Airbus SE — Patent Filing Strategy 2014–2024
8,608 DOCDB patent families and 22,064 applications across 35 patent offices — the filing footprint of the Airbus group, its geographic centre of gravity, its technology profile, and the fields that are gaining weight in the portfolio.
Executive Summary
Between 2014 and 2024 the Airbus group started 8,608 DOCDB patent families, filed as 22,064 individual applications across 35 patent offices. The annual intake of new families runs at roughly 960 in 2014–2016 and settles at 612–885 from 2019 onwards; measured over complete filing years, 2014 to 2022, that is a compound annual change of −4.0%.
The clearest movement in the portfolio is procedural rather than technical. Comparing families whose first Airbus filing fell in 2014–2018 with those first filed in 2019–2022, the share of families carrying an EP application rises from 52.9% to 67.4%, while the share carrying a DPMA (DE) application falls from 21.9% to 13.5%. Filings at the Chinese office rise from 31.1% to 35.6% of families. Across the whole period, 68.6% of families are filed at more than one office.
Technologically the portfolio stays where an aircraft manufacturer belongs: B64D (equipment fitted in or to aircraft) and B64C (aeroplanes; helicopters) each appear in roughly 36–40% of families in both cohorts. What changes is the composition of the energy-related fringe. Families carrying hydrogen or fuel-cell classification move from 1.1% to 4.2% of the cohort, aircraft power-plant installation (B64D 27) from 5.9% to 10.4%, and computing arrangements based on learning models (G06N) from 0.4% to 1.4%. Over the same comparison, plastics shaping (B29C) falls from 10.6% to 7.9% and layered products (B32B) from 5.3% to 3.5%.
Co-filing is largely an internal matter: 1,164 families (13.5%) carry at least one application with more than one applicant, but only 304 families (3.5%) name an applicant from outside the group. Inventors of record sit almost entirely in the four home countries — France (3,626 families), Germany (3,428), the United Kingdom (1,164) and Spain (567).
The portfolio is smaller in annual intake than at the start of the period and more concentrated in how it is filed: fewer national first filings, more EP applications, and a rising share of families in energy-conversion and propulsion-installation classes. The data show a shift in filing route and in technology mix; they do not by themselves explain either.
Filing Trend 2014–2024
New DOCDB patent families by the year of their first Airbus application, and the number of individual applications filed in each year.
New families arrive at roughly 960 per year in 2014–2016, drop to 779 and 780 in 2017–2018, recover to 885 in 2019, and then run between 612 and 736 from 2020 to 2023. Measured across complete filing years only, 2014 (957) to 2022 (691), the compound annual change is −4.0%. The 2023 and 2024 values are provisional and will rise; no growth statement in this report ends in either year.
* 2023 and 2024: incomplete data. Applications are published about 18 months after filing, so the two most recent filing years are still filling up in the PATSTAT Spring 2026 (patstat_2026a) edition. The 2024 application count in particular is a lower bound.
Data table: Filing trend 2014–2024
| Year of first Airbus filing | New patent families | Applications filed that year |
|---|---|---|
| 2014 | 957 | 2,680 |
| 2015 | 961 | 2,598 |
| 2016 | 970 | 2,196 |
| 2017 | 779 | 2,098 |
| 2018 | 780 | 2,355 |
| 2019 | 885 | 2,248 |
| 2020 | 612 | 1,666 |
| 2021 | 647 | 1,771 |
| 2022 | 691 | 1,915 |
| 2023* | 736 | 1,757 |
| 2024* | 590 | 780 |
| Total | 8,608 | 22,064 |
Each family is counted exactly once, in the year of its earliest Airbus application, so the column sums to the portfolio total of 8,608. The application column counts every Airbus application filed in that calendar year and therefore sums to 22,064; a single family contributes to several years when it is filed in several offices over time.
Filing Entities
Which group company appears as applicant on the record. Grouped from the harmonised applicant name (han_name) of every Airbus applicant record in the portfolio.
This chart measures the applicant on the paperwork, not where the work is done. A family is assigned to a group whenever any of its Airbus applicant records carries a name mapped to that group, so a family filed jointly by two group companies is counted in both. The four groups sum to 9,199 assignments against 8,608 distinct families. Movements between the groups can reflect a change in which legal entity signs the filing as much as a change in activity.
Commercial-aircraft entities carry 5,740 of 8,608 families and stay within a band of 387 to 657 per year. Families filed under defence-and-space entity names fall from 219 (2014) to 40 (2024), while families filed under plain group or holding names rise from 79 (2020) to 147 (2021) and stay above 89 thereafter — the two series move in opposite directions in the same years, which is what a reassignment of the filing entity would look like. The subject-matter measure points the same way but far more gently: families classified in B64G (cosmonautics) go from 3.9% of the 2014–2018 cohort to 2.9% of the 2019–2022 cohort, a mild decline, not a collapse. Read the entity series as a record of who files, and B64G as the record of what is filed.
