Investment Research Report: Zhihu Inc. (2390.HK / NYSE: ZH) — Structural Dislocation in the Knowledge Economy

Date: December 28, 2025 Ticker: 2390.HK (Hong Kong), ZH (NYSE) Sector: Communication Services / Internet Content & Information Industry: Interactive Media & Services Current Price (HKEX): HK$9.11\ast \ast MarketCapitalization:\ast \ast HK$2.21 Billion Rating: Speculative Hold / Deep Value

1. Executive Summary

1.1 The Valuation Paradox: Negative Enterprise Value

Zhihu Inc., widely recognized as the preeminent online content community in China, presents one of the most intellectually vexing investment profiles within the global technology sector as of late 2025. The company is currently trading at a valuation that implies a deeply distressed asset, characterized by a market capitalization that is significantly below its net cash position. As of the third quarter of 2025, Zhihu reported holding approximately RMB 4.6 billion (US$673million)incash,cashequivalents,termdeposits,andshort-terminvestments.[1,2,3]Instarkcontrast,thecompany’smarketcapitalizationhoversaroundHK$2.2 billion (approximately US$280 million). This immense disparity results in a negative enterprise value (EV), theoretically suggesting that the market assigns a negative value to the company’s operating business—effectively pricing its core operations as a liability rather than an asset.

This extreme pricing anomaly signals a profound lack of investor confidence, likely anticipating one of two catastrophic outcomes: either a continued, aggressive destruction of capital through sustained operational losses, or a structural inability for shareholders to ever realize the value of the cash hoard due to corporate governance complexities or risks associated with the Variable Interest Entity (VIE) structure. The market is essentially pricing in a scenario where the cash is "trapped" onshore or will be eroded before it can be returned to equity holders.

1.2 Operational Trajectory: A Crisis of Growth

While the balance sheet appears to be a fortress of liquidity, the income statement reveals a business in the midst of a severe contraction. In the third quarter of 2025, total revenues plummeted by 22% year-over-year to RMB 658.9 million. This decline was not isolated to a single vertical but was broad-based, affecting advertising, paid memberships, and the once-promising vocational training segment. Following a brief period of non-GAAP profitability in the first half of 2025—which had sparked hopes of a sustainable turnaround—the company slipped back into a net loss in the third quarter, with losses widening to RMB 46.7 million from RMB 9.0 million in the prior year.

The degradation in fundamentals is further evidenced by the company’s opaque reporting changes; specifically, the cessation of disclosing Monthly Active Users (MAUs), a standard metric for growth stocks, signals a defensive posture. Furthermore, average monthly subscribing members declined to 14.3 million , indicating that the platform's utility or appeal may be reaching a saturation point or suffering from competitive displacement.

1.3 Strategic Pivot: The AI and Storytelling Gamble

Zhihu is undergoing a radical identity transformation. It has largely moved away from its origins as a pure "Quora clone" focused on professional knowledge sharing. The platform now relies heavily on "Yan Selection" (Salt Selection), a paid membership model driven by short-form fiction and serialized stories. This pivot places Zhihu in direct competition with digital entertainment giants rather than functioning solely as a knowledge repository. Simultaneously, the company is aggressively deploying artificial intelligence, with its "Zhihu Zhida" search and AI writing tools achieving significant penetration. The strategic thesis is that AI can lower content creation costs and revolutionize search monetization, yet this transition carries existential risks: if AI answers questions instantly, the necessity for browsing a community forum diminishes, potentially eroding the advertising inventory that underpins a significant portion of revenue.

1.4 Investment Thesis and Recommendation

We initiate coverage with a Speculative Hold rating. The stock offers a massive margin of safety purely on a tangible book value basis, trading at roughly 0.55x Price/Book. However, it represents a classic "value trap" risk. The "Negative Enterprise Value" is a rational market discount applied to a shrinking, loss-making entity operating in a challenging regulatory jurisdiction with a dual-class share structure that limits minority influence. Unless the company demonstrates a sustainable floor for revenue and a clear capital allocation strategy—such as massive dividends or aggressive buybacks beyond the current US$66.5 million executed—the stock is likely to remain stagnant despite its cash richness.

