Backed by Jensen Huang, will AI upstream shortages persist for several years?

来咖智库

joined discussion ·

AI Landscape in Big Tech Earnings Reports: Separate Revenue Lines, Leveraging Risks, and Commercial Promises

In the just-concluded March, major Chinese internet companies successively released their latest quarterly and annual reports.

Browsing through these financial reports, 'AI' is a common keyword, appearing with the highest frequency.Through this, we can also glimpse the progress of AI businesses over the past year as well as this year's AI strategies and considerations. The biggest change lies in AI no longer being just a 'story' during earnings calls but starting to contribute solid revenue and profits.

Through these financial reports, we can also see a new catalog of AI businesses among the tech giants. Some are eager to prove themselves with AI-generated revenue, some aim to catch up by increasing investment, and others have started to reveal their future commercial goals. For instance, Baidu's AI business was listed separately for the first time, accounting for 43% of general business revenue in Q4; Tencent Cloud achieved its first full-year profitability driven by AI; Kuaishou’s AI model generated over $20 million in a single month; Alibaba Cloud’s AI-related product revenue has seen triple-digit growth for ten consecutive quarters, aiming to increase cloud and AI commercialization annual revenue to $100 billion within five years; JD.com uses AI to drive the digital and intelligent transformation of its super supply chain; Meituan's embedded AI assistant 'Little Group' is now fully available to all users, aiming to become the lifestyle service portal in the AI era...

However, the capital market’s reaction has been exceptionally complex—some stocks plummeted after the release of financial results, while others rose against the trend.Behind this lies a fundamental shift in the valuation logic of the internet industry—from 'focusing on investment' to 'focusing on realization.'

Tech Giants’ Financial Reports Highlight AI Commercialization

A common feature of the latest quarterly or annual reports from major tech firms in 2025 is the emphasis on strengthening the impact of AI on overall revenue and business growth, with AI's commercial value continuously being validated.

According to Baidu’s financial report, starting from Q4 last year, Baidu redefined its core operations as general business, including Baidu’s core AI new ventures, traditional businesses, and others.Looking further, Baidu's AI business covers smart cloud infrastructure, AI applications such as Baidu Wenku and Baidu Netdisk, future autonomous driving services, as well as native AI marketing services like digital humans and intelligent agents.

This earnings report also shows that Baidu’s total revenue reached 129.1 billion yuan in 2025, with AI business revenue reaching 40 billion yuan; in Q4 2025, total revenue was 32.7 billion yuan, of which AI business revenue accounted for 43% of Baidu’s general business income at 11.3 billion yuan. This proportion has been continuously and rapidly increasing over recent quarters, making artificial intelligence the core driver of Baidu’s overall revenue growth.

[Image/Baidu Earnings Report]

Compared to Baidu's already scaled AI business, Kuaishou’s AI is 'even a small fly has meat,' using this to showcase the commercial potential of its AI products externally.Kuaishou’s financial report shows that with the continuous advancement in upgrading KeLing AI’s model capabilities and product features, KeLing AI’s revenue reached 340 million yuan in Q4 2025, with annual revenue amounting to approximately 1.04 billion yuan. In December 2025 alone, monthly revenue exceeded 20 million US dollars, achieving an annualized revenue run rate (ARR) of 240 million US dollars. Kuaishou CEO Cheng Yixiao expressed strong confidence in achieving more than 100% year-over-year growth for KeLing revenue this year.

Not only that, Kuaishou also claims that AI technology has deeply empowered core businesses such as online marketing, live streaming, e-commerce, and content ecosystems. For example, in Q4, the amount spent on online marketing services driven by AIGC marketing materials reached 4 billion yuan; generative recommendation models and intelligent bidding models boosted domestic online marketing service revenue by about 5%. Additionally, the continuous iteration and upgrade of the end-to-end generative search architecture OneSearch drove a 3% increase in mall search orders. Capabilities such as recommendations, search, content generation, long-term interest understanding, and order AI diagnostics have been continuously implemented to help merchants enhance cross-domain operational capabilities, optimize conversion efficiency, and improve operational quality.

In the latest earnings reports, although Tencent and Alibaba’s AI businesses have not yet been separately listed, the impact of AI on their various business lines has become more evident.

