AI competition enters a new stage of capital formation capability competition

■Su Xianggao

Since the beginning of this year, the global AI (artificial intelligence) industry has experienced a strong wave of financing. Recently, DeepSeek (Deep Exploration) plans to raise up to 50 billion RMB; leading companies like Zhipu AI and MiniMax have also successively gone to Hong Kong for financing within the year. On the overseas front, OpenAI has completed a new round of committed capital financing worth $122 billion, with a valuation reaching $852 billion; Anthropic is also planning a billion-dollar-level financing to expand computing power.

The reason these financing events have attracted widespread market attention is not only because of their huge amounts but also because the resource structure of AI industry competition is changing: more capital is flowing into infrastructure for computing power, model training, engineering capabilities, and scene ecosystem construction, directly affecting the speed of technological iteration, cost control, and commercialization. In the past, market observation of AI companies mainly focused on model capabilities, algorithm efficiency, and parameter scale; now, capital, computing power, talent, data, and scene ecology together form a new competitive system. From this perspective, global AI competition is shifting from a single-point technological contest to a new stage of competing in capital formation ability.

In the context of AI, capital formation ability refers to a company’s comprehensive ability to effectively convert financial resources into R&D investment, computing infrastructure, application scenarios, and commercial returns. Capital cannot replace the “from zero to one” innovation in technology but can significantly influence the speed, breadth, and resilience of technological iteration. If breakthroughs in cutting-edge technology open the door for capital, then capital formation ability determines whether technological dividends can be continuously amplified. It is reshaping the global AI competition landscape from three dimensions: competitive patterns, transformation pathways, and market pricing.

First, AI competition is shifting from “single-point breakthroughs” to “system competition.”

From the perspective of competitive patterns, capital formation ability is primarily reflected in the organization and allocation of underlying infrastructure and key resources. The enhancement of large model capabilities relies on comprehensive support from computing power, data, engineering, and funding. Whether it is training for capability breakthroughs, inference for stable cost reduction, or application for large-scale deployment, all depend on infrastructure such as chips, servers, data centers, and energy supply. With frequent iterations of frontier models, AI companies not only need top-tier algorithm teams but also require high-intensity capital expenditure, complex organizational capabilities for computing resources, and cross-cycle financial resilience.

Therefore, evaluating the quality of an AI company cannot only look at short-term model performance but must also consider its long-term resource allocation and sustained investment ability. Model capability is an external manifestation; computing power and engineering capabilities are the underlying support; and capital formation ability determines whether these elements can be integrated efficiently. Without continuous capital investment, technological leadership is hard to translate into scale advantages; without a true technological core, even accumulated funds will eventually fall into inefficiency.

Second, AI competition is shifting from “fundraising battles” to “transformation battles.”

From the perspective of transformation pathways, capital formation ability is more deeply reflected in the efficiency of converting funds into core barriers. The accelerated financing of leading domestic AI companies indicates that the industry is moving from early-stage technological exploration into a new phase of scaled investment, commercialization validation, and ecological competition. In this stage, raising funds is just the starting point; the real test lies in whether companies can allocate external funds and internal resources to key links and continuously transform them into competitive barriers.

This transformation involves at least four dimensions: converting into sustained R&D capacity to ensure model iteration speed; converting into organizational computing power to expand cost control boundaries; converting into scene deployment ability to embed technology into real industries; and converting into ecological collaboration to drive resonance across hardware and software supply chains. Ultimately, the scale of financing is only one indicator of a company’s potential; whether it can further solidify into product experience, customer value, and business returns is the core variable determining long-term success.

Achieving these transformations requires not only internal strength-building by companies but also effective adaptation of the capital ecosystem. The growth of China’s AI industry cannot rely solely on short-term capital infusion but needs long-term capital, specialized capital, and industrial capital to work together. Long-term capital can tolerate longer R&D cycles; specialized capital can identify technological barriers; industrial capital can provide real-world scenarios. Only when various types of capital form a synergy can the “heat” of financing truly translate into “industry thickness.”

Third, AI competition is shifting from “concept valuation” to “ability pricing.”

From the market pricing perspective, capital formation ability will reshape the valuation system and competitive ranking of AI companies. In the early stages of the AI industry, the market was more inclined to price based on model releases, parameter scales, user growth, and financing enthusiasm; but as the industry enters a high-investment, long-cycle phase, relying solely on conceptual narratives can no longer support long-term valuation. In the future, markets will pay more attention to whether companies can generate stable operational results through sustained investment, including revenue growth quality, customer stickiness, gross profit improvement, cash flow status, and the sustainability of profit models.

This means that AI competition is not only a contest of technological leadership but also a contest of market recognition and resource allocation ability. Companies with long-term value will receive stronger positive feedback in valuation, financing, talent gathering, and ecological cooperation; those lacking a complete business loop and relying on short-term hype may be re-priced as valuation normalizes. The shift from “concept valuation” to “ability pricing” will drive resources toward more certain and sustainable enterprises, further reshaping the AI industry’s competitive landscape.

In summary, AI competition has entered a new stage of competing in capital formation ability, which is fundamentally a comprehensive test of technological innovation, industrial upgrading, and financial efficiency. Capital advantages can only become a long-term competitive edge if they are transformed into technological accumulation, engineering capabilities, and industrial efficiency. Those who can embed capital into technology, convert technology into industrial efficiency, and realize efficiency into business value will be able to navigate industry cycles and gain true initiative in the global AI race.

(Source: Securities Daily)