* 2023 and 2024 are provisional filing years and will rise.
Data table: Patent families per year by filing entity group
| Year | Commercial aircraft | Defence and space | Group and holding | Helicopters |
|---|---|---|---|---|
| 2014 | 589 | 219 | 98 | 115 |
| 2015 | 558 | 250 | 86 | 115 |
| 2016 | 592 | 227 | 94 | 100 |
| 2017 | 474 | 149 | 89 | 107 |
| 2018 | 550 | 108 | 70 | 89 |
| 2019 | 657 | 115 | 98 | 76 |
| 2020 | 387 | 87 | 79 | 91 |
| 2021 | 429 | 90 | 147 | 70 |
| 2022 | 490 | 93 | 141 | 60 |
| 2023* | 547 | 66 | 125 | 47 |
| 2024* | 467 | 40 | 89 | 29 |
| Total | 5,740 | 1,444 | 1,116 | 899 |
Data table: Largest single applicant names (harmonised)
| Harmonised applicant name (han_name) | Patent families |
|---|---|
| AIRBUS OPERATIONS GMBH | 2,394 |
| AIRBUS OPERATIONS SAS | 2,094 |
| AIRBUS OPERATIONS LTD | 1,026 |
| AIRBUS DEFENCE & SPACE GMBH | 701 |
| AIRBUS SAS | 669 |
| AIRBUS HELICOPTERS | 637 |
| AIRBUS OPERATIONS LIMITED | 303 |
| AIRBUS DEFENCE & SPACE SAS | 300 |
| AIRBUS OPERATIONS SL | 281 |
| AIRBUS HELICOPTERS DEUTSCHLAND GMBH | 272 |
| AIRBUS | 254 |
| AIRBUS OPERATIONS SLU | 212 |
| AIRBUS FRANCE | 155 |
| AIRBUS OPERATIONS, S.L. | 151 |
| AIRBUS SAFRAN LAUNCHERS SAS | 137 |
Harmonisation is PATSTAT's own and is not complete: AIRBUS OPERATIONS LTD and AIRBUS OPERATIONS LIMITED remain separate han_name values, as do AIRBUS OPERATIONS SL and AIRBUS OPERATIONS, S.L. The table is a view of the filing record, not a corporate register. Family counts per name are non-exclusive and sum above 8,608.
Geographic Strategy
Where the portfolio is filed, and how the mix of offices differs between families first filed in 2014–2018 and those first filed in 2019–2022.
Cohort maturity affects the right-hand chart. A family first filed in 2014 has had about twelve years to add foreign members; one first filed in 2022 has had about four, and regional or national entry from a PCT application typically happens at around 30 months. The 2019–2022 cohort is therefore structurally under-counted at every downstream office. Shares that rise against that bias (EP, CN, GB, WO) are the robust readings. DE is largely a first-filing office, so its share is barely truncated. The US figure is the most exposed in the table: 66.3% for the later cohort is a lower bound and should not be read as a decline.
The single largest change is the EP route: 52.9% of the 2014–2018 cohort carries an EP application against 67.4% of the 2019–2022 cohort, a rise of 14.5 percentage points that runs against the maturity bias described above. The German national office moves the other way, from 21.9% to 13.5%. China rises from 31.1% to 35.6% and is the fourth most used office across the whole period, with 2,743 families. Over the full period 68.6% of families (5,909 of 8,608) are filed at more than one office.
