CITIC SEC released a research report stating that next-generation large models (Gemini 3.0/GPT5) will significantly benefit complex reasoning scenarios, marking the true beginning of large-scale AI commercialization. The business model is evolving from AI project delivery to AI-enhanced existing applications and standalone AI functional applications.
Compared to the domestic market, overseas markets lead in technological progress, payment ecosystems, business models, and market potential. Overseas AI applications continue to show growth in Annual Recurring Revenue (ARR), accelerating their implementation. The report highlights the advantages of Chinese companies in product engineering, innovation speed, and iteration capabilities, making them strong contenders in the global AI application market.
Key recommendations include focusing on: 1) AI application companies with high overseas revenue share and rapid growth; 2) Highly productized companies actively preparing for global expansion; 3) Infrastructure-focused firms expanding internationally.
**AI Application Industry Trends: From Technological Breakthrough to Commercialization** The launch of next-gen models (Gemini 3.0/GPT5) will enhance complex reasoning, driving efficiency and cost reduction to accelerate AI adoption. Three phases are identified: - **Phase 1 (2023-): AI Project Delivery** – Early exploratory and non-standard projects dominated post-2023. - **Phase 2 (2025-): AI-Enhanced Existing Applications** – Cost-effectiveness improves, with companies adopting free trials or pilot AI modules. - **Phase 3 (2026-): Standalone AI Functional Applications** – Independent AI products with direct monetization (excluding API fees) are expected to flourish.
According to aiapps100, overseas AI application ARR has grown steadily, reaching nearly $40 billion in July. Agent-type applications like Manus achieved $90 million in annualized revenue, while AI-driven firms like Palantir and Applovin saw accelerated performance growth.
**Overseas Market: Vast Demand Potential** 1) **Technology Advancements**: Global large-model capabilities exceed expectations, with ByteDance and Google reporting surging token usage. AI server investments by top CSPs are projected at $162.4 billion in 2025 (YoY +94%). 2) **Payment Ecosystem**: Domestic software payment habits lag due to user behavior, market dynamics, and IP protection, unlike overseas markets. 3) **Business Models**: Since H2 2024, overseas SaaS firms shifted toward B2B Agent applications, improving user adoption and high-value scenario profitability. 4) **Market Potential**: - **C-end**: Overseas user bases are larger (e.g., U.S. monthly AI app visits: ~2.5 billion vs. China’s ~1.1 billion). - **B-end**: Overseas corporations (e.g., Fortune 500 U.S. firms) boast higher profitability and payment capacity than Chinese peers.
**Chinese AI Expansion: Strategic Imperative** 1) **Product & Engineering**: China’s STEM talent pool (500K graduates in 2025, including 77K PhDs) and proven engineering agility (e.g., Wanxing Seekee) provide a competitive edge. 2) **Success Cases**: Kingsoft WPS’s mobile-centric strategy outperformed Microsoft Office, showcasing localized innovation. AI firms may replicate this via mature product and operational strategies. 3) **Overseas Revenue**: Top AI firms in CITIC’s computer index derive up to 90% of revenue abroad, with growth rates exceeding 10x.
**Risks**: Slower-than-expected AI tech development, data/cybersecurity risks, intensified competition, and geopolitical uncertainties.
Comments