BAT Invests $3 Billion in Kling AI: The Reality Behind the "Anti-ByteDance Alliance"

On July 2, 2026, an announcement on the Hong Kong Stock Exchange revealed that Kling AI, a large-scale video generation language model developed by Kuaishou, had completed a capital raise of up to $3 billion. This marks the largest fundraising event in the history of video generation AI globally. However, market attention has been drawn less to the scale of the funding and more to the participation of China’s three tech giants—Alibaba, Tencent, and Baidu (collectively known as BAT)—in the investor list. Some have labeled this as an “Anti-ByteDance Alliance,” speculating that BAT aims to encircle ByteDance, a narrative gaining traction within the industry.

According to reports from Huxiu, the fundraising was carried out in two phases. The first phase involved 21 investors contributing a total of $2.028 billion, followed by an additional $766 million provided by 15 investors on the same day. The agreement includes a 60-day application period during which new investors can join, but the total cap remains at $3 billion. As a result, the post-investment valuation of Kling AI stands at approximately $18 billion—a roughly 10% drop from the $20 billion speculated in May—indicating a possible cooling of enthusiasm for investments in the AI video domain within the primary market.

BAT’s Stake: Only 2.51%

An analysis of the investor composition reveals five categories: local government capital and sovereign wealth funds, multiple securities firms and banking capital, prominent private equity and venture capital firms, industrial capital, and foreign capital, including Middle Eastern funds.

Specific investment ratios for BAT are as follows: Tencent holds a 1.12% stake through two entities, investing approximately 1.363 billion yuan ($280 million). Alibaba’s subsidiary, Hangzhou Alibaba Cloud Feitian, owns 1.11% with investment matching that of Tencent. Baidu’s Beijing Baidu Netcom holds a 0.28% stake, investing about 341 million yuan ($70 million). The combined equity held by the three companies amounts to a mere 2.51%, with no board seats or veto rights on major decisions.

Notably, Kuaishou’s CEO Cheng Yixiao holds about 1%, while Kling AI’s CEO Gai Kun owns approximately 3% of the shares. The complex investment structure demonstrates that each institution independently decided to participate based on financial returns or industrial needs, rather than forming a unified front. If the full fundraising target is achieved and all stock incentives are distributed, Kuaishou will retain 68.33% equity through its wholly-owned subsidiary, maintaining absolute control. Kling AI’s financial results will continue to be consolidated into Kuaishou’s financial statements.

Diverging Strategic Intentions

While the market is eager to frame BAT’s joint investment as an “Anti-ByteDance Alliance,” an analysis of the three companies’ investment rationales reveals starkly different motivations.

For Alibaba, this represents a strategically balanced offense-and-defense tactic. AI video models require substantial computational resources, and Kling’s expansion drives sustained demand for cloud services, offering stable revenue growth for Alibaba Cloud. Additionally, Alibaba lacks its own short-video community and has limited consumer traffic distribution capabilities. By leveraging Kling’s user channels, Alibaba aims to reach more overseas creators and enterprise customers, promote the export of its model capabilities, and boost its cloud services.

Tencent’s investment logic is geared toward leveraging Kling to erode ByteDance’s market share in AI video and slow down ByteDance’s pace in building market barriers. Tencent’s self-developed Zhiying model lags behind in iteration speed, image quality, and motion control capabilities. Investing in Kling allows Tencent to rapidly acquire mature video generation capabilities, saving development time and trial-and-error costs for tools supporting its video accounts, advertisements, games, and film/TV supply chains.

Baidu’s objective is the most straightforward: addressing its weaknesses in video modeling. While the ERNIE large-scale language model excels in text domains, its progress in video multimodal capabilities has been slow. Baidu’s investment of 341 million yuan ($70 million) aims to quickly acquire video generation API capabilities at low cost, enhancing AI search, agents, and marketing platform multimodal functions, and completing the closed-loop ERNIE ecosystem.

True Strategic Encirclement Faces Challenges

Given these investment rationales, BAT may temporarily share a trench, but their overarching strategy remains one of comprehensive competition. In the cloud services market, Alibaba Cloud, Tencent Cloud, and Baidu Cloud are fierce competitors. In the generalized large-scale language model space, Tongyi Qianwen, Hunyuan, and ERNIE Bot directly confront each other. Similarly, in the agent and AI application sectors, each company is fortifying its own territory.