2. Business Drivers and Ecosystem Analysis

2.1 The Evolution of the "Community-Content-Commerce" Flywheel

Founded in 2010 by Zhou Yuan, Zhihu began as an invitation-only community, a strategy that successfully seeded the platform with a high density of experts, academics, entrepreneurs, and industry insiders. For the first decade of its existence, Zhihu's primary economic moat was its reputation for "professionalism" and "depth." In a Chinese internet landscape dominated by entertainment-first platforms, Zhihu stood out as a sanctuary for serious discourse. This created a virtuous "Community-Content-Commerce" loop:

1. Community: High-quality users generated trust and prestige.

2. Content: This trust facilitated the creation of high-engagement, long-tail content (evergreen answers) that ranked highly on search engines (Baidu, Sogou).

3. Commerce: The high-trust environment allowed for sophisticated "recommendation marketing" (Zhi+ / Chic) and, eventually, paid subscriptions for premium content.

However, as the company sought to expand its total addressable market (TAM) post-IPO, it aggressively pursued mass-market user growth. This expansion diluted the density of professional users, leading to a deterioration in the signal-to-noise ratio. The tension between the "old guard" (who valued rigorous academic debate) and the "new wave" (who consumed mass-market entertainment and stories) has created a fragmented user experience.

2.2 The "Yan Selection" Pivot: From Knowledge to Fiction

A critical, often misunderstood driver of Zhihu’s current business model is its paid membership segment, which generated RMB 385.6 million in Q3 2025. Investors often mistakenly compare Zhihu’s subscription model to LinkedIn Premium (professional access) or The New York Times (news). In reality, the primary driver of Zhihu’s subscription revenue is short-form fiction.

2.2.1 The Mechanism of "Yan Story"

The "Yan Selection" membership gates access to a vast library of user-generated fiction. This content usually originates from open questions such as "What would you do if you woke up in the Qing Dynasty?" or "Tell a story about a revenge plot." Users write "answers" that are effectively serialized novels.

• Monetization: The first few chapters are free (the "answer"), but the conclusion is locked behind the "Yan Selection" paywall.

• Scale: By 2025, this segment had grown to define the platform's consumer identity. However, revenue from this segment fell 16% year-over-year in Q3 2025.

• Strategic Friction: Management has admitted that the segment is in a "transition period." CEO Zhou Yuan noted the challenge of distinguishing "fiction" from "non-fiction" to prevent user confusion, using the analogy of a library needing to separate genres. This decline suggests that the novelty of "Zhihu stories" is wearing off, or that the platform is losing ground to dedicated free reading apps (like Fanqie) and short-drama platforms on Douyin/Kuaishou.

2.3 Marketing Services: Structural and Cyclical Headwinds

Marketing services revenue collapsed by approximately 26% year-over-year in Q3 2025 to RMB 189.4 million. This decline is symptomatic of broader structural shifts in the Chinese digital advertising market, which entered a "regulated maturity phase" in 2025.

2.3.1 The Shift to Hard Conversion

In an environment of economic uncertainty and "consumption downgrades," advertisers in China have shifted their budgets aggressively toward "hard conversion" platforms.

• The Competitors: Platforms like Douyin (ByteDance) and Pinduoduo offer closed-loop ecosystems where an ad view leads directly to a purchase.

• Zhihu's Weakness: Zhihu primarily functions as a "consideration" or "awareness" platform. Users come to read reviews or understand product mechanisms, but the actual transaction happens elsewhere (JD.com, Taobao). In a deflationary environment, brands cut "awareness" budgets first to protect ROI.

• Regulatory Impact: The 2025 implementation of the revised Anti-Unfair Competition Law and stricter internet advertising measures has increased compliance costs and reduced the efficacy of certain aggressive tracking and targeting methods that Zhihu might have utilized.

2.4 Vocational Training: The Failure of a Growth Pillar

Vocational Training was once touted as the "third engine" of Zhihu’s growth, leveraging its demographic of young, educated, upwardly mobile users to sell courses on civil service exams, postgraduate studies, and coding. However, in Q3 2025, the company reclassified this revenue stream into the "Others" segment due to its shrinking scale.

• Strategic Implication: This reclassification is a tacit admission of defeat. Management noted that the segment is "no longer a drag on the bottom line," implying that it has been downsized to break-even rather than scaled for growth.

