed five major compliance failures: commingling personnel and operating premises with controlling shareholders and affiliated entities; excluding fund retention scale from staff performance evaluations; failing to perform due diligence and risk assessments on fund managers and products, leaving certain offerings unrated; lacking conflict-of-interest oversight for sales of related-party private funds; and employing sales, customer service, and key-account staff who lacked required fund practitioner qualifications.²⁹³⁰

Taken together, these findings highlight broader corporate governance challenges. The commingling of facilities and the absence of conflict-of-interest controls on related-party private funds represent structural governance defects rather than clerical oversights. Simultaneously, relying on uncredentialed personnel mirrors the compliance breakdowns cited in the 2024 cloud software action. Domestic financial media covering the enforcement cited the pattern as lvfa lvfan (屡罚屡犯) — repeatedly penalized, repeatedly offending.³⁰

For long-term investors, this pattern points to a structural tension within Hithink's operational model: a high-velocity commercial engine that regularly outpaces internal compliance controls. Across three enforcement actions over a decade — the 2016 crackdown on shadow-leverage software, the 2024 advisory marketing sanctions, and the 2026 fund distribution ruling — regulatory fines have remained modest relative to annual earnings. However, the recurring nature of the infractions indicates persistent exposure.

This recurring friction stems directly from the platform's economics. Hithink's highest-margin offerings operate along the fine boundary separating passive software screeners from active investment advisory services. While a basic data filter functions as a neutral tool, highlighting specific equities alongside client profit testimonials crosses into regulated advice. Because commercial incentives encourage pushing toward that line, compliance risk remains an ongoing operational cost rather than an isolated tail event. Investors must monitor whether future regulatory enforcement escalates from temporary onboarding freezes to formal license suspensions or revocations.

The Rest of the Risk Radar

Turnover cyclicality remains the primary financial risk. Hithink's top line acts as a leveraged derivative of A-share trading activity, with approximately 90% of total revenue tied to advertising placements and consumer subscriptions — both driven by retail participation against a largely fixed wage structure. The 2025 performance, where a 44% revenue increase generated a 76% surge in net profit, demonstrated powerful operating leverage during a record volume year. However, that mechanism operates symmetrically: a severe contraction in average daily turnover would impair profitability far more sharply than revenue, as demonstrated during the 2016–2018 market downturn.

Regulatory scrutiny of AI-driven advisory represents a major strategic vulnerability. The core features that enhance an AI financial agent — personalized insights, specific stock analysis, and direct recommendations — directly expand regulatory exposure under Chinese financial laws. Given China's proactive regulatory stance on algorithmic recommendations, any policy mandating that AI-generated stock output be delivered strictly by licensed human advisors would directly constrain Hithink's premium subscription model.

Threats from major technology firms face structural barriers. Bearish arguments suggest that general-purpose large language models from major platform companies could commoditize natural-language tools like Wencai. However, the primary barrier to entry is not natural language processing, but access to licensed real-time exchange feeds, normalized historical financial datasets, direct brokerage trading APIs, and a tolerance for regulatory oversight. Big Tech platforms face substantial compliance and reputational hurdles when offering direct equity analytics to retail investors, providing Hithink with a defensible niche.

Accounting metrics highlight underlying revenue shifts. Contract liabilities — representing upfront cash collected for unfulfilled consumer software subscriptions — reached ¥1.769 billion at year-end 2025, representing a 19.05% increase.⁶ This deferred revenue figure serves as a leading indicator of subscriber demand. Its slower growth relative to overall revenue confirms that 2025 top-line expansion was driven primarily by institutional advertising spend rather than core retail subscription growth. A continued widening of this gap would signal plateauing engagement within the consumer franchise.

Brokerage industry consolidation creates customer concentration risk. Hithink's advertising division relies heavily on securities firms competing for account conversions. As regulatory policies encourage consolidation among Chinese brokerages, a smaller pool of larger financial institutions increases buyer bargaining power over advertising rates, presenting a gradual challenge to platform monetization.

