Artificial Intelligence Patents in Asia: The Race to Dominate Tomorrow's Technology Landscape

New data from the World Intellectual Property Organization reveals that generative AI patent publications nearly tripled in two years, with Asian nations — led by China and Japan — positioning themselves at the forefront of one of the fastest-moving innovation races ever recorded. Patent trends expose a deepening strategic competition for intellectual property dominance in generative AI across the Asia-Pacific region.

A single set of numbers from the World Intellectual Property Organization (WIPO) has sharpened the conversation about who will own the intellectual architecture of the generative AI era. Published generative AI patent families — a widely used measure of new inventions — rose from approximately 14,000 in 2023 to over 37,800 in 2025. More than 56,000 new generative AI (GenAI) patent families were published in 2024 and 2025 combined, exceeding the cumulative output of the entire period from 2014 to 2023. For Asia's technology powers, those figures are not merely a statistical milestone. They are a strategic signal.

China's Commanding Lead

No country dominates the GenAI patent landscape more visibly than China. China-based inventors published over 43,000 patent families in 2024 and 2025 alone, making it the world's largest source of GenAI patent publications by a substantial margin. Six of the world's top ten GenAI patent applicants are based in China, including Tencent Holdings, Ping An Insurance Group and Baidu — names that span social media, insurance and internet search, reflecting how broadly Chinese industry has moved to embed generative AI across sectors.

China's compound annual growth rate (CAGR) for GenAI patent families between 2023 and 2025 stands at 64 per cent. That figure, while strong in absolute terms, is actually the most modest growth rate among the major inventor nations covered in WIPO's Technology SPARK report — a consequence of China having already built a large base in prior years. The country's lead is structural as much as it is momentum-driven.

The breadth of Chinese applicants is telling. Ping An Insurance Group's presence among the global top ten illustrates a pattern WIPO's data makes explicit: GenAI patents are no longer filed exclusively by technology companies, but increasingly by large enterprises in finance, telecommunications, infrastructure and other digital services.

Japan's Surge: From Fourth to Third

If China's dominance is the headline, Japan's acceleration is the story underneath it. Japan recorded the highest relative growth rate of any top inventor location — a 210 per cent CAGR between 2023 and 2025 — rising from fourth to third place globally.

Japan's SoftBank is the world's leading GenAI patent applicant, publishing nearly 3,000 patent families during 2024 and 2025. SoftBank's position at the very top of the global rankings — ahead of all American and Chinese rivals — underscores how aggressively the conglomerate has repositioned itself around artificial intelligence after its well-documented travails with earlier technology bets.

The Absent Voices: India, South Korea, Taiwan

WIPO's Technology SPARK report does not list India, South Korea or Taiwan among the top inventor locations covered in its findings, which presents its own kind of data point. South Korea and Taiwan, home to semiconductor giants whose hardware underpins virtually all AI model training globally, do not feature among the named leaders in GenAI patent publications. India, despite its vast software engineering workforce and growing domestic AI ambitions, similarly does not appear in the top tier of the WIPO rankings.

This gap between hardware influence and patent-publication leadership is a persistent structural feature of the regional innovation landscape. Designing and manufacturing the chips that power AI systems does not automatically translate into ownership of the algorithms, architectures and application methods that GenAI patents typically protect.

What the Patent Race Reveals

GenAI now accounts for 8.7 per cent of all AI-related patent family publications, up from 6.1 per cent in 2023 and 4.2 per cent in 2017. Seven of the world's top ten inventor locations published more GenAI patent families in the last two years than in the entire preceding decade. The pace of change is, by any historical comparison, extraordinary.

For policymakers across Asia, the patent data functions as a kind of early intelligence. They signal where capital is flowing, which applications industries consider commercially viable, and which nations are building the IP portfolios that will shape licensing terms, trade negotiations and standards-setting for years ahead.

The WIPO findings also complicate the narrative that the generative AI race is simply a contest between American frontier labs and Chinese state-backed challengers. Japan's SoftBank sits at the top of the global applicant rankings. Germany has overtaken the United Kingdom as Europe's leading GenAI inventor location. Canada and Switzerland recorded triple-digit CAGRs. The field is wider, and the competition more genuinely global, than the loudest voices in the debate typically acknowledge.

Remaining Questions

Several dimensions of the Asia GenAI patent story remain unresolved by the available data. The WIPO Technology SPARK report does not specify where South Korea, Taiwan or India rank among global inventor locations, leaving their precise competitive standing unclear. The report also does not break down the quality or commercial impact of patent families by country — a high volume of filings does not necessarily indicate a high proportion of commercially significant inventions. Whether China's 64 per cent CAGR represents a natural plateauing from a large base or a genuine deceleration in innovation momentum is not addressed. The extent to which SoftBank's near-3,000 patent families reflect organic internal R&D versus activity connected to its investment portfolio companies is similarly unverified. Finally, how regional patent filings translate into enforceable IP advantages in cross-border AI product markets — particularly given differing national patent examination standards — remains an area requiring further reporting.