Baidu Reveals AI Commercialization "Report Card": AI Business Revenue Accounts for 43%, Transitioning from Technology Investment to Value Realization

Entering 2026, the frenzy of AI investment remains intense. According to statistics, overseas tech giants such as Google, Microsoft, Meta, and Amazon are projected to collectively invest over $630 billion this year, with capital continuously flowing into large model training, computational power stacking, and infrastructure expansion. However, market attention is shifting: when will AI truly achieve “self-sustaining revenue” and start generating real profits?

As one of the earliest domestic players to declare “All in AI,” Baidu has incorporated its answer into its financial reports. In the fiscal year 2025 Q4 and full-year financial results released on February 26, Baidu disclosed for the first time that AI business revenue accounts for 43% of Baidu’s general business income, surpassing market expectations. In Q4, the company achieved total revenue of 32.7 billion yuan. For the full year 2025, total revenue reached 129.1 billion yuan, with AI business contributing 40 billion yuan.

This first public disclosure of AI revenue share indicates that Baidu is integrating its previously R&D-focused AI strategy—centered on capability development and investment—into a clearer operational and financial framework. Historically, AI was viewed more as a strategic outlook and technological foundation. Now, it has transformed into an independently accounted, continuously trackable revenue segment. Not only has AI-related business experienced substantial growth, but it also occupies a more central position within Baidu’s overall business structure.

Through this public financial window, Baidu’s progress in AI commercialization begins to have verifiable data points. After evaluating AI monetization efficiency, investors will see that Baidu’s growth logic is undergoing profound change, and its value narrative is entering a new cycle focused on AI realization.

Full-stack Blooming, Not Just Point Explosions: Baidu’s New Growth Curve in AI

A deep analysis of Baidu’s AI growth trajectory reveals that its momentum does not stem from a single business explosion but from coordinated advancement across B-end, C-end, and autonomous driving sectors. The core of this layout lies in the leverage effect brought by technology reuse. A single R&D investment can continuously generate value across multiple scenarios, promoting synergistic growth among various business lines, while also gradually reducing marginal costs, ultimately forming a virtuous business cycle.

The latest financial report shows Baidu AI Cloud’s sustained growth, with full-year 2025 revenue up 34% year-over-year. Among these, subscription revenue from high-performance AI computing infrastructure in Q4 grew 143% year-over-year, accelerating from 128% in Q3. According to statistics on major cloud vendor bids related to large models in China for 2025, Baidu Intelligent Cloud ranked first in both the number of projects won and total contract value related to large models, maintaining its position as the top cloud vendor for two consecutive years.

In terms of market share, according to market research firm CCID Consulting, in the first half of 2025, the Chinese AI cloud full-stack service market reached 28.09 billion yuan. Baidu Intelligent Cloud, with its full-stack capabilities from AI infrastructure (AI Infra), agent infrastructure (Agent Infra), to intelligent applications, held a 40.2% market share, ranking first.

B-end markets focus on how to effectively implement technology, and full-stack capability is key. Baidu’s major advantage lies in its自主研发的昆仑芯 (Kunlun chip). In November last year, Baidu Cloud launched the new generation Kunlun chip and Tianchi supernode, announcing that Kunlun chips will be released annually over the next five years. Designed from the ground up for AI scenarios, Kunlun chips can provide large-scale, stable, high-performance computing, compatible with mainstream models and frameworks, helping enterprises deploy and apply solutions more quickly and at lower integration costs.

In today’s environment of rapid technological iteration and market volatility, enterprises need not just a single chip or model but a complete, deployable, sustainable system. For example, Baidu’s self-evolving super intelligent agent “Famous” (伐谋), launched last year for commercial use, leverages large language model reasoning and large-scale evolutionary search technology to simulate billions of years of biological evolution compressed into days or hours, discovering global optima previously unreachable by humans. It can automatically iterate based on changing conditions, providing optimal dynamic solutions. In its first month online, over 2,000 companies applied for trials across logistics, manufacturing, AI for Science (AI4S), and other fields, helping industry clients efficiently explore global optima.

In the fiercely competitive C-end market, Baidu’s AI applications generated over 10 billion yuan in revenue in 2025. Its Wenxin Assistant monthly active users reached 202 million, forming one of China’s three major AI super portals alongside Doubao and Qianwen. Leveraging model advantages and search scenarios, Wenxin Assistant has deeply reconstructed search and AI, integrating deep thinking, multimodal interaction, and full-scenario services into a versatile platform.

The true battleground in the C-end is no longer just user growth but whether users are willing to pay for tangible AI productivity, completing the loop from user scale to commercial value. Baidu Cloud’s no-code application platform, “Suidan,” upgraded to version 2.0, now supports “full-stack application with one-click generation” and “one-stop development and distribution,” enabling ordinary users to create production-level applications. In eight months, the platform has generated over 500,000 commercial applications, with daily new applications increasing by over 150%. Half of these include backend functions, covering more than 200 scenarios such as education, commerce, content creation, and enterprise services, creating over 5 billion yuan in economic and efficiency value.

As AGI exploration deepens, Physical AI is rapidly moving from labs to industry, becoming the next core battleground among global tech giants. Baidu has long been deeply involved in this area. Its autonomous driving service platform, Radish Run (萝卜快跑), is a direct application of Physical AI, opening new growth avenues. The financial report shows that Radish Run’s global autonomous driving trips in Q4 reached 3.4 million, up over 200% year-over-year, with weekly trip peaks exceeding 300,000.

