Zhihu Inc. (ZH)

Executive Summary / Key Takeaways

• Profitability Through Amputation, Not Growth: Zhihu achieved its first full-year non-GAAP profit in 2025 (RMB 37.9M) by slashing costs despite a revenue decline of 23.6%, raising questions about whether this is a sustainable turnaround or a financial adjustment that masks deeper business model challenges.

• The AI Paradox: While Zhihu's professional content is cited by AI assistants 4x more than any competitor, the same AI wave is siphoning entry-point traffic from its core Q&A platform, creating an existential threat that management is attempting to counter by becoming an AI monetization player itself through IP adaptation and data services.

• Balance Sheet Arbitrage Opportunity: Trading at $3.03 with a market cap ($269M) significantly below its RMB 4.5B cash hoard (≈$636M), Zhihu offers downside protection, but this also reflects market skepticism that the business can generate sustainable returns above its cost of capital.

• Segment Divergence Signals Transition Pains: Marketing services revenue bottomed in Q3 and recovered 24% sequentially in Q4, suggesting the adjustment cycle is ending, while paid membership revenue declined 21% YoY as the company prioritized unit economics over scale, and IP monetization grew significantly but remains a small portion of the total mix.

• Critical Variable for 2026: The investment thesis hinges on whether Zhihu's new AI-driven revenue streams—particularly AI data services and short-form drama adaptation—can scale from experimental to material within 12-18 months before its cash cushion erodes.

Setting the Scene: The Professional Knowledge Community at a Crossroads

Zhihu Inc., founded in 2010 and headquartered in Beijing, has spent fifteen years building China's premier online content community for professional knowledge sharing. Unlike its more entertainment-focused peers, Zhihu carved out a defensible niche as the go-to platform for in-depth, expert-driven Q&A content that users trust for career decisions, technical learning, and industry insights. This positioning became more valuable after LinkedIn (MSFT) ceased operations in mainland China in 2023, leaving Zhihu as the de facto professional knowledge hub.

The company generates revenue through three segments: Marketing Services (ad solutions leveraging its expert network), Paid Membership (premium content subscriptions), and Other Revenues (vocational training and IP monetization). For years, Zhihu pursued growth at the expense of profitability, burning cash to expand its user base and content library. That strategy reversed in 2025 when management initiated a disciplined operating strategy focused on cost control, revenue mix optimization, and AI integration across all business lines.

This pivot occurred against a shifting industry backdrop. Generative AI is reshaping content consumption, with users increasingly querying AI assistants directly rather than visiting Q&A platforms. According to the Qubit Research Institute, AI assistants cite community platform content 62.4% of the time in professional domains, with Zhihu ranking highest—cited almost 4 times as much as the second-place platform. This validates Zhihu's content quality but also highlights the threat: if AI can extract and synthesize Zhihu's knowledge without users ever visiting the platform, the core advertising and subscription models face disintermediation.

Zhihu's competitive landscape reflects this squeeze. Bilibili (BILI) dominates video-based knowledge sharing with 300M+ MAUs and grew revenue to RMB 30.35B in 2025 while achieving its first annual profit. Weibo (WB) commands real-time discourse with 500M MAUs and maintains 75.99% gross margins. Baidu (BIDU) integrates search and Q&A with its Ernie Bot AI, generating over RMB 10B in AI revenue. Zhihu's 2025 revenue of RMB 2.75B and 12.2M paid members are smaller in comparison, revealing a scale disadvantage that limits its bargaining power with advertisers and content creators.

Technology, Products, and Strategic Differentiation: The AI Integration Gamble

Zhihu's response to disruption is to become a disruptor itself. The company is accelerating AI integration across its community and commercial models, betting that its unique assets—high-quality content, expert network, and authentic discussion scenarios—can be monetized directly in the AI era rather than being passively scraped by others.

The centerpiece is Zhida AI, which evolved into an agentic mode by September 2025. Zhida now delivers more accurate search results, understands user intent, performs multistep reasoning, and synthesizes information across topics. Its penetration rate exceeded 15% in Q3 2025, nearly 4x higher than the prior year. This transforms Zhihu from a passive content repository into an active AI assistant, potentially recapturing users who might otherwise go to ChatGPT or Baidu's Ernie Bot. The December 2025 upgrade introduced cross-topic content aggregation and hot trend summarization, driving a double-digit increase in click-through rates for AI direct answer cards and increasing average AI search interactions per user.

For content creators, Zhihu rolled out AI-powered publishing tools with one-click enhancement, intelligent editing, automated formatting, and image pairing capabilities. This lowers creation barriers and improves readability, which is critical because Zhihu's community-driven content model depends on a steady flow of high-quality contributions. The "Professional Search" feature launched in October 2024 covers over 50 million Chinese and English literature sources, supporting file uploads and long document parsing—directly targeting researchers and professionals who need verified sources.