Balancing these structural risks, recent financial performance highlights continued near-term momentum. On July 20, 2026, Hithink issued positive interim guidance for the first half of 2026, projecting net profit between ¥878 million and ¥979 million — a 75% to 95% increase over the ¥502 million reported in the prior-year period — driven by strong market turnover and expanding AI integration.²⁶ While long-term compliance challenges persist, the company's core market engine continues to generate substantial cash flow.

---

↑ Back to Top

IX. Strategic Frameworks & Investing Lessons

Strip away the market narrative, and the core strategic question is straightforward: what prevents a competitor from replicating this position?

Where the Power Is — and Isn't

Applying strategic framework analysis clarifies where Hithink's competitive advantages truly lie — and where they fall short.

Counter-positioning is the company's strongest structural advantage. East Money cannot credibly serve as a neutral host for every brokerage's account-opening funnel because it owns a competing brokerage. That constraint is structural rather than strategic. Hithink occupies a market position that its largest rival is legally and commercially disqualified from holding, and any new brokerage-backed entrant faces the same barrier. This represents genuine, durable power — a rare instance where the smaller market participant holds counter-positioning leverage over the larger competitor.

Network effects exist but remain weaker than two-sided platform narratives suggest. A classic network effect requires each additional user to increase platform value for other users. For Hithink, expanding retail user volume increases platform value to brokerages, while comprehensive brokerage access enhances platform utility for users. This creates a genuine cross-side distribution advantage. However, because most securities firms integrate across multiple platforms, brokerage connectivity is not a scarce asset. Furthermore, retail users derive minimal direct benefit from other retail traders on the app, as the platform lacks a strong social graph or user-generated content engine comparable to East Money's investor forums.

Switching costs are substantial, underrated, and driven by behavioral habit rather than contractual lock-in. An investor who spends years organizing watchlists, configuring technical indicators, establishing price alerts, and building muscle memory around a specific chart layout rarely switches to a competing application for incremental feature improvements. Financial software habits are exceptionally persistent — the same dynamic that protects Wind in the institutional sector operates in Hithink's favor within the consumer market.

Cornered resource claims represent the weakest element of the bullish thesis. While Hithink's three-decade financial database is extensive and costly to rebuild, it is not exclusive. Underlying exchange data remains licensable to any paying entity, and institutions like Wind and East Money maintain comparable archives. Similarly, domain-tuned model weights represent an advantage measured in quarters rather than decades. Data volume alone does not constitute a durable competitive moat here.

Scale economies and process power are largely absent. Branding power exists to the extent that the application serves as a category default, but Hithink cannot command premium pricing based on brand alone, as its consumer products compete heavily on price and feature access.

The Five Forces View

Buyer power varies sharply across segments. Retail subscribers are highly fragmented and hold zero bargaining power over subscription pricing. Conversely, brokerage advertisers represent a sophisticated, consolidating customer base equipped with precise analytics on account acquisition costs. Their bargaining power is moderate today and continues to rise as brokerage industry consolidation accelerates.

Supplier power is concentrated and structurally disadvantageous. The Shanghai and Shenzhen stock exchanges maintain monopoly control over Level-2 data feeds, which form the foundation of Hithink's highest-margin consumer subscriptions. The exchanges set pricing terms unilaterally, leaving Hithink with no alternative suppliers or negotiating leverage. Any material increase in exchange data licensing fees would directly compress profit margins in the value-added telecom division.

Rivalry remains intense yet stable, structured as an oligopoly among Hithink, East Money, and a tail of proprietary brokerage applications. The stability of this competitive equilibrium over the past decade underscores significant underlying entry barriers.

Threat of substitution centers not on rival software vendors, but on proprietary brokerage applications improving to the point where investors no longer need third-party aggregators. Major Chinese brokerages now operate capable mobile platforms. Hithink retains its audience because it aggregates across multiple institutions, delivers superior charting tools, and benefits from entrenched user habits.