Currently, Radish Run has accumulated over 300 million kilometers of autonomous driving, with more than 190 million kilometers fully autonomous. The platform is expanding globally; by February 2026, Radish Run has provided over 20 million trips worldwide, covering 26 cities including Dubai and Abu Dhabi.

On January 17, Radish Run and AutoGo officially launched full autonomous commercial operations for the public in Abu Dhabi. Residents and tourists can now summon Radish Run’s driverless vehicles via the AutoGo app. This marks Radish Run’s first public deployment of fully autonomous vehicles overseas, a significant step in China’s autonomous driving technology’s commercialization and internationalization.

From technical validation to large-scale operation, from domestic pilot to overseas deployment, Radish Run is bringing China’s intelligent mobility solutions to users in more cities worldwide.

Intensive Moves in the New Year, Baidu Demonstrates a Strong Combination

During the recent Year of the Horse Spring Festival holiday, China’s tech industry staged a fierce AI-focused kickoff battle. The Spring Festival, the most emotionally charged holiday for Chinese people, has evolved into a key stage for major companies to showcase AI capabilities. Traffic becomes the entry point, interaction amplifies engagement, and commercial conversion is the ultimate goal. Behind this tech showcase is a contest for consumer minds and business monetization.

As the opener of this Spring Festival battle, Baidu APP’s Wenxin Assistant launched the “Early Bird” campaign with AI red envelopes. From January 26 to March 12, users interacting with Wenxin Assistant within Baidu APP had a chance to share 500 million yuan in cash red envelopes—China’s earliest Spring Festival red envelope event involving AI. Unlike others, Baidu’s campaign tightly links red envelope collection with core AI capabilities. Industry analysts note that Baidu Wenxin Assistant’s advantage lies in its integration within Baidu APP, enabling seamless scene switching and leveraging natural traffic, allowing users to switch effortlessly between search and AI use. Official data shows that since the campaign started, MAU for Wenxin Assistant has quadrupled year-over-year.

Meanwhile, Baidu has also been active in capital markets. On February 5, Baidu announced that its board authorized a new share repurchase plan, allowing the company to buy back up to $5 billion worth of shares, valid until December 31, 2028. Baidu stated that with its large cash reserves and sound financial management, this proactive shareholder return measure aims to create and sustain long-term value for shareholders.

Notably, the $5 billion buyback is nearly 10% of Baidu’s current market cap—an unusually high proportion compared to other tech giants like Apple and Alphabet, whose annual buybacks typically account for 2-4% of market cap, and even Meta’s historical high of about 5%.

In this context, a buyback close to 10% of market value is a rare “unconventional move,” signaling not only an effort to boost EPS but also a message to the market: the current stock price does not fully reflect intrinsic value. From a capital efficiency perspective, such a scale of repurchase suggests at least 10% or more upside potential for the stock.

Additionally, Baidu’s board expects to announce its first dividend in 2026 to further enhance shareholder returns. Future dividend amounts, timing, and policies will be determined by the board based on the company’s financial performance, cash needs, market conditions, and other relevant considerations.

Launching dividends amid ongoing high AI investment signals confidence in future business growth and cash flow generation. It indicates that the company believes it can cover ongoing R&D and capital expenditures while also steadily releasing free cash flow.

Historically, tech companies initiating dividends often mark a transition in growth stage and can be a catalyst for valuation reappraisal. For example, Microsoft’s first dividend in 2003, followed by a “dividend + buyback” strategy, helped grow its market cap from about $250 billion to nearly $3 trillion today, becoming a benchmark for tech valuation revaluation.

Beyond shareholder returns, market attention at the start of the year also focused on Baidu’s plan to spin off Kunlun Chip for an independent listing. On January 2, Baidu announced that its non-wholly owned subsidiary, Kunlun Chip, had submitted a confidential listing application to the Hong Kong Stock Exchange on January 1, 2026, aiming for a standalone listing on the HKEX main board.

This move could further unlock the potential value of AI chips, improve market valuation transparency regarding Baidu’s technological strength and long-term prospects, and add new imagination space to Baidu’s overall valuation framework. As its business structure becomes more diversified, market valuation logic is shifting from traditional P/E ratios toward segment-based valuation, more precisely capturing the intrinsic value of each business unit.

In this context, institutional valuations for Baidu have also risen, with nearly 20 firms raising their target prices. Benchmark, for example, raised its target to $215, explicitly citing the Kunlun Chip spin-off as a key catalyst and updating its segment valuation model to include the previously unreflected value of Kunlun Chip. Benchmark considers Baidu one of the most promising stocks for upside in FY2026. J.P. Morgan also raised its December target price to $200, emphasizing that Baidu’s long-term AI investments are translating into more sustainable growth momentum.

As AI becomes embedded in financial statements, Baidu’s valuation system is also being reshaped. It is no longer just an internet platform relying on search portals but a tech giant driven by AI, achieving large-scale monetization across diverse scenarios. Behind this is Baidu’s full-stack AI capability, closing the “technology-business” loop. AI is shifting from a costly expense center to a profit-generating core. When technological investments translate into real income paid by enterprises and users—rather than relying on traffic dividends or short-term subsidies—Baidu’s pricing logic shifts from “storytelling” to “financials.”

2026 may well be this watershed year. Standing at this new temporal coordinate, Baidu’s story is no longer about reversal or revival but about how a long-term-oriented company redefines its position at a pivotal moment in history.