Every cost discussed before grant is a cost of acquisition. What follows grant is a cost of possession, and it behaves quite differently: it is scheduled, escalating, entirely predictable, and it continues for as long as the holder chooses to keep the right in force. Most people planning a filing budget for the first category and discover the second some years later, usually in the form of a notice.
The economic logic is worth stating plainly, because it is not accidental. Granting offices could recover their costs entirely at filing. Instead they charge modestly at the front and heavily at the back, which shifts the burden onto rights that turn out to be worth keeping and lets the rest expire quietly into the public domain. The escalating curve is a filtering mechanism dressed as a fee schedule.
I · The Schedule
Three Windows, Rising Sharply
In some jurisdictions the charge falls due at intervals measured from the grant date. A common arrangement takes payment at three and a half years, seven and a half years, and eleven and a half years, in the region of $2,000, $3,800 and $7,700 respectively for a holder claiming no discount — roughly $13,500 across the life of a single patent, before any professional fee for handling the payments. The small entity discount cuts each by sixty per cent, the micro entity discount by eighty, bringing the lifetime total down to about $5,400 or $2,700.
In many other territories the structure is annual instead. A renewal falls due each year, often beginning in the third or fourth year from the filing date rather than from grant, starting at a modest figure and climbing every year until the final year costs several times the first. The cumulative total over a twenty-year term is broadly comparable, but the cash-flow shape is entirely different, and a portfolio held in a dozen territories under an annual regime generates a payment obligation somewhere in almost every month of the year.
Two exemptions are widely shared. Design rights and plant rights usually carry no renewal obligation at all, running their shorter term on the strength of the issue fee alone. And a patent that has already expired at the end of its term is simply finished; there is no charge for a lapsed right, and no obligation to formally surrender one.
II · The Filter
Half of All Patents Are Allowed to Die
Attrition across the schedule is steep and consistent. Something in the order of eighty-five per cent of holders pay the first renewal, around sixty per cent reach the second, and fewer than half pay the third. By the end of the full term, only a minority of granted patents are still in force. That is not a sign of a system working badly; it is the system working exactly as designed.
The reason is that the value of a patent is discovered after grant, not before. A right granted on a product that never reached market is worth nothing at year eleven, whatever it cost to obtain. A right granted on a component that turned out to sit in the middle of a supply chain is worth many multiples of the fee. Neither outcome is knowable at filing, and the escalating schedule pushes the expensive decision back to the point where the holder finally has the information to make it.
There is a second-order effect that independent holders should understand. Because patent maintenance fees are the same amount for everyone in a given tier, they weigh far more heavily on a portfolio of three than on a portfolio of three thousand. A large holder treats renewals as a line item; a sole inventor treats each one as a decision about whether the invention has earned another four years. The discounts exist precisely because that asymmetry was recognised, and claiming the tier the holder actually qualifies for at each payment is the single largest lever available.
The fee schedule asks the same question three times, four years apart: is this right still worth more than the cheque?
Why the curve escalates rather than flattens
III · Lapse and Revival
What Happens When a Payment Is Missed
Each window opens six months before the due date and closes on it. Miss the date and a grace period of a further six months applies, during which the payment is accepted with a surcharge — modest at the first window, several hundred dollars at the third. Miss the grace period and the patent expires by operation of law, on that day, without any further notice being required.
Revival is possible but deliberately uncomfortable. A petition asserting that the delay was unintentional, filed within a defined outer window and accompanied by both the missed fee and a substantial petition charge in the region of $2,100, will usually be granted. Beyond that window the standard rises to a showing that the delay was unavoidable, which is a genuinely difficult thing to demonstrate.
Even a successful revival leaves consequences. In many systems, a third party who began making or selling the invention during the lapse acquires intervening rights — a defence, sometimes absolute, against enforcement of the revived patent against that activity. A gap in the file is therefore not a clerical embarrassment but a permanent hole in the right's enforceability against whoever walked through it.
The administrative failures that cause lapses are almost always mundane. An address on file that was never updated after a move. An assignment recorded without a corresponding change of correspondence details. A docketing spreadsheet inherited by someone who did not know what the dates meant. No granting office guarantees a reminder, and where courtesy notices exist they are sent to the address of record and not to the inventor. An annuity service charging $50 to $150 per payment exists to remove precisely this failure mode, and for a holder with more than two or three rights it is the cheapest professional service in the entire lifecycle.
IV · The Decision
How to Sit Down and Do the Arithmetic
Each window is a small investment appraisal, and it takes about an hour. Four questions carry most of the weight. Is anything currently being sold under this right, by the holder or a licensee? Has the patent been cited as prior art in anyone else's application, which is the clearest external signal that competitors are working in the same space? Is there a live or plausible licensing conversation? And how much term remains — because the third payment buys roughly eight and a half years, while the first buys four.
The cost-per-remaining-year figure is the one most people never calculate. Paying $7,700 at the eleven-and-a-half-year window for eight and a half years of remaining exclusivity is about $900 a year, which is cheap for a right that protects a shipping product and absurd for one that protects a prototype. A single licence at any realistic royalty covers the entire lifetime schedule several times over; no licence at all makes the schedule pure expenditure.
Where multiple rights cover one product, the family can be pruned rather than abandoned wholesale — keeping the patent with the broadest independent claim and letting narrower siblings lapse. And where a right no longer serves the holder, letting it expire is a legitimate strategic act rather than a failure: the disclosure remains published and continues to function as prior art against everyone else, which is occasionally the only value it ever had. Seen against the whole outlay, and it is worth reading these numbers next to the end-to-end cost of taking an idea from sketch to grant, patent maintenance fees are the only part of the budget the holder can still decline to pay.
The long view is what makes that decision tolerable. Reporting on inventors working on environmental technologies and on inventor-led thinking in the built environment both describe products whose commercial life began years after grant — the exact interval the renewal schedule tests. An account of four decades of work alongside independent inventors makes the same point across many cases, and the popular appetite for stories about invention reliably stops at the moment of grant, which is roughly where the paying actually starts.
End of report