Marketdash
Back to News

AI Is Eating Zhihu's Lunch, And the Company Is Struggling to Find a New Recipe

MarketDash Editorial Team
4 days ago
China's answer to Quora just can't catch a break. After briefly turning profitable, Zhihu slipped back into the red as AI tools siphon off users who once flocked to its knowledge-sharing platform for expert answers. The problem? Why visit Zhihu when ChatGPT can give you a comprehensive answer in seconds?

Here's an uncomfortable truth about the AI revolution: it's great for some businesses and absolutely devastating for others. Zhihu Inc. (ZH), China's equivalent of Quora and a once-thriving knowledge-sharing platform, is learning this lesson the hard way.

The company reported earnings last week that tell a rather grim story. Third-quarter revenue collapsed 22% year-on-year to 660 million yuan ($120 million) from 850 million yuan. Every single revenue stream took a hit. Advertising and marketing services revenue fell 26% to 189 million yuan. Paid membership revenue dropped 16% to 385 million yuan. And "other revenue," which includes the company's vocational training services that once seemed promising, tumbled more than 30% to 83.92 million yuan from 129 million yuan.

This wasn't a one-quarter blip. Revenue declined 23.7% in the first half of the year too, and fell 14.3% for all of last year. The pattern is clear and worrying.

What makes this particularly painful is the timing. Zhihu had finally figured out how to turn a profit after years of losses. It strung together three consecutive profitable quarters. Then the third quarter arrived, and the company slipped back into the red with a net loss of 46.65 million yuan, more than quadrupling from a 10.49 million yuan loss a year earlier. The gross margin also compressed by 2.6 percentage points to 61.3%.

For the first nine months of 2025, Zhihu's revenue totaled 2.11 billion yuan, down 23% year-on-year. It did manage a net profit of 15.7 million yuan for that period, compared to a net loss of 260 million yuan in the same stretch last year. But that brief moment in the sun seems to be fading fast.

The AI Problem Nobody Wants to Talk About

Here's what's actually happening: generative AI is completely reshaping how people consume content and information. For short-video platforms, this is fantastic news. AI makes content creation ridiculously easy, democratizing creativity and fueling growth. For search engines and social platforms, AI makes information retrieval faster and more interactive, which keeps users coming back.

But for a platform like Zhihu? It's an existential threat.

Zhihu built its reputation on high-quality answers and in-depth knowledge sharing from real experts and people with genuine experience. Users would visit the platform specifically to find curated, thoughtful responses to complex questions. The problem is that AI large language models can now deliver similarly high-quality, structured answers in seconds, pulling from vast amounts of data and offering multiple perspectives. Suddenly, Zhihu's entire value proposition looks replaceable.

When users can just ask ChatGPT or any other AI assistant and get an instant, comprehensive answer, why bother navigating to Zhihu, searching through threads, and reading multiple responses? The convenience factor alone is devastating to Zhihu's business.

The user metrics tell the story clearly. By the end of the third quarter, Zhihu's average monthly subscribing members had fallen to 14.3 million, down 13.3% from 16.5 million a year earlier. The company's average monthly active users declined 21.2% year-on-year in 2024. Notably, Zhihu stopped reporting MAU figures this year, which is rarely a sign of good things happening behind the scenes.

The Pivot That Might Not Work

Founder Zhou Yuan understands the problem. In multiple interviews, he's acknowledged AI's impact on content platforms. His argument is that as knowledge becomes commoditized and universally accessible, "trust, expert networks and genuine engagement" will become the scarce resources. He envisions transforming Zhihu into a dual-purpose platform for "information + trust," leveraging its community to become the "trusted information infrastructure" of the AI age.

It's not a terrible theory. But theory and execution are very different things, especially when your revenue is declining by double digits every quarter.

The challenge is that Zhihu's payment incentives have always been more functional than social. Users pay for access to specific knowledge, not to build social identity or status. The platform's advertising context is also less defined than short-video platforms, and trying to embed commercial content within knowledge-focused text faces inherent limitations. Unless Zhou's vision of enhanced social value can translate into actual revenue streams quickly, it's hard to see how this stops the bleeding in the near term.

