China's 2 top AI rivals fuel merger talk but competition is key

A hypothetical merger between two of China’s leading AI firms, DeepSeek and Moonshot AI, is capturing industry attention after both released major open-source models within a four-day span in late April. While a combined entity could create a full-stack Chinese AI powerhouse with a valuation exceeding $38 billion to rival Western leaders, the proposal remains a thought experiment, with intense competition, founder independence, and regulatory hurdles standing in the way.

The discussion was amplified by a social media post questioning the impact of a merger on competitors like OpenAI and Anthropic, according to a speculative analysis from锦缎. The logic for a combination is compelling: DeepSeek’s focus on model efficiency, demonstrated by its V4 model which cut single-token inference costs by 73%, is highly complementary to Moonshot’s Kimi model, which excels at complex agentic tasks and multimodal understanding.

“In a disruptive technology landscape, a closed-source moat is transient,” DeepSeek founder Liang Wenfeng has stated, explaining his open-source logic. This philosophy of leveraging the community for faster iteration is shared by Moonshot founder Yang Zhilin, who believes finding a more optimal data compression method is the key to unlocking higher intelligence under compute constraints.

The speculation comes as both companies are aggressively courting capital and talent. Moonshot AI, maker of the Kimi chatbot, saw its valuation jump to $18 billion after three funding rounds between late 2025 and early 2026. DeepSeek, previously funded internally by quantitative trading firm High-Flyer Quant, began seeking external capital in April 2026 with a reported valuation target over $20 billion. A merger would create an entity with a combined valuation that, while significant, would still be less than one-twentieth of Anthropic's estimated $380 billion private market valuation.

A Case for Consolidation

A merger would address the primary bottlenecks facing both firms: access to computing power, revenue scale, and global branding. Both companies have acknowledged that a shortage of high-end GPUs is their main growth constraint. A consolidated entity would have greater bargaining power with chip suppliers like Nvidia and Huawei, and could streamline efforts to adapt models for domestic hardware like Huawei’s Ascend 950, a stated goal for both firms.

On the commercial front, a unified pricing strategy would prevent a race-to-the-bottom that could erode margins. DeepSeek is known for its aggressive low-cost strategy, with its V4 Pro API priced at just 12 yuan per million tokens, while Kimi recently raised its API prices by 58% after its K2.6 model topped performance charts. Together, they could establish a more stable value benchmark for Chinese open-source models on the global stage, where clients like Cloudflare and Perplexity have already integrated Kimi.

The Independence Barrier

Despite the synergies, a merger is highly improbable. The primary obstacle is the fierce independence of the founders, both of whom have strong technical visions and have successfully guided their teams through intense development cycles. Neither Liang Wenfeng, who bootstrapped DeepSeek with his own capital, nor Yang Zhilin, who steered Kimi through a dramatic turnaround, is seen as likely to accept a secondary role in a combined company.

Furthermore, the complex web of shareholder interests would be difficult to untangle. Tech giants like Alibaba Group Holding and Tencent Holdings have invested in one or both firms, a strategy of hedging bets across the competitive landscape rather than fostering a single champion. Forcing a merger would reduce their strategic flexibility in the AI sector.

Finally, any such deal would almost certainly face intense antitrust scrutiny from Chinese regulators. A merged DeepSeek-Kimi would hold a dominant position in the country's open-source AI model market, potentially marginalizing other startups. The current competitive dynamic, which forces rapid innovation and technological diffusion, is seen as a healthier state for the industry than consolidation under a single giant. The rivalry between DeepSeek and Kimi, much like that of OpenAI and Anthropic in the U.S., serves as the most effective catalyst for progress.

This article is for informational purposes only and does not constitute investment advice.