Data table: Families by filing office, 2014–2024
| Office | Patent families | Applications | Share of 8,608 families |
|---|---|---|---|
| US — United States | 5,481 | 5,678 | 63.7% |
| EP — European Patent Office | 5,075 | 5,257 | 59.0% |
| FR — France | 3,042 | 3,056 | 35.3% |
| CN — China | 2,743 | 2,783 | 31.9% |
| DE — Germany | 1,404 | 1,409 | 16.3% |
| GB — United Kingdom | 1,186 | 1,203 | 13.8% |
| WO — PCT | 927 | 940 | 10.8% |
| CA — Canada | 506 | 513 | 5.9% |
| ES — Spain | 506 | 512 | 5.9% |
| KR — Korea | 217 | 223 | 2.5% |
| JP — Japan | 153 | 158 | 1.8% |
| PL — Poland | 83 | 83 | 1.0% |
| AU — Australia | 65 | 66 | 0.8% |
| BR — Brazil | 48 | 49 | 0.6% |
| RU — Russia | 35 | 35 | 0.4% |
Data table: Office share by first-filing cohort
| Office | Families 2014–2018 cohort | Share of 4,447 | Families 2019–2022 cohort | Share of 2,835 |
|---|---|---|---|---|
| US | 3,026 | 68.0% | 1,880 | 66.3% |
| EP | 2,351 | 52.9% | 1,910 | 67.4% |
| FR | 1,688 | 38.0% | 977 | 34.5% |
| CN | 1,382 | 31.1% | 1,008 | 35.6% |
| DE | 974 | 21.9% | 383 | 13.5% |
| GB | 521 | 11.7% | 445 | 15.7% |
| WO | 472 | 10.6% | 401 | 14.1% |
| ES | 320 | 7.2% | 175 | 6.2% |
| CA | 346 | 7.8% | 142 | 5.0% |
| KR | 138 | 3.1% | 56 | 2.0% |
A family filed in several countries appears once per office, so the office columns sum well above the family total of 8,608. The cohorts stop at 2022 deliberately: 2023 and 2024 are provisional filing years, and a period boundary inside them would read publication lag as a trend.
Technology Profile
IPC subclasses assigned anywhere in the family, aggregated across all Airbus family members, plus the WIPO 35-field view of the same portfolio.
Two aircraft subclasses carry the portfolio: B64D (equipment fitted in or to aircraft, 3,276 families, 38.1%) and B64C (aeroplanes; helicopters, 3,148 families, 36.6%). Ground and deck installations (B64F) follow at 962. Below that the profile is an aerospace supplier profile rather than an airframe one: plastics shaping (B29C, 773), digital data processing (G06F, 464), gas-turbine plants (F02C, 441) and fastening elements (F16B, 391). In the WIPO 35-field view a single field, Transport, covers 5,769 families (67.0%); the next largest, Mechanical elements, covers 1,267 (14.7%).
A family carries several IPC subclasses and is counted under each one, so the columns below sum far above the portfolio total of 8,608. The same holds for the WIPO fields. No pie or doughnut is used for this reason. This is a breakdown of an applicant-defined set — the family set comes from applicant names, not from a code list — so no classification-code expansion applies to it.
Data table: IPC subclasses in the Airbus portfolio
| IPC subclass | Title | Families | Share of 8,608 | 2014–2018 cohort | 2019–2022 cohort |
|---|---|---|---|---|---|
| B64D | Equipment for fitting in or to aircraft; flight suits; parachutes | 3,276 | 38.1% | 1,625 | 1,133 |
| B64C | Aeroplanes; helicopters | 3,148 | 36.6% | 1,648 | 1,100 |
| B64F | Ground or aircraft-carrier-deck installations | 962 | 11.2% | 452 | 371 |
| B29C | Shaping or joining of plastics | 773 | 9.0% | 470 | 225 |
| G06F | Electric digital data processing | 464 | 5.4% | 283 | 120 |
| F02C | Gas-turbine plants; air intakes for jet-propulsion plants | 441 | 5.1% | 193 | 163 |
| F16B | Devices for fastening or securing constructional elements | 391 | 4.5% | 204 | 129 |
| B32B | Layered products | 365 | 4.2% | 235 | 100 |
| G05D | Systems for controlling or regulating non-electric variables | 360 | 4.2% | 208 | 110 |
| G08G | Traffic control systems | 329 | 3.8% | 193 | 95 |
| G01C | Measuring distances, levels or bearings; surveying; navigation | 281 | 3.3% | 178 | 83 |
| B64G | Cosmonautics; vehicles or equipment therefor | 281 | 3.3% | 174 | 81 |
| F02K | Jet-propulsion plants | 280 | 3.3% | 164 | 91 |
| H04L | Transmission of digital information | 270 | 3.1% | 153 | 91 |
| F16L | Pipes; joints or fittings for pipes | 259 | 3.0% | 115 | 84 |
| B29L | Indexing scheme relating to particular articles (plastics) | 247 | 2.9% | 144 | 85 |
| G01S | Radio direction-finding; radio navigation; radar | 237 | 2.8% | 137 | 72 |
| H04B | Transmission | 226 | 2.6% | 118 | 80 |
| G01N | Investigating or analysing materials | 217 | 2.5% | 143 | 55 |
| H01M | Batteries and processes for direct conversion of chemical into electrical energy | 214 | 2.5% | 69 | 94 |
| G01M | Testing static or dynamic balance of machines or structures | 198 | 2.3% | 92 | 73 |
| F01D | Non-positive-displacement machines or engines | 178 | 2.1% | 102 | 55 |
Data table: WIPO 35 technology fields
| Technology field | Families | Share of 8,608 | 2014–2018 cohort | 2019–2022 cohort |
|---|---|---|---|---|
| Mechanical engineering — Transport | 5,769 | 67.0% | 2,916 | 2,005 |
| Mechanical engineering — Mechanical elements | 1,267 | 14.7% | 639 | 418 |
| Instruments — Measurement | 1,148 | 13.3% | 662 | 351 |
| Mechanical engineering — Other special machines | 916 | 10.6% | 554 | 271 |
| Instruments — Control | 809 | 9.4% | 448 | 258 |
| Mechanical engineering — Engines, pumps, turbines | 797 | 9.3% | 419 | 267 |
| Electrical engineering — Electrical machinery, apparatus, energy | 773 | 9.0% | 360 | 268 |
| Electrical engineering — Computer technology | 681 | 7.9% | 367 | 220 |
| Chemistry — Surface technology, coating | 488 | 5.7% | 295 | 151 |
| Mechanical engineering — Machine tools | 425 | 4.9% | 248 | 131 |
| Electrical engineering — Telecommunications | 381 | 4.4% | 203 | 143 |
| Electrical engineering — Digital communication | 350 | 4.1% | 199 | 119 |
Technology Shift Between Filing Cohorts
Share of families carrying each IPC subclass, families first filed 2014–2018 (n = 4,447) against families first filed 2019–2022 (n = 2,835).
