Kunlun Tech Co.,Ltd. (300418.SZ) has released a financial report that leaves the market with mixed feelings, submitted just before the end of the earnings season.

According to the financial data, the company achieved operating revenue of 81.98 billion yuan in 2025, with a net loss attributable to shareholders of 15.93 billion yuan and an adjusted net loss attributable to shareholders of 17.25 billion yuan.

This marks the second consecutive year of substantial losses for Kunlun Tech. The company's significant losses have continued under its AI strategy since 2024.

In the first quarter of 2026, this trend of losses was not effectively curbed; instead, the scale of losses expanded year-over-year, with a single-quarter net loss attributable to shareholders of -8.87 billion yuan.

Once regarded by the capital market as the "King of Globalization" and a "Gaming Giant," Kunlun Tech's fate seems to have become tied to a high-stakes gamble since Chairman Fang Han proclaimed the company's commitment to "All in AGI and AIGC" and publicly declared its ambition to become an "AI Realist."

Now, halfway through this strategic shift, with resources nearly depleted, the market is left with an awkward situation: revenue growth without profit growth, mounting pressure on cash flow, and core businesses yet to achieve operational viability.

**Performance "Bleeding" and Profit Black Hole**

On the surface, Kunlun Tech appears to be on a fast track of rapid growth.

In 2025, the company's operating revenue reached 81.98 billion yuan, a year-over-year increase of 44.78%. Revenue for the first quarter of 2026 reached 25.7 billion yuan, up 45.69% year-over-year. However, beneath the impressive revenue growth lies a concerning "bleeding" reality.

The consecutive two-year losses are not accidental but a direct reflection of severe imbalances in the company's financial structure.

In 2025, Kunlun Tech's net loss attributable to shareholders remained largely unchanged from the previous year. This indicates that despite a nearly 45% increase in revenue, the company has not achieved any substantial profit recovery.

More worryingly, cash flow has deteriorated.

In 2025, net cash flow from operating activities was -7.36 billion yuan, a significant decline of 354.09% compared to the positive 2.9 billion yuan in 2024. This suggests that the company's core business is not generating sufficient cash flow to sustain itself but is instead consuming substantial existing funds to support its ambitious expansion plans.

This "bleeding" intensified in the first quarter of 2026. Although revenue exceeded expectations, losses expanded by 15.34% year-over-year. Apart from losses from changes in fair value due to factors such as geopolitical conflicts affecting investment operations, the "burn rate" of the core business remains the fundamental cause of the losses.

**Strategic "Missteps"**

**Flaws in Aggressive Expansion**

Kunlun Tech's predicament largely stems from its aggressive and fragmented strategic direction.

Under the high-profile advocacy of executives like Fang Han, the company has attempted to build a comprehensive AI ecosystem covering "computing power—models—applications." However, this seemingly grand "4+3" AGI strategy, which aims to support three AI-native platforms with four foundational models, has exposed critical strategic flaws in execution.

As a mid-sized internet company, Kunlun Tech's resources are inherently less abundant than those of internet giants. Yet, the company has simultaneously charged into multiple sectors, including AI video, AI music, AI social networking, AI gaming, and AI office agents. This "scattershot" approach has severely diluted resources, preventing the company from concentrating its efforts to establish dominance in any single niche.

Currently, the market sees that while Kunlun Tech has launched numerous products, aside from its short-video platform gaining some traction due to industry trends, other products like the AI social app Linky have monthly active users hovering in the millions, failing to create a true "killer app" with tens of millions of monthly active users.

To support its extensive AI application matrix, Kunlun Tech has adopted an extremely aggressive "burn money for market share" strategy. Financial reports show that in 2025, the company's sales expenses reached 41.82 billion yuan, an increase of 81.53% year-over-year, with a sales expense ratio reaching an astonishing 51%. This means that for every 1 yuan of revenue earned, the company spends 0.5 yuan on promotion and traffic acquisition.

This growth, reliant on massive traffic purchasing, is essentially an inefficient cycle. Once spending is reduced or halted, user numbers and revenue are likely to decline rapidly. High customer acquisition costs severely erode profit margins, meaning that despite the appearance of high revenue growth, the company is effectively "working for" traffic platforms.