Zhihu has tried to fight back with AI tools of its own. Last year, it launched Zhihu Zhida, an AI-powered feature that generates answers by synthesizing content from its platform and the broader web, offering both concise and detailed responses. It also introduced a knowledge base subscription product that emphasizes traceable content sources to boost credibility.

But here's the problem: these features offer only marginal differentiation from rapidly improving general AI models. They don't answer the fundamental question: why should someone use Zhihu instead of a general-purpose AI assistant? Without a compelling answer to that question, these initiatives haven't produced significant or sustained growth in either traffic or paid users since launch.

What the Market Thinks

Investors aren't buying the turnaround story. Zhihu's Hong Kong-listed shares dropped more than 8% the day after the earnings release. Over the past six months, the stock has fallen roughly 7%, even as many other Chinese tech stocks have rallied on AI optimism. That divergence tells you everything about market sentiment.

The valuation metrics are particularly telling. Zhihu's Hong Kong-listed shares trade at a price-to-sales ratio below 1. Compare that to Kuaishou at 2.1 and Baidu at 2.2. The market has clearly decided that Zhihu is not an AI beneficiary but rather a legacy content company losing relevance during the AI transition.

Zhihu's real challenge isn't technological innovation. Plenty of companies can build AI features. The challenge is articulating and delivering unique value that can't be easily replicated in an AI-dominated world. The company needs to move beyond its original identity as a knowledge supply platform and build something new around trust and human connection that AI can't duplicate.

That's easier said than done. Until Zhihu can identify its irreplaceable niche at the intersection of AI capabilities and human insight, and then convert that positioning into sustainable revenue, the company will likely continue struggling. The brief taste of profitability earlier this year might end up being just that: a brief taste, rather than the beginning of a sustainable turnaround.

Back to News

AI Is Eating Zhihu's Lunch, And the Company Is Struggling to Find a New Recipe

MarketDash Editorial Team
4 days ago
China's answer to Quora just can't catch a break. After briefly turning profitable, Zhihu slipped back into the red as AI tools siphon off users who once flocked to its knowledge-sharing platform for expert answers. The problem? Why visit Zhihu when ChatGPT can give you a comprehensive answer in seconds?

Here's an uncomfortable truth about the AI revolution: it's great for some businesses and absolutely devastating for others. Zhihu Inc. (ZH), China's equivalent of Quora and a once-thriving knowledge-sharing platform, is learning this lesson the hard way.

The company reported earnings last week that tell a rather grim story. Third-quarter revenue collapsed 22% year-on-year to 660 million yuan ($120 million) from 850 million yuan. Every single revenue stream took a hit. Advertising and marketing services revenue fell 26% to 189 million yuan. Paid membership revenue dropped 16% to 385 million yuan. And "other revenue," which includes the company's vocational training services that once seemed promising, tumbled more than 30% to 83.92 million yuan from 129 million yuan.

This wasn't a one-quarter blip. Revenue declined 23.7% in the first half of the year too, and fell 14.3% for all of last year. The pattern is clear and worrying.

What makes this particularly painful is the timing. Zhihu had finally figured out how to turn a profit after years of losses. It strung together three consecutive profitable quarters. Then the third quarter arrived, and the company slipped back into the red with a net loss of 46.65 million yuan, more than quadrupling from a 10.49 million yuan loss a year earlier. The gross margin also compressed by 2.6 percentage points to 61.3%.

For the first nine months of 2025, Zhihu's revenue totaled 2.11 billion yuan, down 23% year-on-year. It did manage a net profit of 15.7 million yuan for that period, compared to a net loss of 260 million yuan in the same stretch last year. But that brief moment in the sun seems to be fading fast.