The airframe core holds and slightly widens: B64D moves from 36.5% to 40.0% of cohort families and B64F, ground and deck installations, from 10.2% to 13.1%. The material-processing block moves the other way — plastics shaping B29C from 10.6% to 7.9%, layered products B32B from 5.3% to 3.5%, materials investigation G01N from 3.2% to 1.9%. The largest relative gain in the table belongs to H01M, which doubles its share from 1.6% to 3.3% on 69 and 94 families respectively; gas-turbine plants F02C rise from 4.3% to 5.7%.
Shares are computed against the number of families in each cohort, not against each other, and a family counted in several subclasses contributes to each of them. Both cohorts end before 2023 so that neither boundary sits inside a provisional filing year. H01M and G01N rest on fewer than 150 families per cohort; read them as direction, not as precise magnitude.
Data table: IPC subclass share by filing cohort
| IPC subclass | Families 2014–2018 | Share of 4,447 | Families 2019–2022 | Share of 2,835 | Change (percentage points) |
|---|---|---|---|---|---|
| B64D — Equipment fitted in or to aircraft | 1,625 | 36.5% | 1,133 | 40.0% | +3.5 |
| B64C — Aeroplanes; helicopters | 1,648 | 37.1% | 1,100 | 38.8% | +1.7 |
| B64F — Ground or deck installations | 452 | 10.2% | 371 | 13.1% | +2.9 |
| B29C — Shaping or joining of plastics | 470 | 10.6% | 225 | 7.9% | −2.7 |
| G06F — Electric digital data processing | 283 | 6.4% | 120 | 4.2% | −2.2 |
| F02C — Gas-turbine plants | 193 | 4.3% | 163 | 5.7% | +1.4 |
| B32B — Layered products | 235 | 5.3% | 100 | 3.5% | −1.8 |
| H01M — Batteries; direct conversion of chemical into electrical energy | 69 | 1.6% | 94 | 3.3% | +1.7 |
| G01N — Investigating or analysing materials | 143 | 3.2% | 55 | 1.9% | −1.3 |
| B64G — Cosmonautics | 174 | 3.9% | 81 | 2.9% | −1.0 |
Emerging Technology Areas
Seven code-defined themes measured inside the Airbus portfolio, as a share of families in each filing cohort. Classification symbols are matched on IPC and CPC together, so a family counts if any of its members carries the code at any office.
Two themes grow clearly against the cohort baseline. Families carrying an aircraft power-plant installation code (B64D 27) move from 5.9% to 10.4% — 263 against 294 families, on a cohort that is itself 36% smaller. Hydrogen and fuel-cell codes (H01M 8, C01B 3, F17C) move from 1.1% to 4.2%, from 49 to 119 families. Computing arrangements based on learning models (G06N) go from 16 to 40 families, 0.4% to 1.4%. Additive manufacturing and unmanned aircraft both lose a little share, 2.9% to 2.4% and 2.6% to 2.2%.
Two data-quality notes. First, 1,096 families in the portfolio carry a Y02T 50 tag (climate-change mitigation in aeronautics). Y tags are applied by examiners as a secondary scheme and their coverage of recent filing years is incomplete, so this report states the total only and reads no trend from it. Second, the unmanned-aircraft theme unions the subclass B64U with its predecessor group B64C 39/02, because B64U was introduced only in the 2023 CPC and IPC editions; without that union the series would show reclassification progress rather than filing behaviour.