The more significant monetization bet lies in two new AI-driven avenues. First, AI-enabled short-form drama and comic adaptation leverages Zhihu's Yan'an Stories IP library and active creator ecosystem. With text-to-video and image-to-video models streamlining production, Zhihu can adapt its high-quality copyrighted stories into premium video content at near-zero marginal cost. The adapted short dramas "Fang and Xia" and "The Seventh Year Of Secret Love For My Childhood Friend" achieved high popularity, proving demand exists. Second, AI data services position Zhihu to provide high-value, traceable, structured data solutions for AI developers. Given its expert network and authentic discussion scenarios, Zhihu can offer differentiated data that improves model training quality—a business management explicitly states has clear potential for positive cash flow.

These initiatives represent a fundamental shift from advertising-dependent revenue to direct AI monetization. However, they remain nascent. IP monetization revenue grew 5x year-over-year in Q4 2025 and doubled for the full year, but its absolute size is still small relative to Zhihu's RMB 2.75B total revenue. The risk is that these promising experiments fail to scale before the core business deteriorates further.

Financial Performance & Segment Dynamics: Evidence of Strategic Execution

Zhihu's 2025 financial results provide evidence of management's strategic priorities: profitability over growth, quality over quantity, and AI investment over cash returns. The company achieved its first-ever full-year non-GAAP profit of RMB 37.9M, a turnaround from a RMB 96.3M loss in 2024. However, this profit margin of 1.3% on RMB 2.75B revenue reveals the achievement as primarily a result of cost savings. The question is whether this leaner cost structure can support future growth or merely preserves a shrinking business.

Total revenue declined 23.6% year-over-year, with Q4 2025 revenue falling 25% to RMB 643.5M. This reflects deliberate decisions to optimize revenue mix. The marketing services segment underwent an adjustment cycle to eliminate low-quality content, which reduced consumption of such content by over 90% by Q1 2025. This strengthened the platform's credibility. The strategy shows signs of traction: after declining 40% YoY in Q1, marketing services revenue bottomed in Q3 and surged 24% sequentially in Q4 to RMB 234.8M, which management called an inflection point. This recovery was driven by stronger client quality and deeper industry penetration in verticals like technology and healthcare.

Paid membership revenue declined 21% YoY in Q4 to RMB 333.5M as average monthly paid members fell to 12.2M from 14.1M a year ago. Management prioritized unit economics over scale, optimizing the member structure to attract users with longer life cycles. The strategy shows early signs of success: Q4 ARPU increased 1.4% sequentially, and overall quarterly renewal rates improved by 2.7 percentage points. Voice live streaming, a premium benefit launched in Q2, saw ARPU exceed expectations. This focus on profitable subscribers rather than vanity metrics could build a more durable subscription business.

The "Other Revenues" segment, which now consolidates vocational training and IP monetization, reflects a tale of two strategies. Vocational training revenue declined 35% YoY in Q1 and was reclassified in Q3 to improve profitability. However, the segment's operating profit increased 90% YoY in Q2 despite revenue decline, demonstrating the effectiveness of the strategic refinement. Meanwhile, IP monetization revenue grew more than fivefold in Q4, suggesting Zhihu's content library has significant untapped value. The core advantage of this model is that once a high-quality story is created, the marginal cost of reuse across multiple scenarios is low while revenue can be amplified.

Gross margin was 53.6% in Q4 2025 compared to 62.9% a year earlier, primarily due to ongoing efforts to broaden and enhance content offerings. Full-year gross margin of 59.9% was only slightly below 2024's 60.6%, showing cost discipline. Operating expenses increased in Q4 due to a one-time RMB 126.3M goodwill impairment charge from prior acquisitions, but underlying expenses continued declining year-over-year. Selling and marketing expenses fell 13% YoY, and R&D expenses decreased 16% YoY, reflecting improved efficiency.

The balance sheet remains strong. As of December 31, 2025, Zhihu held RMB 4.5B in cash and equivalents, down from RMB 4.9B a year earlier due to share repurchases and operating losses. With a market cap of $269M versus $636M in cash, the company trades at 0.42x its cash value. This provides a floor on valuation and gives management flexibility to invest through the transition. During 2025, Zhihu repurchased 31.1M Class A shares for $66.5M, and the Board approved a new program authorizing up to 10% of issued shares.

Outlook, Management Guidance, and Execution Risk

Management's 2026 guidance frames a strategic choice: invest in AI-driven opportunities rather than simply maximizing short-term profitability. This signals confidence that Zhihu's AI potential is larger than current scale reflects, but also introduces execution risk as the company balances new investment against maintaining its newly achieved profitability.

For marketing services, management aims to drive continued recovery and sustainable long-term growth in 2026, building on the Q4 inflection. They expect each quarter's revenue to remain above the Q3 2025 baseline, supported by a healthier commercial ecosystem and AI-accelerated product upgrades. This guidance appears achievable given the sequential momentum, but risks remain if the ad market remains soft or if AI search further disintermediates content consumption.

The paid membership business is in a transition period with ongoing experiments to improve retention and ARPU. Management is willing to sacrifice subscriber count for quality, which could mean continued revenue pressure in H1 2026 before stabilization. The critical variable is whether improved renewal rates and premium content formats like voice live streaming can offset the smaller user base.

