Patent Friction
What it costs, how long it takes, whether a lone inventor can get in, and whose applicants it serves. Measured the same way at five patent offices, expressed as one signed number: how far the United States sits from its peers.
Four-pillar composite · median parity
Zero is parity with the peer-office median. Negative means the US is the worse place to be a filer.
Drop the weakest pillar and the headline moves to -49.14 rather than reversing. How a number becomes a score, and what the figure looks like under three other parity rules, is set out in the methodology.
Sources: WIPO World IP Indicators 2025 · IP5 Statistics Report · USPTO, EPO, CNIPA, JPO and KIPO fee schedules · USPTO PatentsView and PatEx.
Data vintages: fee schedules priced 2026 · pendency and foreign share 2024 · independents 2020. Index published August 2026.
The standings
Before the pillars, the table. Each column is one pillar measured identically at all five offices; the rank beside each value is that office's position on the pillar alone. Rows are ordered by median rank. The United States is first in nothing, and bottom-two on all three pillars the offices all publish.
The Independents pillar is missing here because only three of the five offices publish it; it enters the composite but not this table. Median rank orders the rows and nothing else: the headline number remains the four-pillar composite above.
Sources: WIPO World IP Indicators 2025 · IP5 Statistics Report · USPTO, EPO, CNIPA, JPO and KIPO fee schedules.
The four pillars
One quantity per pillar, scored 0 to 100 against the offices that publish it. A score is a position among those offices, not an absolute standard. Definitions and caveats, pillar by pillar, are in the methodology.
Second-dearest of five
Official fees to obtain and hold one representative patent to month 120, in USD. Attorney and translation fees excluded. 5 of 5 offices · lower is better · Fees priced 2026.
Statutory fees only, the one cost component all five offices set in law and publish.
Slowest of five
Average pendency to a final office decision, in months, 2024, all five offices from one publisher. 5 of 5 offices · lower is better · 2024.
The five offices start this clock at different events, and none publishes a harmonized figure.
Half the rate of South Korea
Whether an independent inventor gets in at all: the share of an office's own home-country grants that go to a person rather than a company. 3 of 5 offices · higher is better · 2020.
Only three of the five offices publish this, so parity here is South Korea and China.
Read the full storyA majority of US grants go abroad
Share of the office's own patent grants that go to applicants living outside its territory, 2024. 5 of 5 offices · lower is better · 2024.
Scored lower-is-better: a register where most grants go abroad is serving the domestic inventor less.
The clock
The Time pillar is not a bad year, it is a trend. Pendency has an annual history, and on it the United States moves the wrong way while at least one peer moves the right way. Waiting is also where the pillars start to interact: the longer the queue, the more it costs to stay in it, and the drill-down below shows who drops out first.
Months of waiting each office spares (+) or adds (−) against the peer-median line, 2021 to 2024, all five offices on one publisher's questionnaire.
It is not a staffing problem. The US office has 8,705 examiners, more than twice Europe's, and the longest wait of the five; Japan reaches a decision in 12.9 months with 1,669. The zero line is the composite's own parity rule, the median of the four peer offices, recomputed each year. The five offices start this clock at different events. Clock-by-clock detail is in the methodology.
Source: WIPO World Intellectual Property Indicators 2025, Table A62 (pendency and examiner headcounts); corroborated against the IP5 Statistics Report.
The foreign paradox
The Foreigners pillar says 55.50% of US patent grants go to applicants resident abroad, and that share has a history: it crossed half in 2008 and never came back. Concentration makes the same point from the other end. Rank the largest recipients of 2025 by headquarters and the pattern repeats at the very top of the register, only stronger.
Domestic share of grants, 1991 to 2024. China and the United States trade places across three decades.
South Korea is absent from this series. China's climb is denominator-driven: since 2016 its resident grants grew 210% while non-resident grants grew 4%.
Source: WIPO IP Statistics Data Center, patent grants by origin; cross-checked against World IP Indicators 2025, Table A61.
The 2025 top ten by utility patent grants, colored by headquarters country. Red bars are foreign-headquartered organizations; dark bars are American.
Five of the ten largest recipients are headquartered abroad, and they hold 22,947 of the top ten's 36,054 patents: 63.6%, an even higher share than the 55.50% the register shows as a whole.
Headquarters is the public domicile of the named assignee's parent, a different basis from the WIPO figure above, which counts the applicant's residence. The two measures point the same way. Counts follow the assignee names as PatentsView records them, which can split corporate families, so family level totals would only be higher.
Source: USPTO PatentsView, utility patent grants through December 31, 2025; headquarters from public company records.
Inside the US register
A comparison between offices cannot reach the split inside a single register: whether the cost and the wait land the same way on a large organization and on a person filing alone. They do not.
The Independent Inventors page follows that through the office: who gets a patent, who gives up before a decision, and who has been missing from fifty years of growth and from the AI boom.