Data table: Themed technology areas by filing cohort
| Theme (codes matched) | Families 2014–2024 | 2014–2018 cohort | Share of 4,447 | 2019–2022 cohort | Share of 2,835 |
|---|---|---|---|---|---|
| Aircraft power-plant installation (B64D 27) | 682 | 263 | 5.9% | 294 | 10.4% |
| Hydrogen and fuel cells (H01M 8, C01B 3, F17C) | 265 | 49 | 1.1% | 119 | 4.2% |
| Electrical power supply and conversion (H02J, H02K, H02M, H02P) | 218 | 95 | 2.1% | 76 | 2.7% |
| Additive manufacturing (B33Y, B22F 10) | 214 | 130 | 2.9% | 69 | 2.4% |
| Unmanned aircraft (B64U and B64C 39/02) | 186 | 116 | 2.6% | 63 | 2.2% |
| Batteries and electric storage (H01M 4, H01M 10, H01M 50) | 92 | 44 | 1.0% | 34 | 1.2% |
| AI and machine learning (G06N) | 79 | 16 | 0.4% | 40 | 1.4% |
| Climate-change mitigation, aeronautics (Y02T 50) | 1,096 | — | — | — | — |
Themes overlap by construction — a hydrogen fuel-cell family in an aircraft installation carries codes from two rows — so the column does not sum to the portfolio total. The smallest themes rest on fewer than 50 families per cohort; they are reported as direction and magnitude of share, not as precise counts of inventions.
Grant Rates by Office
Applications filed 2014–2019 and their grant status in the PATSTAT Spring 2026 snapshot. Offices with at least 100 applications in the cohort.
Grant rate here is an outcome count, not a quality score. It varies with the procedure of each office as much as with the application, and the spread in this chart — 15.4% to 99.5% — is wider than any plausible difference in the underlying inventions. One possible explanation for the low DE (24.9%) and GB (15.4%) figures is the priority-filing route, in which a national application secures a filing date and is later allowed to lapse once an EP or PCT continuation is on file; this report does not measure why applications end and cannot confirm it. WO is excluded from the section entirely: an international application is never granted, it only enters national phases, and its 7 recorded grants out of 509 applications would otherwise appear as a 1.4% rate.
Among offices that examine substantively, the cohort reaches 81.0% at the USPTO (2,600 of 3,210 applications), 82.6% at the French office (1,665 of 2,016) and 76.3% at the EPO (2,027 of 2,658). China sits lower at 59.3% (875 of 1,475), Canada at 54.3% and Japan at 64.2%. These are counts of individual applications, not families: a family granted in one office and refused in another contributes to both columns.
The 2014–2019 filing cohort is used because grant rates over more recent years measure examination speed as much as outcome — applications filed from 2020 onwards are still substantially pending, and a rate computed over them is artificially low. Status is as recorded in PATSTAT Spring 2026 (patstat_2026a); a pending application counts as not granted.
Data table: Grant rate by office, applications filed 2014–2019
| Office | Applications filed 2014–2019 | Granted | Grant rate |
|---|---|---|---|
| US — United States | 3,210 | 2,600 | 81.0% |
| EP — European Patent Office | 2,658 | 2,027 | 76.3% |
| FR — France | 2,016 | 1,665 | 82.6% |
| CN — China | 1,475 | 875 | 59.3% |
| DE — Germany | 1,165 | 290 | 24.9% |
| GB — United Kingdom | 682 | 105 | 15.4% |
| ES — Spain | 377 | 375 | 99.5% |
| CA — Canada | 359 | 195 | 54.3% |
| KR — Korea | 146 | 114 | 78.1% |
| JP — Japan | 120 | 77 | 64.2% |
Collaboration Network
Applicants outside the Airbus group that appear on the same application as an Airbus applicant. Matched at application level, not family level, and with every Airbus name variant — Latin and non-Latin — excluded from the partner side.
Co-filing inside the group and co-filing outside it are two different things. 1,164 families (13.5%) carry at least one application with more than one applicant, but only 304 families (3.5%) name an applicant from outside the Airbus group; the remaining 860 carry no applicant that this report counts as external. Among the external partners the largest single group is public research: CNES (23 families), CNRS (19), Fraunhofer (15) and DLR (14). The PATSTAT sector field marks 92 of the 304 families as government or non-profit and 42 as university. The largest external partnership overall is with CNES, at 23 shared families.