The most ambitious element of the 2026 plan involves scaling AI-driven commercialization. Management is pursuing a dual-track approach: enhancing core community efficiency with AI while developing new revenue streams in AI-enabled content adaptation and data services. They explicitly state these initiatives will focus on visible ROI potential and a clear path to healthy cash flow. This disciplined approach contrasts with the growth-at-all-costs mentality of prior years, but the timeline to material revenue remains uncertain.

Key assumptions underpinning the outlook include continued AI adoption in China, where the government's "AI+" strategy prioritizes integrating AI across the economy. Zhihu's professional AI-related content grew 46% year-over-year, and its high-quality content is increasingly cited by large language models. This positions the company to benefit from the AI wave, but also exposes it to competition from better-funded players like Baidu and ByteDance.

Risks and Asymmetries

The central risk is that Zhihu's AI transformation arrives too late to offset core business decline. The company faces transition pains with consecutive quarterly losses in Q3 and Q4 2025 despite full-year profitability. If marketing services recovery stalls or paid membership revenue continues falling, the company could burn through its cash cushion before new AI revenue streams mature. The one-time goodwill impairment charge of RMB 126.3M in Q4 signals that prior acquisitions failed to deliver expected value.

A second material risk is competitive displacement. As users grow accustomed to querying AI directly, Zhihu's crucial entry-point traffic is being siphoned off, reflected in declining user metrics. Baidu's Ernie Bot, ByteDance's Douyin, and other platforms are integrating Q&A capabilities that could make Zhihu redundant. While Zhihu's content quality advantage is real, it may not be defensible if users never visit the platform to begin with. The company's smaller scale—12.2M paid members versus Weibo's 500M MAUs—limits its network effects.

Third, the AI data services opportunity faces execution challenges. Management believes Zhihu can provide differentiated data solutions for AI developers, but this requires building enterprise sales capabilities and competing against established data providers. The IP monetization strategy depends on AI-generated content achieving user willingness to pay, which remains nascent.

On the positive side, significant asymmetry exists if Zhihu executes successfully. The company's net cash position provides downside protection, while successful scaling of AI data services or IP monetization could unlock value. If marketing services recovery accelerates and paid membership stabilization yields ARPU growth, the combination of revenue growth and maintained cost discipline could drive profit expansion from the current 1.3% margin.

Valuation Context

Trading at $3.03 per share, Zhihu presents a unique valuation profile. With a market capitalization of $269M and an enterprise value of -$337M (net cash position), the stock trades at just 0.42x its cash and equivalents of approximately $636M. This establishes a tangible floor on valuation, effectively meaning the market is pricing the operating business at a negative value.

From a multiples perspective, traditional metrics are less meaningful given the company's marginal profitability. The price-to-book ratio of 2.97x reflects asset-light internet economics, while the absence of meaningful earnings makes P/E ratios less relevant. The price-to-sales ratio of approximately 0.69x (based on $388M TTM revenue) compares to Bilibili's 2.25x and Baidu's 2.01x, suggesting the market has priced in significant pessimism.

The gross margin of 59.94% is healthy and competitive with Weibo's 75.99% and well above Bilibili's 36.62%, indicating Zhihu maintains pricing power in its content business. However, the operating margin of -21.37% and profit margin of -7.02% reveal the core challenge: while the company can produce high-margin revenue, it has not yet achieved the scale to generate sustainable operating profits.

The balance sheet strength is the key valuation support. With a current ratio of 3.73 and quick ratio of 3.65, and essentially no debt, Zhihu has over two years of runway even if it continues burning cash at the 2025 rate. This gives management time to execute the AI transformation without facing near-term liquidity constraints. The aggressive share repurchase program, with 10% of shares authorized for buyback, signals management's conviction that the market undervalues the business.

Conclusion

Zhihu stands at a critical inflection point where strategic discipline has achieved accounting profitability but has yet to prove the business can grow sustainably. The company's first full-year non-GAAP profit in 2025 was driven by cost cuts and revenue sacrifice rather than underlying business expansion. This creates a fragile thesis: the market has priced the stock at less than half its cash value, offering downside protection, but the operating business must demonstrate it can generate sustainable returns above its cost of capital.

The central investment case hinges on whether Zhihu's AI transformation can convert its content quality advantage—validated by being cited 4x more than competitors by AI assistants—into material new revenue streams before its core business erodes further. The sequential recovery in marketing services and growth in IP monetization provide early evidence that the strategy may work, but both remain too small to offset the 23.6% revenue decline in the near term.

For investors, the critical variables to monitor are: (1) the pace and scale of AI data services revenue generation; (2) whether marketing services recovery sustains into 2026, validating the adjustment cycle completion; and (3) if paid membership ARPU growth can compensate for subscriber losses, proving the unit economics focus is working.

The risk/reward is asymmetric: limited downside given the net cash position and aggressive buybacks, but substantial upside if AI monetization scales. However, the risk of a value trap is real—if the core Q&A platform continues losing relevance and new initiatives fail to materialize, Zhihu could slowly burn cash while the market cap remains anchored below book value. The next 12-18 months will determine whether this is a turnaround story or a slow-motion liquidation of a once-valuable community.