**Industry Pressure**

**"Pseudo-Moat" in a Tight Spot**

Fang Han has repeatedly emphasized Kunlun Tech's global leadership in AI technology, citing rankings of SkyReels-V4 and Mureka V8 on certain evaluation charts. However, this technological edge has not translated into commercial advantages.

In the foundational large model sector, Kunlun Tech's "Tiangong" large model remains in the second or even third tier domestically, with a noticeable gap compared to products from giants like ByteDance, Tencent, and Alibaba. In terms of data, computing power, and top-tier talent, Kunlun Tech cannot compete head-to-head with these giants.

Industry observers argue that the so-called "technological moat" appears unstable without independent underlying computing power and top-tier algorithmic talent. The brief tenure and subsequent departure of computer vision and machine learning expert Yan Shuicheng seemed to reflect a lack of confidence in the company's underlying technological foundation.

Kunlun Tech's missteps are also closely tied to its awkward position in the industry. In the current AI landscape, the company faces dual pressure: "suppression from giants ahead and pursuit from startups behind."

Upstream, giants like Google, Microsoft, and domestic players such as ByteDance, Alibaba, and Tencent, with their vast user data, substantial capital reserves, and top-tier computing clusters, can quickly replicate Kunlun Tech's AI application models while also leveraging iterative advancements in underlying technology to gain a competitive edge. Downstream, numerous vertical AI startups like Cursor and Kimi are agile and quick to make decisions, capturing niche market opportunities with focused breakthroughs.

Kunlun Tech, however, has lost the agility of a startup while lacking the resource depth of a giant. This "middle-ground dilemma" leaves the company struggling strategically: it must maintain high R&D investments to catch up with giants while also paying substantial marketing expenses to fend off encroachment from startups, ultimately putting immense pressure on its financial performance.

**Commercialization Challenges**

**Company Weaknesses and Hidden Concerns**

If there is any bright spot in the financial report, it is that Kunlun Tech's AI short-video business generated revenue of 16.17 billion yuan in 2025.

However, behind this seemingly impressive performance lies more the overall boom in the overseas short-video market rather than exclusive barriers created by AI technology. Moreover, this business heavily relies on traffic purchasing, and the cost reductions promised by AI have not fully materialized in profits, offset instead by high marketing expenses. Other AI businesses, such as music and social networking, remain small in revenue scale and cannot shoulder the burden.

Additionally, Kunlun Tech has become overly reliant on overseas markets in recent years.

In 2025, overseas revenue accounted for 94.2% of Kunlun Tech's total revenue. While this reflects its global capabilities, it also means the company lacks a buffer in the domestic market. Should overseas markets face tightened regulations, geopolitical friction, or intensified competition, the company would have limited means to hedge risks.

Furthermore, a critical weakness for Kunlun Tech is the decline of its traditional businesses and diminishing cash-generating capabilities. Once the company's "cash cow," the gaming business has continued to shrink, with its revenue share falling to less than 8% in 2024. While businesses like the Opera browser remain stable, their growth potential is limited. The decline of old businesses has outpaced the cash-generating ability of new ones, forcing the company to rely on financing or existing funds to sustain its AI gamble.

Behind all these performance metrics, issues with executive resolve and execution are evident. Fang Han's high-profile statements starkly contrast with the company's actual performance. Despite two consecutive years of massive losses and deteriorating cash flow, the company persists in a full-frontal assault, lacking the flexibility to adjust its strategy based on market feedback. This "AI for AI's sake" obsession may be jeopardizing the company's future.

At present, Kunlun Tech's "All in AI" strategy is undoubtedly a high-stakes gamble. Two consecutive years of losses, worsening cash flow, and an unproven business model all serve as warnings. Grand strategy cannot mask inefficient execution, and technological vision cannot replace commercial logic. For Kunlun Tech, 2026 will be a critical year. Otherwise, this costly AI gamble may end up leaving nothing but chaos in its wake.