Data table: External co-applicants with at least 5 shared families
| Co-applicant (harmonised name) | PATSTAT sector | Shared families | 2014–2018 cohort | 2019–2022 cohort |
|---|---|---|---|---|
| Centre National d'Études Spatiales (CNES) | Gov / non-profit | 23 | 13 | 7 |
| ArianeGroup SAS | Company | 20 | 18 | 2 |
| CNRS | Gov / non-profit | 19 | 16 | 1 |
| Fraunhofer-Gesellschaft | Gov / non-profit | 15 | 7 | 6 |
| Deutsches Zentrum für Luft- und Raumfahrt (DLR) | Gov / non-profit | 14 | 8 | 4 |
| C Series Aircraft Limited Partnership | Company | 10 | 8 | 2 |
| Bombardier Inc. | Company | 9 | 9 | 0 |
| Short Brothers PLC | Company | 9 | 9 | 0 |
| STELIA Aerospace | Unknown | 9 | 3 | 6 |
| CEA — Commissariat à l'énergie atomique et aux énergies alternatives | Unknown | 6 | 2 | 4 |
| Vibro-Meter SA | Company | 6 | 0 | 3 |
| Liebherr-Aerospace Lindenberg GmbH | Company | 6 | 4 | 2 |
| ArianeGroup GmbH | Company | 5 | 5 | 0 |
| Hydac Technology GmbH | Company | 5 | 5 | 0 |
| University of Surrey | University | 5 | 4 | 1 |
| The Manufacturing Technology Centre Ltd | Unknown | 5 | 0 | 1 |
| Manufacturing Technology Center Limited | Unknown | 5 | 0 | 1 |
Data table: External co-applicants by PATSTAT sector
| PATSTAT sector of the co-applicant | Families with such a co-applicant |
|---|---|
| Company | 145 |
| Government / non-profit | 92 |
| Unknown | 78 |
| University | 42 |
| Individual | 13 |
| Government / non-profit and university | 2 |
| Not set | 4 |
| Distinct families with an external co-applicant | 304 |
Sector rows sum to 376 against 304 distinct families, because a family with two external partners in different sectors is counted in both. Partner names come from PATSTAT's harmonised applicant name, which does not always merge spellings: The Manufacturing Technology Centre Ltd and Manufacturing Technology Center Limited are the same organisation recorded twice, and the 5-family rows are shown as PATSTAT records them rather than merged by hand. The two ArianeGroup rows and STELIA Aerospace appear here as external partners because this report scopes Airbus by the name on the application rather than by ownership. The same joint venture therefore sits on both sides of that boundary: filings under ArianeGroup are external, filings under its pre-2017 name Airbus Safran Launchers SAS are in scope and appear in the Filing Entities table with 137 families. The Methodology section sets out the rule and what it leaves out.
Inventor Locations
Country of record of the inventors named on the portfolio, counted as the number of families with at least one inventor registered in that country.
Invention stays close to the group's four home countries. France appears on 3,626 families (42.1%) and Germany on 3,428 (39.8%), followed by the United Kingdom with 1,164 (13.5%) and Spain with 567 (6.6%). Everything else is a long tail: the United States appears on 91 families, India on 56, Canada on 42 and China on 30. The contrast with the filing map is the point — China is the fourth most used filing office in this portfolio, with 2,743 families, and the eighth-largest inventing country in the table below, with 30. The pattern is consistent with filing for market coverage rather than for locally generated inventions.
Data table: Families by inventor country of record
| Country | Families with an inventor there | Share of 8,608 |
|---|---|---|
| FR — France | 3,626 | 42.1% |
| DE — Germany | 3,428 | 39.8% |
| GB — United Kingdom | 1,164 | 13.5% |
| ES — Spain | 567 | 6.6% |
| US — United States | 91 | 1.1% |
| IN — India | 56 | 0.7% |
| CA — Canada | 42 | 0.5% |
| CN — China | 30 | 0.3% |
| NL — Netherlands | 22 | 0.3% |
| IT — Italy | 15 | 0.2% |
| CH — Switzerland | 12 | 0.1% |
| AT — Austria | 12 | 0.1% |
A family with inventors in two countries is counted in both, so the twelve rows sum to 9,065 against 8,608 distinct families. The country is the one recorded on the inventor's person record in PATSTAT, which is an address of record rather than a workplace, and it is not populated for every inventor; families whose inventors carry no country are absent from this table.
Methodology
Search strategy, counting rules, data basis and known limitations
This report was compiled on 15 August 2026 against
EPO PATSTAT Global, Spring 2026 (patstat_2026a), with classification
labels taken from IPC edition 20260101 (in force 1 January 2026) and
CPC edition 202608 (in force 1 August 2026). Every figure in the report
comes from a query executed against that data basis; the full SQL is shipped alongside the
report in queries.sql.
