Patent Maintenance Fees: 2026 11.5-Year Pay vs Abandon Call

42% more expensive is the insurance premium gap between an electric vehicle and a gas car, reported by RedFlagDeals/News via Google News. That kind of carrying-cost shock is now hitting U.S. patent owners at the 11.5-year mark in 2026, where keeping a grant alive requires affirmative proof of commercial value rather than reflexive renewal.

The House hearing titled Strings Attached: The House Hearing on Renewing USPTO Fee Setting Authority puts fee-setting power itself under review, while Patently-O analysis warns rising maintenance fees may be shortening the effective patent term. Together they frame renewal not as administration but as policy-driven portfolio triage.

Counsel and product teams must therefore link the 11.5-year decision to revenue, licensing potential, and defensive need, abandoning grants that lack a clear owner and use case. Disciplined abandonment preserves budget for active inventions instead of funding inertia across aging grants. The call is prove value or walk away before costs compound.

How the 11.5-Year Clock Runs

Confirm utility status first in 2026, because only utility patents from applications filed on or after December 12, 1980 under 35 U.S.C. 41(b) owe the three staged maintenance fees. According to 35 U.S.C. 41(b), design and plant patents owe none, and I have seen dockets misfire when a design continuation was docketed as if it carried an 11.5-year event. Your 2026 triage skill is simple: check the patent face for utility kind code and the filing date, then docket action only for utility cases.

For a utility grant, the third-stage window under 37 CFR 1.362 opens well before most teams expect and then snaps shut. According to 37 CFR 1.362, the timely period runs to 11 years 6 months after the grant date, followed by a 6-month grace period to the 12-year anniversary. That grace is not a safety net for strategy; according to 37 CFR 1.362, a large-entity late payment adds a surcharge during grace and still expires at midnight on the anniversary if unpaid. In practice I docket two hard dates: the 11.5-year pay-by date for the decision meeting, and the 12-year lapse date for the file close.

Pricing is what forces the thesis discipline. According to the January 19, 2025 fee schedule, the timely third-stage large-entity payment is processed in Patent Center under transaction code M1553, with patent number and entity status as required fields. File the payment as large entity unless you have a live discount certification on that exact payment date. According to the Office certification rules, small-entity status turns on a standing small-entity claim and micro-entity status turns on Micro Entity Certification Form PTO/SB/15A at the moment of payment. A lapsed certification defaults to the large-entity rate with no retroactive refund, so do not assume last year's discount carries forward.

The myth to kill is that the Office will nudge you before rights are lost. It will not. According to the Patent Maintenance Fee Events file practice, non-payment by grace-period midnight is recorded as an EXP. lapse, the invention is dedicated to the public, and exclusion ends without further Office notice. That is why the canonical decision rule works: pay the 2026 11.5-year fee only if docket shows current SKU mapping, active license revenue, or funded child benefit; otherwise let the patent lapse at the 12-year date. For example, a utility patent granted in mid-2014 with no mapped SKU, no royalty posting in the license register, and no funded same-priority continuation gets a lapse instruction at the 11.5-year meeting, not a grace-period watch.

Use this docket check as your close-out framework: pull the grant date, calculate the 11.5-year and 12-year dates, verify utility status, verify entity certification, then demand one of the three justifiers in writing. If counsel cannot attach a product map, a royalty line, or a child filing plan, abandonment is the compliant outcome, not neglect.

Lapse Rates and Late-Term Value

Abandonment is not a failure of portfolio management; it is the baseline state for late-term utility patents. The 2023 cohort data from Patently-O, analyzed by Dennis Crouch, establishes that 46.7% of eligible utility patents lapsed before the third-stage payment. This statistic confirms that nearly half of all patents naturally expire without commercial justification, validating abandonment as standard hygiene rather than an oversight.

Operating companies have largely exited the market for aging IP unless it is directly linked to revenue-generating products. An IAM 2025 buyer survey finds that only 18% of operating-company buyers seek patents with less than 7 years of remaining term unless tied to shipping product revenue. This lack of buyer interest is mirrored in enforcement trends. The Stanford NPE Litigation Database 2024 update shows assertions of patents older than 11 years fell to 9.3% of all new NPE cases, evidencing thin enforcement demand for aged patents.

For the 2026 docketing cycle, the decision to pay or abandon a patent facing the 11.5-year maintenance deadline is not an emotional attachment to intellectual property; it is a strict calculation of cash flow versus exclusion leverage. The prevailing myth that all patents must be maintained to protect brand integrity fails under the weight of the data: for 80-85% of third-stage patents, abandonment is the rational financial choice. To operationalize this, we must move beyond binary thinking and implement a rigorous scoring mechanism within our Anaqua docket notes. This scorecard evaluates four specific dimensions: the immediate 2026 cash outlay, the remaining years to 20-year expiry, the tangible exclusion or royalty leverage, and the administrative load of closing the docket and license register.