• Root Cause: The vocational education market in China is fiercely competitive, dominated by specialists like Offcn and Fenbi. Zhihu’s "traffic-based" model (selling courses to users browsing content) likely suffered from lower conversion rates and high customer acquisition costs compared to purpose-built education platforms.

2.5 AI Integration: "Zhihu Zhida" and the LLM Pivot

Facing these headwinds, Zhihu has bet its future on Artificial Intelligence. The company launched "Zhihu Zhida," an AI-powered search and answer engine, and integrated AI writing tools for creators.

• Adoption: AI writing tools have surpassed 20% penetration among creators, indicating strong uptake.

• The Double-Edged Sword:

  • The Threat: The existential risk is that Large Language Models (LLMs) like Baidu's Ernie Bot or global competitors render "Q&A sites" obsolete. If a user can get a synthesized, accurate answer from an AI agent instantly, the utility of browsing through Zhihu’s forum threads diminishes. This "zero-click" future destroys the page-view-based advertising model.

  • The Opportunity: Zhihu possesses a unique proprietary dataset—14 years of high-quality, conversational, Chinese text data—which is invaluable for training LLMs to understand nuance, logic, and culture. However, monetization of this data (through licensing) has not yet materialized as a significant revenue line item in the financial statements.

3. Macroeconomic and Regulatory Environment

3.1 The "Regulated Maturity" of China's Ad Market

The China Internet Advertising Market Report for H1 2025 characterizes the sector as entering a "regulated maturity phase," with steady but low growth (5.6%).

• Consolidation: The market is consolidating around "super apps" (WeChat, Douyin). Mid-tier platforms like Zhihu are being squeezed out of media plans.

• Demographic Pressure: Zhihu’s core demographic—educated urban youth—is at the epicenter of China's youth unemployment challenge. This reduces the purchasing power of the audience, making them less attractive to premium advertisers (automotive, electronics, luxury) that traditionally supported Zhihu’s ad rates.

3.2 Regulatory Landscape: A Minefield of Compliance

• Content Regulation: As a text-heavy platform, Zhihu faces constant scrutiny under China's strict cybersecurity and content laws. The platform has previously faced fines and temporary suspensions. The 2025 regulatory outlook emphasizes "social stability" and "consumer protection," necessitating heavy investment in moderation (both human and AI) to prevent "harmful" content from proliferating.

• Anti-Unfair Competition Law (2025 Revision): The new rules prohibiting "involution-style" competition and enhancing platform responsibility could theoretically benefit Zhihu by curbing the predatory practices of larger rivals. However, it also raises the compliance bar for Zhihu’s own commercialization algorithms.

• VIE Structure Risks: Despite the resolution of the PCAOB audit inspection dispute, the Variable Interest Entity (VIE) structure remains a legal gray area. Foreign investors do not own the operating assets in China but rather a contractual right to the profits. The massive discrepancy between Zhihu’s onshore cash (held by VIEs) and its offshore market capitalization highlights the "China discount" applied to this structure. Investors fear that in a worst-case geopolitical or regulatory scenario, these contracts could be declared void, leaving offshore shareholders with nothing.

3.3 The AI Regulation Frontier

China has moved quickly to regulate Generative AI. Platforms are responsible for the output of their AI agents. This creates a significant liability for Zhihu as it rolls out "Zhihu Zhida." If the AI generates historically nihilistic or politically sensitive content, Zhihu bears the penalty. This necessitates conservative tuning of the models, which may reduce their utility compared to unrestricted global competitors.

4. Financial Performance (2024-2025)

4.1 Revenue Decomposition: The Shrinking Top Line

The financial data from 2024 through the third quarter of 2025 paints a picture of a business in secular contraction.

Table 1: Quarterly Revenue Trend (RMB Millions)

Analysis: The acceleration of the decline is particularly worrying. From a full-year contraction of 14.3% in 2024 to a 22% drop in Q3 2025, the negative momentum is building. The company is finding it increasingly difficult to monetize its user base in the current economic climate.

4.2 Profitability Analysis: The Illusion of Cost Cutting?

Zhihu achieved a significant milestone in the first half of 2025 by recording Non-GAAP profitability, a first in its history.

• Q2 2025: Adjusted Net Profit of RMB 91.3 million.

• Q1 2025: Adjusted Net Profit of RMB 6.9 million.

• Q3 2025 (The Reversal): Adjusted Net Loss of RMB 21.0 million.