Search strategy, counting rules and data source — in detail
How the Airbus portfolio was delimited
PATSTAT holds no company identifier, only names. The scope was therefore built on
person_name — the name as filed — and never on the harmonised
han_name, for the reason set out below.
-
Kept: applicant records whose own name carries an Airbus-group brand.
Latin script:
AIRBUS,EUROCOPTER,ASTRIUM,CASSIDIAN,EADSfollowed by a corporate word, and the spelled out form European Aeronautic Defence and Space. Non-Latin script: 空中客车 / 空中客車 / 空中巴士 / 空客, エアバス, 에어버스. A discovery pass across all of these name forms returns 3,148 person records, which the review below then narrows. -
Why non-Latin spellings were searched separately: Chinese, Japanese
and Korean applicant records are not harmonised onto a Latin
han_name— for these recordshan_namerepeats the original-script string — so a Latin-only search cannot reach them. Measured effect in this portfolio: 352 families in the 2014–2024 window carry an original-script Airbus applicant record, and 11 of those carry no Latin-script Airbus record at all. The addition is small here, but it also matters on the partner side: without it, the Chinese and Japanese spellings of Airbus itself would surface as external co-applicants. -
Excluded — natural persons named Eads: a bare
EADStoken matches a further 251 person records that sit in a blank or unknown sector, and the sampled names there are natural persons (Eads, Brian Donovan, Brandon Eads, James M. Eads). The PATSTATpsn_sector = INDIVIDUALflag marks only 13 records in the whole match set, so it cannot be relied on here; the pattern was tightened instead to require a corporate word after the token. -
Excluded — design rights and utility models: the filter
ipr_type = 'PI'restricts the set to patents of invention. In the 2014–2024 filing window, 127 utility-model families and 40 design families with an Airbus applicant fall outside the report because of it.
Why han_name was used as a cross-check and not as the scope
-
It brings in foreign organisations. In the live edition, the harmonised
name
AIRBUS UK LTDis attached to person records reading AlliedSignal Ltd., AlliedSignal Limited, Allied Signal Limited and Honeywell UK Ltd., Bristol. Building the scope onhan_namewould have imported them. -
It also recovers genuine members, so it was run in the other direction
too:
han_nameconnects Dutch Space B.V. to Airbus Defence and Space Netherlands and PlantCML, Inc. to Cassidian Communications. The measured contribution of that route is 2 families, none of them unique to it — so nothing was lost by leaving it out of the scope definition. -
Within the fixed scope,
han_nameis used purely for display and grouping in the Filing Entities section, where the risk of importing a stranger no longer exists.
Family counting
-
The unit of count is the DOCDB patent family
(
docdb_family_id), so one invention filed in ten countries counts once. - The population is families whose earliest Airbus application falls in 2014–2024 — 8,608 families. Each family is placed in exactly one year, the year of that earliest application, computed over all Airbus filings of the family regardless of year. That is why the annual column in the Filing Trend section sums exactly to 8,608.
- Application counts (22,064) cover Airbus applications filed 2014–2024 within those families. A family contributes to several offices and several years.
-
Only the applicant role is counted (
applt_seq_nr > 0); the inventor role is used solely in the Inventor Locations section. - Derived rates: PCT rate = families with at least one WO member ÷ total families (927 ÷ 8,608). Grant rate = granted applications ÷ applications, per office, on the 2014–2019 filing cohort. Co-application rate = families with a co-applicant ÷ total families.
Multi-assignment note
- A patent family carries several IPC and CPC symbols, is filed at several offices, may name inventors in several countries, and may be filed by several group companies. Every breakdown in this report is therefore non-exclusive and its column sums above 8,608 — IPC subclasses and WIPO fields, filing offices, filing entities (9,199), inventor countries (9,065) and co-applicant sectors (376 against 304).
- For this reason no pie or doughnut chart appears in the report; all such breakdowns use bars, which do not imply parts of a whole.
Classification scope (QA check 8b)
- The patent set itself is defined by applicant names, not by a code list, so the IPC and WIPO-field breakdowns are descriptions of an already-fixed set. No code-set completion applies to them.
-
The Emerging Technology Areas section is code-defined, and the three standard
completions were handled as follows. Subtrees: matching is a prefix
match on the raw symbol, which is subtree-closed by construction —
H01M 8captures every group beneath it. Predecessors: the unmanned-aircraft theme unions the 2023-introduced subclassB64Uwith its predecessor groupB64C 39/02; without it the series would measure reclassification progress. Retired symbols: nois_activefilter was applied anywhere, so deprecated symbols that still sit on documents are retained. -
Codes were matched against IPC and CPC together, so a family counts if any of its
members carries the code at any office. The
Y02T 50tag is CPC-only and under-represents offices that do not assign CPC; it is reported as a total without any trend reading.