Threat of new entry remains low for direct platform replicas, though adjacent financial technology entrants present ongoing competition.

Myth vs. Reality

Three prevailing market narratives require evaluation against documented results.

"Hithink is primarily an AI company." Disclosed performance does not support this framing. Financial expansion in 2025 was driven by record retail trading volume and a 71% surge in brokerage advertising spend. Artificial intelligence appears in corporate reporting as an operational feature and a cost-efficiency narrative — specifically the 4% R&D spending reduction attributed to AI-assisted developer output — rather than a distinct, high-margin revenue line item. AI integration represents a forward-looking growth option rather than a current earnings engine.

"Neutral platform positioning is inherently superior." Neutrality yields superior returns on deployed capital, but generates significantly lower absolute profit compared to an integrated brokerage model, as demonstrated by East Money's earnings lead. Platform neutrality is a defensible strategic choice rather than an absolute victory. Its resilience faces its hardest test during retail trading downturns, when institutional marketing budgets shrink while integrated brokers continue earning interest income on client cash balances.

"The balance sheet represents an unassailable fortress." The financial position is exceptionally clean, with no interest-bearing debt, zero balance-sheet goodwill, and cash and equivalents comprising roughly 88.65% of total assets. However, accumulating cash reserves does not constitute a competitive moat. Holding ¥14 billion in liquid assets alongside an 87% dividend payout ratio reflects substantial capital without an articulated reinvestment strategy beyond legacy operations.

The Activist's Angle

An institutional shareholder evaluating corporate governance would focus on three primary issues:

First, the fund distribution subsidiary remains subscale and subject to regulatory sanctions. Despite holding a distribution license and hosting China's largest retail trading audience, Hithink's fund sales division generates roughly one-fourteenth of East Money's wealth management revenue. Furthermore, regulatory findings in 2026 documented commingled personnel and facilities alongside an absence of conflict-of-interest controls for related-party funds — indicating operational underperformance combined with structural governance defects.

Second, capital allocation requires a formalized policy rather than ad-hoc distributions. Delivering an 87% dividend payout ratio returns capital effectively, but management has not articulated a long-term capital framework. Additionally, accompanying capital-reserve bonus issues — such as the four-for-ten share distribution — expand total share count without changing underlying economic value, adding friction to historical per-share analysis.

Third, regulatory compliance represents a persistent board-level risk. Three enforcement actions over ten years — two occurring within an eighteen-month span with recurring compliance themes — point to systemic operational friction. Ongoing governance evaluations must focus on structural internal control reforms rather than short-term financial penalties.

The Bull and Bear Cases

The bull case relies on counter-positioning durability, expanding Chinese retail market participation, AI driving consumer subscription ARPU and institutional terminal market share gains against Wind, and an operating structure that converts incremental revenue into high gross margins. Under this scenario, Hithink continues generating strong cash flows, returning capital to shareholders while expanding its institutional software footprint.

The bear case posits that 2025 performance reflected a cyclical market peak rather than a permanent structural shift. Highly cyclical advertising revenue drove top-line growth, while absolute R&D spending declined and regulatory scrutiny intensified. As brokerage customers consolidate, a prolonged drop in retail trading turnover would trigger significant negative operating leverage — unsupported by net interest income, brokerage balance-sheet assets, or counter-cyclical revenue streams.

Both perspectives are supported by current operational evidence, establishing Hithink's long-term trajectory as a key analytical debate.

The KPIs That Matter

Evaluating Hithink's performance requires monitoring three primary indicators:

One: A-share average daily turnover. Trading volume serves as the single most critical driver of Hithink's financial performance. Brokerage advertising spend tracks new account openings, which fluctuate with trading turnover; consumer subscription demand similarly mirrors active trading engagement. Tracking average daily turnover provides a direct leading indicator of top-line business momentum.