Insight: The return to losses in Q3 2025 is a critical red flag. It suggests that the profitability achieved in H1 was likely driven by aggressive, one-time cost cuts (layoffs, slashed marketing spend) rather than true operating leverage. Revenue fell faster than costs could be cut in Q3. The Gross Margin deteriorated to 61.3% in Q3 2025 from 63.9% a year prior , indicating lost pricing power or a mix shift toward lower-margin revenue streams.

4.3 Balance Sheet Analysis: The Fortress and the Trap

The most compelling, yet frustrating, aspect of Zhihu is its balance sheet.

• Cash Position: As of June 30, 2025, the company held RMB 4.82 billion (US$673 million) in cash, cash equivalents, term deposits, and short-term investments.

• Liability Profile: The company has negligible debt. The primary liabilities are accounts payable and accrued expenses.

• Market Cap vs. Cash: With a market cap of ~US$300 million, the stock trades at roughly 0.45x Net Cash.

• Cash Burn: The Q3 Net Loss was RMB 46.7 million. Even at this run rate, the cash runway is theoretically decades long.

The "Trust Discount": Why is the market ignoring $670 million in cash?

1. Trapped Cash: Much of this cash is likely held onshore in PRC subsidiaries or VIEs. Repatriating it to pay dividends to offshore shareholders involves significant withholding taxes and regulatory hurdles regarding foreign exchange controls.

2. Management Intent: The company stated in its annual report that it "currently intends to retain most, if not all, of our available funds" for operations and expansion. The lack of a massive special dividend or a more aggressive buyback signals that management wants to keep the war chest for "strategic initiatives" (potentially AI investments), which the market fears will result in low-ROI spending.

4.4 Shareholder Returns

• Repurchase Program: As of Q3 2025, Zhihu had repurchased approximately 31.1 million Class A shares for a total price of US$66.5 million.

• ADS Ratio Change: In May 2024, Zhihu changed its ADS ratio from 1 ADS = 0.5 Class A shares to 1 ADS = 3 Class A shares (effectively a 1-for-6 reverse split). This was likely done to maintain compliance with NYSE minimum listing price requirements, a move often associated with distressed companies.

5. Risk Assessment

5.1 Governance and Corporate Structure

• Dual-Class Share Structure: CEO Zhou Yuan holds Class B ordinary shares, which entitle him to 10 votes per share, whereas Class A shareholders (public investors) get one vote per share. This structure concentrates approximately 43% of the aggregate voting power in the hands of the founder. This renders minority shareholders effectively powerless to influence strategy, replace management, or force a sale/liquidation to realize the cash value.

• Key Man Risk: The company’s vision is inextricably linked to Zhou Yuan. His resignation or removal would cause significant disruption, yet his current strategy is being heavily questioned by the market.

5.2 The "Value Trap" Risk

A classic value trap occurs when a stock looks cheap (low P/B, high cash) but the underlying business fundamentals continue to deteriorate, eroding the "value" over time.

• Erosion: If Zhihu loses RMB 200 million a year, the cash pile shrinks slowly. However, if the enterprise value remains negative, the stock price may stagnate indefinitely.

• Catalyst Vacuum: Without a dividend, significant buyback expansion, or privatization offer, there is no mechanical force to "unlock" the discount.

5.3 Technical Analysis and Market Sentiment (December 2025)

The technical indicators reinforce the fundamental pessimism.

• Moving Averages: The stock is trading below its 50-day and 200-day Moving Averages (MA). The 200-day MA is acting as a strong resistance level at approximately HK$9.78. The "death cross" (short-term MA crossing below long-term MA) signals a sustained bearish trend.

• RSI: The Relative Strength Index (14) is hovering between 39-45, indicating the stock is nearing oversold territory but lacks the momentum to break out.

• Trend: Technical analysis summaries classify the stock as a "Strong Sell" across most timeframes due to the lack of buying pressure.

6. 5-Year Scenario Analysis (2025-2030)

We model three potential trajectories for Zhihu to assess the investment expected value.

Scenario A: The "Slow Bleed" (Base Case - 50% Probability)

• Narrative: Revenue continues to decline at a managed rate of 5-10% annually as the ad business shrinks and membership saturation persists. Management executes enough cost-cutting to maintain break-even or slight losses. The cash pile is slowly drawn down for "AI experiments" that yield mediocre returns but prevent total obsolescence.