For instance, in Tencent’s earnings report, nearly all sectors—games, advertising, cloud services—mention the impact of AI on their business. Particularly, driven by enterprise AI demand, Tencent Cloud achieved profitability at scale for the first time. Tencent stated that game AI effectively accelerates content production, improves user experience, and enhances marketing effectiveness; while ad pricing benefits from AI-driven precise targeting, advertisers using AI to create more ads, and the increasing share of closed-loop advertisements. Additionally, AI has also boosted the activity levels of products within the WeChat ecosystem.

For Alibaba, AI is one of the main growth engines for the group’s core businesses. In Q4 2025, Alibaba Cloud experienced robust growth of 36%, with AI-related product revenue achieving triple-digit growth for the tenth consecutive quarter. The integration of the Qwen app with various application scenarios in Alibaba's consumer ecosystem significantly drove new user growth and transaction activity. The MAU of Qwen across all platforms has surpassed 300 million, marking the entry into a scaled era for AI agents.

Currently, with the establishment of the ATH division, Alibaba has formed a complete chain from chips, cloud infrastructure, models, to applications. By integrating full-stack AI capabilities with its own commercial ecosystem, it is advancing efforts in both AItoB and AItoC directions.

Unlike Baidu, Alibaba, and Tencent whose AI businesses are mainly outward-facing, JD.com and Meituan's operations are more inward-focused and pragmatic.For example, the realization of JD.com's AI industrial value can be described as 'horizontal and vertical integration.' Horizontally, it deeply penetrates four core businesses: retail, logistics, health, and industry. Vertically, it integrates deeply into JD.com's 'super supply chain,' including R&D, production, marketing, customer service, and internal operations.

Data shows that by 2025, the annual active user base of JD.com's AI agents will exceed 150 million, with a user penetration rate surpassing 20%. It is expected to double this user base by 2026. The JoyAI large model has been deeply implemented in over 2,000 business scenarios at JD.com, achieving full-chain intelligence.

Meanwhile, Meituan is also continuously increasing its AI investment in 2025 to build an AI foundation and operational capabilities for the physical world. The company’s annual R&D investment reached 26 billion yuan, a year-on-year increase of 23%. On one hand, Meituan continues to invest in logistics and robotics-related technologies, including drones and unmanned vehicles. By the end of 2025, Meituan drones will have launched 70 routes in multiple cities both domestically and internationally, cumulatively completing over 780,000 orders. On the other hand, Meituan leverages its years of accumulated service scenarios and data advantages in the physical world to enhance user experience through AI technology innovation, helping every merchant utilize their own AI assistant. Based on its self-developed multimodal LongCat series large language model and open-source models, Meituan introduced AI assistants 'Xiao Mei' and 'Xiao Tuan' for users, enabling AI technology to be applied in real consumer scenarios.

Overall, whether it is AI business itself or AI's empowering effect on other businesses, both are driving revenue growth for major companies to varying degrees.The difference lies in each company's focus. For instance, Baidu places more emphasis on fundamental technology R&D and full-stack layout in AI, such as the applications of Wenxin Yiyang and Smart Cloud. Tencent relies on its vast social and content ecosystem to deeply integrate AI, enhancing user experience and commercial value. Alibaba's layout in the AI field is more comprehensive, covering everything from models to applications and chips. JD.com focuses on applying AI technology to supply chain management to improve efficiency. Kuaishou drives content creation and community operations through AI technology. Meituan emphasizes applying AI to the local life services sector. Based on their respective business characteristics and strengths, major companies have taken different paths in developing AI businesses.

A new round of AI roadmaps

With the release of financial reports and executives' interpretations, while major companies are presenting their AI business performance for 2025, their AI development roadmaps for this year and beyond are gradually emerging.

Baidu CEO Robin Li mentioned the recent release of the 2025 upgraded version of the Wenxin large model, along with proactive adjustments to the organizational structure to adapt to the rapidly changing market environment.For example, the large model R&D team was split according to different directions: one team continues to tackle the cutting-edge capabilities of foundational large models, maintaining technological leadership and driving continuous model iteration and optimization; another team focuses more on specific business needs and application scenarios, concentrating on cost reduction, efficiency improvement, and speeding up processes, or selecting the most suitable model for specific scenarios, with the core aim of helping enterprises better implement artificial intelligence based on actual needs.

Currently, Baidu's core AI capabilities are being reused across various scenarios, from empowering B-end industries to C-end super apps and even Robotaxi, which connects the virtual and physical worlds. This cross-scenario application is demonstrating a remarkable leverage effect, with the potential to transform into a more sustainable driver of business growth. In the future, revenue from core artificial intelligence operations is expected to account for half of Baidu's main business.