Two: monthly active users relative to East Money. User volume represents the core attention inventory monetized through advertising and subscriptions. Monthly active user levels indicate overall platform scale, while the competitive gap against East Money reveals whether neutral positioning successfully retains audience share against an integrated brokerage rival.

Three: value-added telecom service growth relative to trading turnover. Subscription revenue growth offers a benchmark for business model decoupling. Consumer subscriptions represent Hithink's most stable revenue stream and the primary channel where AI enhancements should drive monetization. If value-added telecom revenue outpaces market turnover across a full cycle, it validates the AI expansion thesis; if it tracks market volume, the business remains primarily a leveraged play on retail trading sentiment.

↑ Back to Top

X. Epilogue & Outro

There is a version of this story that presents a straightforward triumph, but that framing falls short of the full picture.

The full story is stranger and more instructive. In 1994, a 24-year-old electrical engineering graduate from Zhejiang University started a business to sell ¥120,000 analytical software to futures trading desks—a product for a narrow market that was soon curtailed by regulators. Thirty-two years later, that same enterprise operates the most-used financial application in a market with 250 million registered investors, generates a gross margin above 90%, carries no interest-bearing debt, has avoided major acquisitions, and returns nearly all of its earnings to shareholders.

Connecting those two eras was not a single grand strategy, but a consistent pattern of decisions that appeared unglamorous at the time: building backend brokerage plumbing when consumer portals were more prominent; cannibalizing a profitable desktop subscription line to launch a free mobile app; passing on a securities brokerage license during a peak market expansion; and refraining, across 17 years as a public company, from major M&A.

Each of those choices sacrificed immediate visibility for structural positioning. The cumulative result is a business with a focused profile: it operates a pure information platform, avoids taking custody of client assets, and converts top-line revenue into operating cash flow at a rate rarely matched in traditional financial services.

The strategic trade-offs, however, are equally distinct. Hithink generates roughly one-fifth of East Money's net income despite hosting a comparable user audience, having chosen not to monetize direct equity transactions. Consequently, its earnings remain more cyclical than those of integrated peers, leaving fewer counter-cyclical buffers when trading turnover declines. Operationally, regulatory oversight remains a persistent friction, evidenced by two enforcement decisions from regional regulators within an 18-month span targeting business lines adjacent to advisory licensing limits. Furthermore, governance friction emerged in September 2025 when the controlling founder announced a personal share reduction framed as a market liquidity enhancement, only to cancel the plan a day later following public criticism.

For long-term investors, the central analytical question is not whether Hithink represents a profitable enterprise—its returns on capital demonstrate that—but what category of asset it actually constitutes. The company operates as a capital-light, founder-controlled tollbooth on Chinese retail trading activity, backed by a counter-positioning advantage that integrated rivals cannot replicate, a forward-looking option on financial artificial intelligence, and an earnings profile that amplifies broader A-share turnover in both directions.

Over the next several years, two main issues will determine the trajectory: whether AI integration yields measurable subscription revenue or remains an internal efficiency tool, and whether a platform operating along regulatory licensing boundaries can maintain its market position without triggering stricter enforcement.

Both questions can be tracked through empirical reporting, quarter by quarter—providing a clear benchmark for evaluating the long-term thesis.