Data source
- EPO PATSTAT Global, Spring 2026 edition, dataset
patstat-mtc.patstat_2026a, on Google BigQuery. - Classification titles and hierarchy from
patstat-mtc.classification: IPC edition 20260101, CPC edition 202608. - Core tables:
tls201_appln,tls206_person,tls207_pers_appln,tls209_appln_ipc,tls224_appln_cpc,tls230_appln_techn_field,tls901_techn_field_ipc. - Analysis and visualisation: patstat-mcp server and Claude AI; report published by mtc.berlin.
Scope Limitations
- Family counts are a lower bound. Name harmonisation is never complete; Airbus filings recorded under a name that carries no group brand are not reachable by this profile.
- Separately branded subsidiaries are outside the scope, and were measured. Applicant names without an Airbus brand contribute, in families that this report does not contain: Premium AEROTEC 66, STELIA Aerospace 50, Aerolia 11, APWorks 7, Sogerma 1 — 135 families in total, about 1.6% of the portfolio size.
- Joint-venture filings that carry no Airbus brand are outside the scope. ArianeGroup contributes 357 families and MBDA 416 families whose applicants carry no Airbus-branded name at all; Airbus is one of several owners of both, and counting those 773 families as Airbus filings would credit the group with portfolios it does not hold alone. The scope rule is the name on the application, not the ownership of the company, so a joint venture that files under an Airbus-branded name is inside the report: Airbus Safran Launchers SAS — the name ArianeGroup filed under until 2017 — contributes 137 families to the portfolio, and Airbus OneWeb Satellites SAS is in scope on the same rule. The consequence is that one joint venture can appear on both sides of the boundary under its two names, as ArianeGroup does: an external co-applicant under that name, an in-scope filing entity under the older one.
- Design rights and utility models are excluded (127 utility-model and 40 design families in the window).
- The two most recent filing years are provisional. Applications publish about 18 months after filing, so 2023 and 2024 will both rise. No growth rate in this report ends in either year, and no period boundary sits inside them.
- Later cohorts are less mature. A family first filed in 2022 has had far less time to acquire foreign members than one first filed in 2014, which systematically depresses downstream office shares in the 2019–2022 cohort. Rises measured against that bias are robust; falls in offices that are entered late are not necessarily real.
- Grant status is a snapshot. A pending application counts as not granted. Grant rates are computed on the 2014–2019 filing cohort for that reason and still carry a residual pendency effect.
- External validation is limited. A timeboxed public-source search returned no per-year family count for Airbus that is comparable with the definition used here, so the plausibility check rests on internal consistency: the annual family series sums exactly to the reported total, and office and technology sub-totals behave as expected for a non-exclusive breakdown.
- Inventor countries are addresses of record, not workplaces, and are not populated for every inventor record.
Glossary — patent terms explained
- DOCDB patent family
- A group of applications covering the same invention in different countries, linked by identical priority claims. Counting families instead of applications prevents an internationally filed invention from counting many times.
- Application vs. family
- One family typically contains several applications — one per office it is filed at. This report holds 8,608 families and 22,064 applications.
- Filing authority (appln_auth)
- The office an application is filed at: EP for the European Patent Office, WO for an international PCT application, US, CN, FR, DE and so on.
- PCT / WO application
- An international application under the Patent Cooperation Treaty. It is never granted itself; it buys time before national or regional filings must be made.
- IPC and CPC
- International and Cooperative Patent Classification — hierarchical technology codes. A subclass such as B64C is the four-character level used for technology profiles.
- WIPO 35 technology fields
- A standard mapping of IPC codes onto 35 broad fields grouped in five sectors, used for cross-industry comparison.
- han_name
- PATSTAT's harmonised applicant name. Useful for grouping, unreliable as a search key — see the cross-check note above.
- psn_sector
- PATSTAT's classification of an applicant as company, university, government or non-profit, or individual. Frequently unset.
- Grant rate
- Share of applications recorded as granted at the time of the data snapshot. Recent filings depress it because examination is still running.
- PATSTAT edition
- PATSTAT is published twice a year. Figures move between editions, most of all for the two most recent filing years, so the edition is named wherever a number is quoted.
Report Files
The report ships with queries.sql, containing every SQL statement behind the
figures, with the PATSTAT and classification editions recorded in its header and each query
annotated with the section it feeds. Re-running the file against the same edition reproduces
the numbers exactly; re-running it against a later edition will not, and that is the point of
naming the edition.
Like what you see?
This report was built with a fully reproducible pipeline: EPO PATSTAT Global on BigQuery, a custom MCP server, and Claude AI for analysis and visualisation. Every figure traces back to a query shipped with the report.