• Financials: Revenue settles at ~RMB 2.5 billion by 2030. Net Income remains negligible.

• Valuation: The market continues to price the company at roughly 0.5x Book Value, viewing it as a stagnant utility. The stock price drifts sideways or slowly downward, tracking the erosion of book value.

• Outcome: "Dead money" for investors.

Scenario B: The "Privatization" (Bull Case - 30% Probability)

• Narrative: The CEO or a consortium (potentially backed by Tencent or Baidu) recognizes the absurdity of the negative enterprise value. They launch a take-private bid to delist the company.

• Catalyst: Continued share price weakness makes the buyout cheaper than the cash on hand. A bid is made at a premium (e.g., HK$14-16), which is still below the cash value per share but offers a 50%+ upside from current distressed levels.

• Rationale: Being a private company would allow Zhihu to restructure without the quarterly scrutiny of public markets.

• Outcome: Significant one-time gain for current holders.

Scenario C: The "AI Renaissance" (Optimistic Growth - 20% Probability)

• Narrative: Zhihu successfully monetizes its proprietary data for LLM training, signing high-margin licensing deals with major AI labs. The "Yan Story" IP becomes a major content source for video streaming adaptations (similar to the success of The Bad Kids or Three Body Problem). "Zhihu Zhida" becomes a preferred search engine for professional/academic queries, taking share from Baidu.

• Financials: Revenue growth returns to +15% CAGR. Net profit margins expand to 10% as high-margin data licensing kicks in.

• Valuation: The market reprices the stock to 15x P/E plus the cash balance.

• Outcome: Multi-bagger return (3x-5x).

7. Qualitative Scorecard

Table 2: Investment Quality Scorecard (1-5 Scale)

8. Conclusion

Zhihu Inc. is a textbook example of the disconnect between price and value in the current Chinese equity market.

• The Price: ~HK$9.11 per share implies the business is practically worthless and the cash on the books is discounted by nearly 50%.

• The Value: A platform with ~14 million paying members, a massive proprietary database suitable for AI training, and US$673 million in the bank.

However, cheapness is not a catalyst. The company is shrinking, and its core utility as a search/knowledge engine is being challenged by Generative AI. The Q3 2025 return to net loss is a severe disappointment after the H1 profitability narrative, damaging management's credibility regarding a "sustainable turnaround." The strategic pivots—from Q&A to Stories to AI—feel reactive rather than proactive.

Investment Verdict: Zhihu is uninvestable for growth investors. The "growth story" is currently broken. For deep value and special situations investors, the stock warrants a position only if one has high conviction that a privatization or aggressive capital return policy (buybacks/dividends) is imminent. The "Negative Enterprise Value" provides a theoretical floor, but without a change in trajectory or capital allocation, the ceiling is equally low.

Action: We recommend avoiding the stock unless it drops further to extreme levels (e.g., <0.3x Cash), or waiting for a definitive signal of capital return (e.g., announcement of a special dividend or a massive tender offer).

Detailed Analysis

1. Company Overview & History

Zhihu, which translates to "Do you know?" in Classical Chinese, launched in 2011. It began as an invitation-only community for entrepreneurs, developers, and intellectuals. This initial scarcity created a culture of high-quality, long-form discourse that differentiated it from mass-market forums like Baidu Tieba.

1.1 Evolution of the Business Model

• 2011-2016 (The "Utopia" Phase): Pure User Generated Content (UGC) community. Monetization was minimal. The focus was on building the "Community" and "Content" pillars of the flywheel.

• 2016-2019 (The "Knowledge Pay" Phase): Introduction of Zhihu Live (interactive talks), e-books, and tipping. This was an attempt to monetize the expertise of its "Big V" (verified influencers) users directly.

• 2019-2023 (The "Story" Phase): Launch of "Yan Selection" (Salt Selection) Membership. This marked the pivot from "Knowledge" to "Stories." The platform incentivized users to write serialized fiction disguised as answers. This significantly lowered the barrier to content creation (anyone can tell a story; not everyone is an expert in quantum physics) and expanded the TAM.

• 2024-Present (The "AI & Efficiency" Phase): Facing user saturation and a slowing ad market, Zhihu is attempting to use AI to lower costs and improve search utility while managing a declining topline. The focus has shifted from "Growth at all costs" to "Quality commercialization."