---

References

1. 1.72亿人！11月炒股软件流量回暖，14家用户数破600万 — 证券时报 STCN ↩↩

2. A股投资者，已超2.5亿！2025年新增投资者1386.95万 — 证券时报 STCN ↩

3. 2025年A股收官：420万亿成交创纪录，科技、资源双主线领航 — 21世纪经济报道, 2026-01-01 ↩

4. 同花顺：2025年年度报告摘要 — 新浪财经 / CNINFO 公司公告 ↩↩

5. 同花顺的前世今生：营收60.29亿行业排名第8，净利润32.05亿位居榜首，毛利率91.54%远超行业平均 — 新浪财经, 2026-04-30 ↩

6. 同花顺2025年业绩亮眼，广告业务成增长主引擎 — 新浪财经, 2026-03-10 ↩↩↩↩↩↩↩↩

7. 东方财富2025年净利润增长25.75% 行业龙头地位持续巩固 — 东方财富网, 2026-03-20 ↩↩↩

8. 证监会拟重罚恒生铭创同花顺，三公司总计罚没6亿元 — 澎湃新闻 The Paper ↩↩

9. 同花顺子公司暂停新增3个月，对业绩影响有限 — 新浪财经, 2024-11-18 ↩↩

10. 70后金融科技大亨易峥：50万起步到290亿身家 — 搜狐 ↩↩↩↩

11. 同花顺的发展历程 — 《管理案例精选》第3辑 ↩↩↩↩

12. 同花顺(300033) 新股发行 — 新浪财经 ↩↩

13. 中国创业板首批28家企业集体挂牌上市 — 中国新闻网, 2009-10-30 ↩↩

14. 场外配资清理收尾：恒生电子、同花顺双双被立案调查 — 新浪财经, 2015-08-19 ↩

15. 东方财富拟逾40亿收购同信证券 — 人民网财经, 2015-04-16 ↩↩

16. 东方财富收购同信证券完成过户 有望获国内券商牌照 — 中国新闻网, 2015-12-11 ↩

17. 西藏同信证券正式更名东方财富证券，万2.5佣金"扎眼" — 新浪财经, 2016-03-26 ↩

18. 【开源非银】同花顺2025年报点评：市场交投活跃驱动广告业务，预计一季报高增 — 新浪财经, 2026-03-11 ↩↩↩↩↩↩↩

19. 同花顺2025年净利大增76%，拟每10股派51元并转增4股 — 新浪财经, 2026-03-10 ↩

20. "高位接盘邀请函？"同花顺实控人减持理由"让渡机会"引爆舆论 — 21世纪经济报道, 2025-09-07 ↩↩↩↩

21. 同花顺：董事长提前终止减持计划 — 新浪财经, 2025-09-09 ↩

22. 大模型落地实践：同花顺大模型技术应用及优化 — InfoQ 写作社区 ↩

23. 内测申请开启！同花顺问财大模型—HithinkGPT来了！ — 中国证券网 CNSTOCK, 2024-01 ↩

24. 同花顺问财大模型点评：首家通过备案金融大模型，金融AI生态加速建设 — 方正证券研报 / 搜狐 ↩↩

25. 金融数据终端"四强"逐鹿AI新战场：从"卖水"到"卖铲" — 证券时报 STCN ↩↩↩

26. 上半年净利润预计增长75.00%—95.00% 同花顺大涨5.68% — 证券时报 STCN, 2026-07-20 ↩↩↩

27. 关于对浙江同花顺云软件有限公司采取责令改正、责令暂停新增客户措施的决定 — 中国证监会浙江监管局, 2024-11-13 ↩

28. 同花顺（300033）2026年一季报简析：营收净利润同比双双增长，盈利能力上升 — 腾讯新闻, 2026-04-24 ↩

29. 关于对浙江同花顺基金销售有限公司采取责令改正措施的决定 — 中国证监会浙江监管局, 2026-04-10 ↩

30. 屡罚屡犯：同花顺基金再遭监管点名，合规失守与业绩边缘的双重困局 — 新浪财经, 2026-04-13 ↩↩

31. 同花顺SZ300033：中国证监会浙江监管局责令改正、暂停新增客户措施是否结束？（投资者互动） — 新浪财经, 2025-02-18 ↩

32. Hithink RoyalFlush Information Network — 公司官网 10jqka.com.cn ↩

33. 同花顺 300033 公司信息披露页面 — 巨潮资讯网 CNINFO / SZSE ↩