To maintain high investment in artificial intelligence and support business growth, Baidu is also exploring diversified financing methods to meet its capital needs, such as operating leases and financial leases. At the same time, it can secure low-cost bank loans, with some loans and leases carrying interest rates below 2%.

For JD.com, there is also an emphasis on continuing to increase investment in AI technology, deepening supply chain efforts, and gradually building an end-to-end leading AI e-commerce company from supply chain to consumers.Specifically, this includes: on the demand side, AI-driven search and recommendations are reshaping the shopping process; on the supply side, AI is helping procurement and sales teams improve efficiency in pricing and inventory management, significantly replacing manual work; in logistics scenarios, there is enormous potential for automated logistics fulfillment, though large-scale application will still take time; on the after-sales service side, AI customer service has already been widely implemented. Additionally, AI and new technologies have brought about a large number of innovations and new product category opportunities. JD.com will also introduce more innovative products through operations, accelerating the expansion of AI products into thousands of households by upgrading traditional products with intelligent features.

In the face of the AI technology wave, Wang Xing stated that Meituan positions this as a core strategic opportunity, focusing on the implementation of local services rather than blindly chasing concepts.He clearly expressed that Meituan is not in a rush to build a 'token factory' but instead will use AI to improve and revolutionize its core business, enhancing operational efficiency and user experience. Specifically, Meituan will firmly advance the development of its self-researched large language model LongCat, and in the future, it plans to upgrade the Meituan app into an AI-powered app, which will serve as the AI entry point for future local life demands.

Kuaishou also estimates that the group's overall capital expenditure investment will reach approximately RMB 26 billion by 2026, an increase of around RMB 11 billion compared to 2025.These investments include computing power input for the KeLing large model and other foundational models, as well as regular server procurement expenditures for offline data storage and processing, and investments in data and computing power center construction projects.

During the earnings call, Cheng Yixiao provided two clearer assessments: On one hand, KeLing’s competitiveness lies not only in the model's performance but also in how it is rapidly evolving from single-point capabilities into a more integrated, professional, and industry-demand-aligned creation system. On the other hand, Kuaishou will continue to focus on investing heavily in foundational models, Agents, and computing infrastructure in 2026, pushing for deeper AI-driven operational value release in advertising, e-commerce, and content ecosystem scenarios.

Compared to short-term goals, Tencent and Alibaba are adopting a more systematic and long-term approach to their AI strategies.Tencent President Martin Lau disclosed on the earnings call that Tencent has formulated a clear AI strategy plan. In the next two to three quarters, several key initiatives will be implemented, and measurable business progress will follow. At the large model level, over the past few months, Tencent has completed intensive team adjustments and upgrades, restructured its organizational structure and workflows to better align with AI research and development pace, while also rebuilding infrastructure for pre-training and reinforcement learning, significantly enhancing data quality. Based on these efforts, Tencent is confident in building even more intelligent large models and further accelerating overall iteration speed. Additionally, Tencent will continue to deepen its focus in video models and world models.

In terms of capital investment, Tencent is continuously increasing its commitment to the AI field. In 2025, Tencent's investment in new AI product R&D will reach 18 billion yuan, and this figure will at least double in 2026, possibly even reducing the scale of stock repurchases accordingly.

Notably, Tencent explicitly mentioned that one of the core plans for AI's next phase is to create a dedicated AI entity within WeChat. This entity will leverage the WeChat ecosystem to establish deep connections with users, integrating the entire chain of capabilities including mini-programs, social networking, and payment systems. It will not only enable natural conversations with users but also complete various complex scenario-based tasks.

Currently, Alibaba is rapidly restructuring its AI portfolio, elevating AI operations to a core business line. Alibaba CEO Wu Yongming elaborated on Alibaba’s full-stack AI strategy – an AI infrastructure layer built on chips and cloud computing; and an AI model and application layer featuring Token Hub as the main thread, consisting of large models, MaaS services, and both B2B and B2C applications, together forming comprehensive capabilities from AI Infra to applications. Meanwhile, Alibaba set a commercial goal for its AI strategy: within five years, annual revenue from cloud and AI commercialization will surge from over 100 billion yuan this year to 100 billion US dollars (approximately 690 billion yuan), with a compound annual growth rate of around 47%. This indicates that cloud and AI are expected to become Alibaba’s largest revenue drivers in five years.

From this earnings report, it is evident that a new round of AI mobilization has begun, with major companies shifting their AI investments from early model competitions to deeper challenges in computational power infrastructure, full-scenario penetration, and commercial implementation.Regardless of previous levels of investment or development pace, all players have now clearly demonstrated their determination and accelerated timelines to increase investment, truly entering a race for comprehensive AI leadership.