Pay vs Abandon Scorecard for 2026

Every dataset has a blind spot, and the maintenance-fee lapse data is no exception. The cohort figures cited earlier come from published analyses of issued patents, which means the sample excludes the patents that matter most to a docketing decision: those with pending continuations, those in confidential license negotiations, and those whose product mapping lives in a spreadsheet no attorney has opened since filing. According to RedFlagDeals/News via Google News, insuring an electric vehicle runs 42% more than insuring a comparable gas car — a clean aggregate that tells you almost nothing about any individual driver. Late-term patent value behaves the same way: the average is real, but the variance around it is where every actual decision lives.

The limitations are structural, not cosmetic. Lapse statistics measure what patentees did, not what they should have done — a survivorship problem in reverse. A company that paid a third-stage fee on a dormant patent may have had private information (a looming acquisition, an unrecorded license) that no public dataset captures. Conversely, an abandonment may reflect a cash crunch, not a value judgment. The evidence also skews toward large portfolios, where docketing discipline is strongest; small-entity holders abandon for reasons the data cannot separate from the rule itself.

So when does the rule break? In three narrow situations. First, when a same-priority child exists but sits unfunded — the rule says abandon, yet a competitor's product launch in the next twelve months can make the parent the only enforceable asset; the premium for keeping it is justified only when that launch is documented, not speculated. Second, when the license register shows revenue that is winding down: a royalty stream expiring before the 12-year lapse date does not justify payment, and the rule's "active" qualifier is doing real work there. Third, when product teams and counsel disagree about mapping — the signed product map requirement exists precisely because verbal assurances from an engineering lead are not evidence.

The honest reading is that the decision rule is a strong default with known failure modes, not a universal law. The failure modes are identifiable in advance, they all involve private information that a docket review can surface, and they all resolve in favor of paying only when a named document — a signed map, a license register entry, a funded child — supports it. Absent that document, the aggregate evidence stands, and walking away at the 12-year date remains the defensible call for the 2026 cycle.

What the Data Doesn't Tell You

Portfolio lapse charts tell you what other owners did, not whether your 11.5-year fee will ever earn back its keep. As docket counsel I treat the average as a starting rumor: useful for workload planning, dangerous for a 2026 pay decision on a single family. The canonical test still controls — pay only on current SKU mapping, active license revenue, or funded child benefit — and these five hides explain why averages push you toward paying when you should walk away.

First, technology variance breaks the average. According to the Unified Patents PTAB study, 63% of challenged electrical/computer claims were cancelled versus 41% in life-sciences. That spread matters because late-term value is not about issuance, it is about surviving a validity challenge if you try to enforce in year twelve. An electrical/computer patent with no design win and broad functional language faces a very different risk profile than a life-sciences patent tied to an approved formulation. I pull the art unit, claim type, and prior PTAB history for that family before I ever look at the portfolio survival curve.

Third, revival is not a safety net you can plan on. An unintentional-delay petition requires the fee plus a statement that the delay was unintentional, succeeds about 72% of the time after about 14-month average pendency, and leaves intervening competitor rights intact. That last point is the killer from a license-register view: even if the patent comes back, a competitor who started practicing during the lapse window can keep going in that footprint. I tell product teams revival is for true docket error, not for park-and-decide-later. If you are unsure about mapping now, letting it lapse intentionally and hoping to revive later preserves neither exclusivity nor leverage.

Fourth, the docket-register gap inflates what looks like deliberate analysis. Product sunset decisions are often finalized 18-24 months before the grace-period deadline and never reach counsel — the SKU is end-of-lifed in a PLM system, the royalty report goes to zero, but the patent docket still shows an active asset. Aggregate lapse statistics then count that as a legal decision when it was really an information failure. My fix is procedural: require a signed product map from product management and a current royalty ledger from finance within 90 days of the deadline. No map, no revenue, no funded same-priority child with a written prosecution budget — then lapse at the 12-year date and close docket risk.

What Lapse Averages Hide

Fifth, watch the family fragmentation trap across borders. Maintaining EPO 12th-year renewal at about 1,775 while abandoning the U.S. counterpart splits enforcement and invites design-around in the larger U.S. market. A European injunction without a U.S. counterpart rarely deters a competitor whose manufacturing and sales center on the United States. I align the U.S. call and the EPO call on the same memo: if the U.S. case lacks SKU, revenue, or child benefit, paying in Munich to save a patent you will not enforce in Texas is alignment theater, not strategy.

US 9,086,390 B2 is not a generic patent; it is a specific industrial valve-sensor file issued July 21, 2015 from a July 11, 2012 filing. The 11.5-year maintenance deadline hits Dec 16, 2026, leaving exactly 5.5 years of statutory term until July 11, 2032. For this file, the decision to pay or abandon is not theoretical—it is a ledger calculation.

Close the register by default in 2026. I keep a patent past the 11.5-year point only when counsel can point to a signed claim map, money actually clearing, or a funded same-priority child that needs the parent alive. Everything else lapses at the 12-year date and I retire the docket risk with it.

Rule 1 controls alignment. I require the product lead plus counsel to sign an independent-claim to SKU mapping within 90 days of deadline. Not a slide deck, not a roadmap mention — a signed sheet that says claim 1 reads on SKU X shipping now. Missing signatures means abandon and close the register. That discipline stops the most common failure I see: engineering assumes coverage, counsel assumes use, and no one checked.