2. Market Dynamics: The "Knowledge Economy" in China

2.1 The Competitive Landscape: A War on Multiple Fronts

Zhihu operates in a fragmented attention economy, fighting battles on three distinct fronts.

2.1.1 The Video Competitors: Bilibili and Kuaishou

Zhihu competes for the same "educated youth" demographic as Bilibili.

• Engagement: Bilibili offers significantly higher engagement through video (average time spent >90 mins vs. Zhihu’s lower engagement).

• Valuation Comparison: Bilibili (9626.HK) trades at much higher multiples (P/S ~1.8x vs. Zhihu ~0.7x). This premium is awarded because Bilibili has successfully transitioned to profitability while maintaining growth, whereas Zhihu is shrinking.

• Content Format: The shift from text to video is a secular trend. Zhihu’s attempts to introduce video have largely failed to gain traction against Bilibili’s dominance in "long-form" and Douyin’s dominance in "short-form."

2.1.2 The "Search" Competitors: Baidu and Douyin

• Douyin/TikTok: The apex predator of time spent. Douyin’s search function is increasingly replacing traditional search engines and Zhihu for "how-to" queries (e.g., "how to cook a steak," "best travel itinerary"). The visual nature of Douyin is often preferred for practical advice.

• Baidu: While often criticized for quality, Baidu remains the default search engine. With its "Ernie Bot" integration, Baidu is upgrading its search experience to provide direct answers, bypassing the need for users to click through to Zhihu links.

2.1.3 The Lifestyle Competitors: Xiaohongshu (RedNote)

Xiaohongshu has encroached on Zhihu’s territory for product reviews and practical advice.

• Demographics: Dominated by female users and lifestyle content.

• Commercialization: Xiaohongshu offers a more visual and "shoppable" experience. Advertisers prefer it for beauty, fashion, and travel brands because the path to purchase is shorter and clearer.

2.2 The AI Threat to Q&A: An Existential Crisis

The fundamental utility of a Q&A site is "Information Retrieval."

• Old Model: User searches Google/Baidu -> clicks Zhihu link -> reads 5 different answers -> synthesizes truth.

• New Model (AI): User asks LLM -> LLM reads Zhihu (backend) -> synthesizes truth -> User never visits Zhihu.

• Impact: This disintermediation threatens Zhihu’s ad revenue (fewer pageviews) but potentially increases the value of its data (licensing to LLMs). However, China’s data laws make licensing complex, and Zhihu must navigate the "tragedy of the commons"—if it blocks LLMs, it loses relevance; if it allows them, it loses traffic.

3. Financial Deep Dive: The Numbers Behind the Narrative

3.1 Revenue Analysis: Decomposition of the Decline

The -22% YoY revenue drop in Q3 2025 is composed of specific failures in key segments.

• Paid Membership (-16% YoY): This segment was supposed to be the stable, recurring revenue engine. The decline suggests that the price elasticity of the "Yan Selection" membership is being tested. Users may be cancelling as economic pressure mounts. Furthermore, the "Transition" to blend fiction and non-fiction indicates product confusion. If a user subscribes for "Knowledge" and is bombarded with "CEO Romance Novels," they churn. If a user subscribes for novels and gets academic papers, they churn. Zhihu is trying to be everything to everyone and failing at both.

• Marketing Services (-26% YoY): This is a beta-play on the Chinese economy. Zhihu’s ads are a mix of performance-based and brand-based. In a downturn, "brand building" (Zhihu’s strength) is cut in favor of "sales conversion" (Douyin’s strength). The decline is steeper than the broader market, implying market share loss.

3.2 Cost Structure and Margins

• Gross Margin: Dropped to 61.3% (Q3 2025) from 63.9% (Q3 2024).

  • Reasoning: Revenue sharing costs (paying content creators) likely remained fixed or increased as a percentage of revenue. As the high-margin ad revenue falls, the lower-margin membership revenue (which requires payouts to authors) constitutes a larger portion of the mix, compressing the blended margin.

• Operating Expenses:

  • SG&A: Selling and Marketing expenses have been slashed aggressively (-19.4% in Q3). This is the primary driver of the narrowing losses in previous quarters. However, cutting marketing too deep accelerates the revenue decline—a vicious cycle where the company saves its way to irrelevance.