The harsh reality behind stock price volatility

As major companies make bold strides toward AI, the market remains concerned about the time lag and uncertainty between 'investment' and 'returns.' During this critical period of transitioning between old and new growth drivers, capital markets have responded differently, causing noticeable divergence in stock price fluctuations among major firms.

On the Hong Kong stock market, shares of Baidu, Tencent, Alibaba, and Kuaishou started to reverse downward the day after their earnings reports were released, each hitting new lows.Data shows that from the release of earnings reports until April 1, Baidu's share price hit a six-month low, Tencent and Kuaishou reached 11-month lows, and Alibaba's stock price hit a 17-month low.

Upon analysis, the stock price fluctuations of these four companies are primarily due to the uncertainty brought by reinvesting in AI, but there are also different fundamental factors at play.For instance, Baidu is facing a 'power interruption' during the transition between old and new engines. On one hand, its AI business, which represents the future, is growing rapidly, but on the other hand, its traditional advertising business, the 'cash cow,' is bleeding faster, and the new business cannot yet compensate for the profit gap caused by the decline in the old business. Tencent's issue, however, is that 'although its traditional core business is highly profitable, it suddenly decided to significantly reduce payouts to shareholders while heavily investing in AI to gamble on an uncertain future.'

Alibaba, on the other hand, has been affected by its own financial reports showing 'revenue growth without profit growth' and a sharp 71% drop in free cash flow. Moreover, the dual challenges of AI and retail in the future will make the situation even more severe. There is no sign of a profitability turning point in the short term despite massive spending. Although management explained it as 'we are spending to buy the future,' investors have become seriously divided over the return cycle and certainty of this bold gamble.

Kuaishou's situation is more unique. Unlike Tencent, it lacks a solid gaming foundation, and compared to Baidu and Alibaba, it doesn't possess technological or ecosystem advantages. Therefore, it can only bet everything on AI, even resorting to leveraging up. For example, Kuaishou may earn 20.6 billion in profits by 2025, but this year it is spending 26 billion to gamble on an uncertain AI future — this expectation of 'not earning enough to cover the burn' can only lead investors to choose to exit first.

[Image/Kuaishou stock price]

Now let’s look at Meituan. After the earnings report was released, although there was a slight drop in stock price, it remained generally stable, forming a stark contrast with the significant declines of companies like Baidu, Alibaba, and Kuaishou. The core reason lies in the fact that the market isn’t focusing on the bad news of 'a 23.4 billion loss,' but rather on the signal of certainty that 'the worst is over'—regulators have put the brakes on price wars, losses are narrowing, and new businesses and AI are building momentum for the next phase.

Among the recent stock price fluctuations of major tech firms, JD.com may have delivered the most surprising performance. Before the earnings release, JD.com's stock had been on a downward trend, hitting a new 18-month low during the period. However, after the earnings announcement, JD.com’s stock reversed its downtrend and began gradually rising, moving back toward this year’s high.

[Image/JD.com stock price]

Analyzing JD.com’s upward reversal shows that the market is less concerned with the company’s quarterly book losses and more focused on the clarity of its strategic investments—losses have peaked, core businesses remain stable, AI value is starting to materialize, and management has provided a clear path to reducing losses. Compared to Baidu, which raises concerns about an 'endless money pit for AI,' JD.com convinces the market that 'spending is for making money faster.'

The capital market's reaction to these tech giants’ AI results shows a clear divergence, reflecting investors' differing assessments of the pace at which AI value will be realized.But one harsh and clear fact is that today's AI is no longer about 'telling stories' but delivering tangible 'results,' with investors looking for AI-driven revenue, AI profits, and AI-powered growth.

Conclusion

The 2025 earnings reports will mark the transition of major companies' AI efforts from the 'preparation' phase to the 'real-world application' phase. However, AI is neither a lifeline nor a narrative boost for these giants; instead, it represents a new battleground. On this arena, former leaders may fall, and new powerhouses may rise. Each company will choose differentiated growth paths based on its own business characteristics and strengths. Whether they can deliver satisfactory results to the market will determine their standing and fate in the AI era.

$MEITUAN-W (03690.HK)$ $JD-SW (09618.HK)$ $BABA-W (09988.HK)$ $KUAISHOU-W (01024.HK)$ $BIDU-SW (09888.HK)$

26K Views