Optimistic market projections envision Alibaba providing computational resources, Tencent contributing traffic, and Baidu offering application scenarios, with the trio cooperating to strengthen Kling and counter ByteDance. However, forming a substantive encirclement of ByteDance is unlikely. Core assets like Alibaba Cloud’s computational power, Tencent’s social traffic, and Baidu’s search entry points are vital for each company’s survival and serve as strategic trump cards. It is improbable that these core resources would be freely shared with a third-party entity holding only about 1% equity. Cooperation will likely remain limited to business-level convenience and scenario-specific connections, far from deep ecosystem-level collaboration.

ByteDance has already established a self-reinforcing loop for creation, distribution, monetization, and data feedback through TikTok traffic, CapCut tools, and the Seedance model. In overseas markets, CapCut, with over 1 billion monthly active users, has rapidly expanded its user base following its integration with Seedance 2.0. Kling AI’s monthly active user count is only about one-fourth of ByteDance’s Seedance, and the ecosystem gap cannot be bridged by funding alone.

Kling AI’s Growth Achievements and Challenges

Kling AI has pursued a differentiation strategy, achieving notable growth. It relies on the professional creator market and focuses on overseas paid users and the entertainment industry. It has garnered professional recognition, including awards at the Cannes Lions International Festival of Creativity.

In 2025, Kling AI’s revenue reached approximately 1.04 billion yuan ($210 million), and its Q1 2026 revenue exceeded 650 million yuan ($130 million). Annual recurring revenue (ARR) surged from $100 million to about $500 million, growing fourfold in just one year.

However, Kling AI faces three practical challenges. First, its technical barriers are weak, and the loss of key personnel is undermining its competitiveness. Currently, it ranks below ByteDance’s Seedance 2.0 and Alibaba’s HappyHorse-1.0. Second, over 70% of its revenue comes from overseas markets, exposing it to ByteDance’s global competition. Customer acquisition costs are significantly higher than competitors, narrowing overseas opportunities. Third, its investment in foundational computational resources pales in comparison to ByteDance, necessitating ongoing adjustments to balance investment and outcomes.

The Essence of Separation Lies in Valuation

Reconstruction

Kuaishou’s flagship business is slowing down, with its current market capitalization hovering around $25.4 billion. Kling AI’s post-investment valuation of $18 billion accounts for over 70% of Kuaishou’s market cap. With its AI business contributing less than 1% to revenue, this structure represents a typical valuation reconstruction based on an AI narrative.

There are uncertainties underpinning Kling AI’s valuation. ARR discrepancies exceed 60%, making the valuation highly sensitive to short-term fluctuations. Applying generalized large-scale language model valuation logic to vertical AI domains involves systemic mismatches. Achieving the target of growing ARR from $300 million to $1.3 billion within 12 months is highly challenging. The agreement includes a bet clause requiring an IPO by October 2031, with investors entitled to an 8% annual minimum buyback rate, indicating prolonged digestion of the valuation.

Editorial Opinion

In the short term, Kling AI’s fundraising is set to intensify competition within China’s AI video generation market. BAT’s limited resource investments in Kling may serve as a counter-strategy against ByteDance’s Seedance. However, with low equity stakes, the incentive for BAT to deploy their core resources remains limited, raising doubts about whether this will lead to substantial competitive enhancement.

From a long-term perspective, the AI video generation sector will be defined by traffic foundations and user ecosystems. Whether Kling AI can successfully differentiate in the professional market or be absorbed into ByteDance’s consumer ecosystem will be a critical juncture. Kuaishou’s heavy valuation dependence on Kling highlights one aspect of the AI bubble that warrants close scrutiny. If Kling fails to sustain its technological edge, the gap between investor expectations and reality may lead to an adjustment phase.

The editorial team emphasizes that the BAT collaboration is not a true “Anti-ByteDance Alliance” but rather a financial investment driven by short-term individual strategies.

References

• Huxiu: Is BAT Banding Together to Counter ByteDance? — Published July 6, 2026