Patent Survival Curve: How Many Patents Reach Full Term?
Only 41.4% of US utility patents reach full term. Of patents granted in 2014, 85.8% survived the 3.5-year maintenance fee, 64.6% survived the 7.5-year fee, and 41.4% had all three fees paid. The remaining 59% expired early because an owner decided the patent was no longer worth the fee. These figures are computed from 27,273,654 USPTO maintenance fee records covering 8,262,336 US utility patents granted since 1981.
Survival Rates at Each Maintenance Fee Stage
A US utility patent requires three maintenance fee payments to reach its full 20-year term — at 3.5, 7.5 and 11.5 years after grant. Each one is a decision point where the owner weighs the fee against the value of continued exclusivity. Most patents do not survive all three.
If You Have Already Paid, You Are More Likely to Pay Again
The overall rate understates what happens to any individual patent, because it averages across owners who abandoned at the first fee and owners still paying at the third. The conditional rate is the more useful number if you already hold a patent and have paid before.
Paying a maintenance fee is a revealed valuation. An owner who has paid twice has twice judged the patent worth keeping — which is why 63% of them pay a third time, against an overall full-term rate of 41.4%. If you are approaching a fee deadline and unsure, the question is not what most patents do. It is what your patent is worth.
Patent Survival by Grant Year
Survival rates by the year the patent was granted. A cohort is only reported for a stage once every patent in it has passed that fee window plus the six-month grace period — blank cells mean the data is not yet observable, not that the rate is zero.
| Grant Year | Patents Granted | Survived 3.5 yr | Survived 7.5 yr | Reached Full Term |
|---|---|---|---|---|
| 2000 | 157,420 | 85.7% | 68.7% | 51.1% |
| 2001 | 165,957 | 86.3% | 68.1% | 50.3% |
| 2002 | 167,298 | 87.5% | 67.4% | 47.7% |
| 2003 | 168,984 | 88.5% | 69.4% | 48% |
| 2004 | 164,257 | 88.7% | 70.7% | 47.4% |
| 2005 | 143,784 | 87.5% | 69.3% | 46% |
| 2006 | 173,753 | 85.9% | 67.4% | 44.2% |
| 2007 | 157,253 | 87.3% | 67.4% | 44.3% |
| 2008 | 157,726 | 88.1% | 66.9% | 42.8% |
| 2009 | 167,275 | 87.3% | 66.5% | 42% |
| 2010 | 219,522 | 86.6% | 65.9% | 43.1% |
| 2011 | 224,391 | 85.7% | 64.9% | 42.4% |
| 2012 | 253,064 | 85.7% | 64.6% | 42.1% |
| 2013 | 277,764 | 86.1% | 64.4% | 41.4% |
| 2014 | 300,626 | 86.2% | 65.8% | 41.4% |
| 2015 | 298,367 | 86.4% | 66.5% | — |
| 2016 | 303,006 | 86.2% | 66.1% | — |
| 2017 | 318,773 | 85.9% | 65.3% | — |
| 2018 | 307,722 | 87% | 64.6% | — |
| 2019 | 354,394 | 87.3% | — | — |
| 2020 | 351,978 | 87% | — | — |
| 2021 | 327,316 | 86.3% | — | — |
| 2022 | 322,962 | 85.8% | — | — |
Ipiry Patent Survival Curve v1.0, computed from the USPTO Patent Maintenance Fee Events file dated 2026-08-24. Source: USPTO. Updated quarterly.
Methodology
Source
The USPTO Patent Maintenance Fee Events file records every maintenance fee event for patents granted since 1 September 1981. It is a US government work in the public domain, refreshed weekly. Version 1.0 of this curve uses the file dated 2026-08-24: 27,273,654 event rows.
Unit and denominator
The unit is the individual granted US utility patent, not the patent family. The denominator for each stage is utility patents by grant year. Design, plant and reissue patents and statutory invention registrations are excluded — 561,958 rows dropped — because design and plant patents pay no maintenance fees at all. Including them would understate every survival rate by several points.
What counts as survival
A patent counts as surviving a stage if a payment for that stage is recorded. Payments made late, within the six-month grace period, count as survival, because the patent remained in force.
Censoring
A cohort is reported for a stage only once every patent in it has passed that stage's window plus the grace period. As of 2026-08-24, the 3.5-year figure is complete through grant year 2022, the 7.5-year figure through 2018, and the 11.5-year figure through 2014. Later cohorts are shown blank rather than zero.
Validation
Cohort sizes match published USPTO utility grant totals — the 2019 cohort computes to 354,394 against approximately 354,000 reported — confirming the file covers substantially all utility patents rather than a self-selected subset.
Known limitations
- Refunds are not netted out. Where a payment was later refunded, v1.0 still counts the payment. The expected effect is small but has not been quantified.
- Reissue patents are excluded. They do pay maintenance fees but run on the original patent's term and need separate handling.
- The unit is the patent, not the family. An owner may abandon a US patent while keeping a continuation alive, so this reads as more abandonment than the underlying invention experiences.
- Two fee codes covering payments under 37 CFR 1.28(c) are not assignable to a stage and are excluded.