  • R&D: Continued investment in AI prevents deeper cuts here. R&D is sticky due to the "AI arms race." Zhihu cannot afford to stop developing "Zhihu Zhida" lest it fall completely behind Baidu.

3.3 The Share Repurchase Program

Zhihu has authorized significantly more buybacks than it has executed.

• Status: Repurchased ~31.1 million Class A shares for US$66.5 million as of Q3 2025.

• Capacity: The program allows for more, but the pace is cautious.

• Impact: At current prices, $66 million represents nearly ~20% of the float. The fact that the stock price has fallen despite this buying pressure indicates overwhelmed selling pressure from early investors (VCs) or employees cashing out their stock-based compensation.

4. Technical Analysis (Dec 2025)

The technical indicators paint a bearish picture in the short term, consistent with the fundamental deterioration.

• Support/Resistance Levels:

  • Support: HK$8.83(Short-termMAfloor).Abreachofthislevelopensthedoortothe52-weeklowofHK$8.30.

  • Resistance: HK$9.21(Dayhigh)/HK$9.78 (Long-term MA). The stock has repeatedly failed to break above the HK$10.00 psychological barrier.

• Oscillators: The RSI (14) at ~42 suggests the stock is weak but not yet at "panic selling" levels (usually <30). This implies a slow bleed rather than a capitulation, which is often more dangerous for investors as there is no "V-shape" bounce signal.

• Volume: Volume has been relatively light , suggesting a lack of institutional interest. The stock is drifting on retail flow and algorithmic trading.

5. Valuation Models

5.1 Relative Valuation

Comparing Zhihu to peers Bilibili (BILI), Kuaishou (1024), and Weibo (WB).

Table 3: Valuation Multiples Comparison (2025 Estimates)

Analysis: Zhihu is the cheapest by far, but it is also the only one shrinking. Weibo is a "mature cow" (low growth, high profit), Bilibili is a "growth star" (high growth, turning profit). Zhihu is in "distress" (negative growth, loss). The discount is warranted.

5.2 Liquidation Value (Net Net Working Capital)

• Cash & Short Term Investments: RMB 4.82 Billion.

• Total Liabilities: ~RMB 1.5 Billion (Estimate based on typical payables/accruals for this revenue base).

• Net Cash Value: ~RMB 3.3 Billion.

• Current Market Cap: ~RMB 2.2 Billion.

• Margin of Safety: The stock trades at ~66% of its Net Cash value. Even if we assume $0 value for the brand, platform, and IP, the stock is undervalued if the cash can be accessed.

• Burn Adjustment: Assuming a burn of RMB 200M/year, the company has 16.5 years of cash runway. This confirms that the risk is not bankruptcy, but value trapping.

6. Strategic Recommendations for Management (Hypothetical)

If we were advising the Board of Directors, we would recommend:

1. Massive Tender Offer: Use RMB 2 Billion of the cash to launch a tender offer for shares at HK$12.00. This would instantly close the valuation gap, reward long-suffering shareholders, and accrete value for remaining holders. It sends a signal that management believes the stock is undervalued.

2. Dividend Policy: Institute a fixed dividend yield of 5-7% to attract income investors and put a floor under the stock. With $600M+ in cash, they can afford a $30-40M annual dividend indefinitely.

3. Spin-off AI/Data: Separate the data-licensing business into a distinct entity to highlight its value to potential acquirers (like Microsoft/OpenAI equivalents in China).

4. Kill the "Others": Fully exit vocational training if it distracts from the core. Focus purely on "High-End Q&A" (retention) and "Yan Story" (revenue).

7. Conclusion

Zhihu Inc. is a company in search of a soul in the AI era. It has money, it has data, but it is losing its users and its identity. The financials for Q3 2025 confirm that the turnaround is not yet successful. The revenue contraction is too steep to be ignored, even with the cost-cutting.

However, at 0.55x Book Value, the market has priced in bankruptcy. Zhihu is not going bankrupt. It is simply shrinking. Therefore, the risk/reward is asymmetric to the upside only for patient capital willing to wait for a capital allocation event (buyout/dividend). For everyone else, the opportunity cost of holding this "melting ice cube" is too high compared to faster-growing peers like Bilibili or safer yields like Weibo.

Final Rating: Speculative Hold.