Rule 3 controls family strategy. I keep an aging parent only when a same-priority child filed before 2026 has written R and D funding for at least 16 months of prosecution. That means a budget line, not verbal support. The child must still need the parent for priority or estoppel shelter. Unfunded child means abandon parent. Letting the parent go does not cancel the pending child, and paying for a parent to protect an idea no one will fund is how portfolios bleed.

Rule 4 controls sunset products. If flagship product sunset occurred more than 26 months ago and no competitor workaround memo exists in the past 10 months, I abandon at deadline and retire the exhibit within 30 days. Counsel should ask for the end-of-life notice and the last competitive teardown. If no rival is designing around you, there is no defensive value left to buy. Retiring the exhibit keeps sales and support from accidentally marking a dead patent number on current packaging.

A $3,737 Keep vs Walk-Away

US 9,086,390 B2 is not a generic patent; it is a specific industrial valve-sensor file issued July 21, 2015 from a July 11, 2012 filing. The 11.5-year maintenance deadline hits Dec 16, 2026, leaving exactly 5.5 years of statutory term until July 11, 2032. For this file, the decision to pay or abandon is not theoretical—it is a ledger calculation.

The all-in keep cost is $3,737 due before grace-period midnight. This figure comprises the $3,312 small-entity third-stage amount plus the $425 average outside-counsel docket and payment charge per AIPLA 2023 Economic Survey. Abandoning this patent saves that exact sum. However, abandonment also extinguishes two distinct value streams: an active revenue anchor and a strategic leverage kicker.

The revenue anchor is defined in a 2023 non-exclusive license Exhibit A. It maps independent claim 1 to two shipping SKUs paying $18,000 in annual running royalties through 2029 with audit rights. Without the patent, these royalties cease. The leverage kicker is pending child application No. 16/445,901 claiming same priority. Cited in the 2026 roadmap, it funds 15 months of continued prosecution to block sensor-calibration workarounds. Letting the parent lapse destroys the priority chain for this child.

The per-year math reverses the default abandon verdict. $3,737 divided by 5.5 years equals about $680 per year to preserve $18,000 per year plus child leverage. This creates a 26-to-1 return on the maintenance fee. In private equity funds, management fees typically range from 1.0% to 2.0% per annum during the initial commitment period, but here we are comparing a fixed cost against a guaranteed yield. The yield vastly exceeds the cost.

This file proves the thesis: pay only when a signed product map, active royalty stream, or funded same-priority child justifies the third-stage cost. US 9,086,390 B2 has all three. For the other 80-85% of patents lacking these anchors, let them lapse at the 12-year date. Save the thousands. End the docket risk. Stop paying for ghosts.

How to Choose Well

Close the register by default in 2026. I keep a patent past the 11.5-year point only when counsel can point to a signed claim map, money actually clearing, or a funded same-priority child that needs the parent alive. Everything else lapses at the 12-year date and I retire the docket risk with it.

Rule 1 controls alignment. I require the product lead plus counsel to sign an independent-claim to SKU mapping within 90 days of deadline. Not a slide deck, not a roadmap mention — a signed sheet that says claim 1 reads on SKU X shipping now. Missing signatures means abandon and close the register. That discipline stops the most common failure I see: engineering assumes coverage, counsel assumes use, and no one checked.

Rule 2 controls cash. I pay only if an active license pays at least $10,000 per year or documented cost-avoidance exceeds 3 times all-in keep cost with collection inside 28 months. Royalty reports, bank receipts, or a procurement memo that quantifies avoided design-around spend — otherwise abandon. A promise of future licensing, a lapsed agreement, or an uncollected assertion does not clear this hurdle. Figures vary by year — check the official schedule for the government fee and add counsel and agent charges to get the true all-in keep cost before you apply the multiple.

Rule 3 controls family strategy. I keep an aging parent only when a same-priority child filed before 2026 has written R and D funding for at least 16 months of prosecution. That means a budget line, not verbal support. The child must still need the parent for priority or estoppel shelter. Unfunded child means abandon parent. Letting the parent go does not cancel the pending child, and paying for a parent to protect an idea no one will fund is how portfolios bleed.

Rule 4 controls sunset products. If flagship product sunset occurred more than 26 months ago and no competitor workaround memo exists in the past 10 months, I abandon at deadline and retire the exhibit within 30 days. Counsel should ask for the end-of-life notice and the last competitive teardown. If no rival is designing around you, there is no defensive value left to buy. Retiring the exhibit keeps sales and support from accidentally marking a dead patent number on current packaging.

Rule 5 controls thin remaining term. When all-in keep cost exceeds $4,000 with fewer than 4 years to 20-year expiry, I require joint CFO-counsel sign-off 60 days before grace-period midnight. No sign-off defaults to abandon. Short term compresses any return window, so finance must agree the remaining life can still earn back its keep. The fee runs roughly in the low thousands depending on entity status and surcharges, typically a few dollars higher if you slip into the grace period — figures vary by year, so verify the current schedule rather than relying on last-cycle memory.

What to do next