China Holds 60% of Global AI Patents
TrendForce, citing a government report presented at an August 14 press briefing by Chinas State Council Information Office, reports that China accounts for 60% of the worlds artificial intelligence patent filings. The same coverage attributes several related metrics to Chinese authorities: average daily token consumption exceeded 30 trillion as of end of June, China has built more than 35,000 high-quality datasets totaling over 400 PB, and the cumulative transaction value of high-quality datasets reached nearly RMB 4 billion, per TrendForce's account. The report also cites research showing over 400,000 data companies in 2024 and a data industry scale of RMB 5.86 trillion, and highlights Huaweis HarmonyOS ecosystem exceeding 1.19 billion devices, according to TrendForce.
What happened
TrendForce reports, citing a government report and a press conference hosted by Chinas State Council Information Office on August 14, that China accounts for 60% of global artificial intelligence patent filings. Per TrendForce's article, which quotes Chinese officials and national research bodies, average daily token consumption in China exceeded 30 trillion as of the end of June. The same reporting attributes the construction of more than 35,000 high-quality datasets with a combined volume above 400 PB to national efforts, and lists the cumulative transaction value of high-quality datasets at nearly RMB 4 billion. TrendForce also cites research from the National Data Development Research Institute stating the number of data companies exceeded 400,000 in 2024 and that the data industry's scale reached RMB 5.86 trillion. The piece highlights Huaweis HarmonyOS ecosystem, which TrendForce reports has surpassed 1.19 billion devices.
Editorial analysis - technical context
High patent volume, while not a direct measure of deployed capability, is a visible indicator of R&D activity and intellectual-property accumulation. Industry observers often treat concentrated patent portfolios as factors that affect licensing, interoperability, and freedom-to-operate discussions across jurisdictions. Rapid reported growth in token consumption and large dataset inventories are consistent with increased model training and data-pipeline activity, although patent counts do not reveal model quality, deployment, or production-readiness.
Context and significance
Reporting frames these figures as part of broader metrics used by state bodies to illustrate progress under China's 14th Five-Year Plan. Observers tracking global AI competition will weigh patent share against other measures such as compute capacity, chip self-sufficiency, open publications, and commercial product adoption. Patent counts concentrated in one jurisdiction can create asymmetric leverage in specific technology areas, but cross-border enforcement and the relevance of individual patents vary by portfolio composition.
What to watch
Indicators to follow include independent breakdowns of the patent data by technology subfield, cross-border patent filings (for example, filings with WIPO or USPTO), third-party analyses of dataset quality and provenance, and independent measures of compute infrastructure growth. Also monitor whether subsequent reporting provides detailed patent-class mappings that clarify where Chinas patent share is largest.
Scoring Rationale
The reported **60%** patent share is a notable indicator of concentrated IP activity and data-scale growth in China, which matters to practitioners tracking global capacity and licensing risk. The story lacks direct technical benchmarks or independent validation, and some source material dates to 2025, reducing immediacy.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