The AI Problem Nobody Wants to Talk About

Here's what's actually happening: generative AI is completely reshaping how people consume content and information. For short-video platforms, this is fantastic news. AI makes content creation ridiculously easy, democratizing creativity and fueling growth. For search engines and social platforms, AI makes information retrieval faster and more interactive, which keeps users coming back.

But for a platform like Zhihu? It's an existential threat.

Zhihu built its reputation on high-quality answers and in-depth knowledge sharing from real experts and people with genuine experience. Users would visit the platform specifically to find curated, thoughtful responses to complex questions. The problem is that AI large language models can now deliver similarly high-quality, structured answers in seconds, pulling from vast amounts of data and offering multiple perspectives. Suddenly, Zhihu's entire value proposition looks replaceable.

When users can just ask ChatGPT or any other AI assistant and get an instant, comprehensive answer, why bother navigating to Zhihu, searching through threads, and reading multiple responses? The convenience factor alone is devastating to Zhihu's business.

The user metrics tell the story clearly. By the end of the third quarter, Zhihu's average monthly subscribing members had fallen to 14.3 million, down 13.3% from 16.5 million a year earlier. The company's average monthly active users declined 21.2% year-on-year in 2024. Notably, Zhihu stopped reporting MAU figures this year, which is rarely a sign of good things happening behind the scenes.

The Pivot That Might Not Work

Founder Zhou Yuan understands the problem. In multiple interviews, he's acknowledged AI's impact on content platforms. His argument is that as knowledge becomes commoditized and universally accessible, "trust, expert networks and genuine engagement" will become the scarce resources. He envisions transforming Zhihu into a dual-purpose platform for "information + trust," leveraging its community to become the "trusted information infrastructure" of the AI age.

It's not a terrible theory. But theory and execution are very different things, especially when your revenue is declining by double digits every quarter.

The challenge is that Zhihu's payment incentives have always been more functional than social. Users pay for access to specific knowledge, not to build social identity or status. The platform's advertising context is also less defined than short-video platforms, and trying to embed commercial content within knowledge-focused text faces inherent limitations. Unless Zhou's vision of enhanced social value can translate into actual revenue streams quickly, it's hard to see how this stops the bleeding in the near term.

Zhihu has tried to fight back with AI tools of its own. Last year, it launched Zhihu Zhida, an AI-powered feature that generates answers by synthesizing content from its platform and the broader web, offering both concise and detailed responses. It also introduced a knowledge base subscription product that emphasizes traceable content sources to boost credibility.

But here's the problem: these features offer only marginal differentiation from rapidly improving general AI models. They don't answer the fundamental question: why should someone use Zhihu instead of a general-purpose AI assistant? Without a compelling answer to that question, these initiatives haven't produced significant or sustained growth in either traffic or paid users since launch.

What the Market Thinks

Investors aren't buying the turnaround story. Zhihu's Hong Kong-listed shares dropped more than 8% the day after the earnings release. Over the past six months, the stock has fallen roughly 7%, even as many other Chinese tech stocks have rallied on AI optimism. That divergence tells you everything about market sentiment.

The valuation metrics are particularly telling. Zhihu's Hong Kong-listed shares trade at a price-to-sales ratio below 1. Compare that to Kuaishou at 2.1 and Baidu at 2.2. The market has clearly decided that Zhihu is not an AI beneficiary but rather a legacy content company losing relevance during the AI transition.

Zhihu's real challenge isn't technological innovation. Plenty of companies can build AI features. The challenge is articulating and delivering unique value that can't be easily replicated in an AI-dominated world. The company needs to move beyond its original identity as a knowledge supply platform and build something new around trust and human connection that AI can't duplicate.

That's easier said than done. Until Zhihu can identify its irreplaceable niche at the intersection of AI capabilities and human insight, and then convert that positioning into sustainable revenue, the company will likely continue struggling. The brief taste of profitability earlier this year might end up being just that: a brief taste, rather than the beginning of a sustainable turnaround.

    AI Is Eating Zhihu's Lunch, And the Company Is Struggling to Find a New Recipe - MarketDash News