The company, sometimes called the Quora of China, is banking on growth from a new AI tool to improve expert-supplied answers to user questions
Key Takeaways:
Zhihu reported its first-ever profit in the fourth quarter, though its revenue fell by 25% as its marketing and education services both reported big declines
The knowledge-sharing platform may have found a winner in its new Zhida.ai platform, which uses AI to enhance its core Q&A services
After years of searching since its founding in 2010, knowledge-sharing platform Zhihu Inc. (2390.HK; ZH.US) finally found an answer to its profitability challenge in last year’s fourth quarter, to the delight of investors who waited more than a decade for the breakthrough.
The profit came as Zhihu became more selective about its users, forgoing less profitable ones to focus on bigger spenders. The company also held out big potential for AI to take its knowledge-sharing content to the next level, saying it is using the technology to complement answers traditionally provided by experts to user questions.
Zhihu reported a profit of 86.4 million for the quarter, reversing a net loss of 103.1 million a year earlier, despite a relatively steep 25% year-over-year revenue decline to 859.2 million yuan over that period, according to its landmark report. Its move to the black owed in part to a gross margin that improved by 3.8 percentage points to 62.9%.
The revenue plunge came as Zhihu’s monthly average users fell by 18% year-on-year to 81.4 million from 99 million a year earlier, reflecting a more common theme among Chinese companies lately to focus on quality users over simply quantity. Zhihu remained in the red for the year with an annual loss of 169 million yuan, though that was a big improvement on the 840 million yuan loss a year earlier.
“The fourth quarter of 2024 was a true milestone for us, an inflection point that underscores our strategic focus and strong execution,” said founder and Chairman Zhou Yuan on the company’s earnings call.
Zhihu’s CFO Wang Han, who took the job in February 2022, may have been an important engineer behind the move to the black. Among other things, the company managed to reduce its cost of revenue by 31.5% in the fourth quarter year-on-year, outpacing the 25% revenue decline. But he also hinted the move to the black might be temporary, saying on the call that Zhihu this year might “choose to maintain a slightly overall loss temporarily to continue pursuing and capturing the potential opportunities arising from deeper integration between Zhihu and AI.”
Zhihu breaks its business down into three main areas: its paid membership and core marketing services that cater to traditional advertisers; and its newer vocational training education services. Marketing services took a big hit in the fourth quarter, plunging by about a third year-on-year to 316 million yuan, from 465 million yuan a year earlier. That decline propelled paid subscribers to its top revenue source, though even that segment fell 8% year-on-year to 420 million. Vocational training fared the worst, tumbling by about half to 84 million yuan.
Zhihu explained that the drop in market services revenue was part of a “deliberate” choice to forego some “peripheral” users and “sunsetting lower quality revenue streams”.
Zhihu has frequently been compared to Quora, a U.S.-based Q&A website. It’s trying to take that knowledge-sharing core function to the next level with generative AI, aiming to transform its user generated content platform by delivering top-tier, verifiable data for professionals. It says such positioning differentiates it from its peers and can help attract advertisers who are always seeking the latest targeted way to reach their audience.
Tough to monetize
CFO Wang explained that Zhihu’s highly knowledgeable users and in-depth content were historically difficult to monetize in a traditional internet paradigm that prioritized high traffic and lengthy user engagement more often seen in entertainment-focused platforms.
“However, in the AI era, this has been completely reversed. And now the AI large language models prioritize high-quality answers generated by top-tier experts,” he said, maintaining that advertisers are increasingly focused on the quality of content and whether it turns up in AI generative engine optimization.
Zhihu started using generative AI in its business last June, when it released its first AI-based search platform, called Zhida.ai or “direct answer,” which has already attracted 10 million MAUs. Zhida.ai has its own content based on input from a Zhihu community of experts together with data sources, allowing users to look for answers to their questions from individual experts, as well as data-backed sources.
As of the end of the fourth quarter, Zhihu had 71.7 million cumulative content creators or users who provide content by answering questions, and 874.6 million “pieces” of content.
Chairman Zhou said Zhihu’s number of AI-focused content creators grew 50% year-over-year in the fourth quarter and included “many leading entrepreneurs and industry pioneers who engage in multi-dimensional, professional and in-depth discussions.” Income-generating creators increased by 22% year-over-year, he added, with Zhihu prioritizing onboarding of high-tech content creators, including many podcasters.
Zhihu’s shares jumped by 13% on Feb. 11 after it announced Zhida.ai would integrate DeepSeek’s R1 model. Zhihu’s share price has remained elevated since then, though it’s still far from its all-time highs since its Nasdaq IPO in March 2021 and second listing in Hong Kong a year later. At its current level, the stock is still down by more than half from its Nasdaq IPO price, though the stock is up about 15% since the start of the year.
Zhou painted Zhida.ai in glowing terms on the earnings call, hoping to sell investors and analysts on the technology’s potential to truly transform the company.
“We have implemented completely new ways for AI to integrate with our community,” he said. “We will continue deepening this integration to amplify the impact of our expert network. And second, AI will empower many of our community workflows, boosting the efficiency of high-quality content creation and distribution. We believe this is a new beginning for us.”
CFO Wang conceded that even though many people are using AI, the technology still has some credibility issues due to “hallucination problems,” making real-world experts still important. He added that Zhihu is unique from other entertainment-style platforms in China as the only one with a focus on experts and professionals.
Seven analysts surveyed by Yahoo Finance share Zhou’s and Wang’s enthusiasm, with six rating Zhihu a “buy” or “strong buy” despite their expectation that it will return to the red this year. The stock looks undervalued by most standards, with a price-to-sales (P/S) ratio of 0.79, well below the 1.86 for short-video and livestreaming site Kuaishou (1024.HK), and 1.75 for search leader Baidu (BIDU.US; 9998.HK), which also counts advertising as its main revenue source. But those rivals are all profitable, and Zhihu may also need to show it can operate profitably on a consistent basis before investors value it higher.
© 2025 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.
