Orphaned oil and gas wells — 141,959 documented, and the bond floor is headed back down

What is happening?

The Interstate Oil and Gas Compact Commission counted 141,959 documented orphaned oil and gas wells on state and private land in a 2023 survey, up 54 percent from 92,203 in 2020, and estimated a further 250,000 to 740,000 undocumented wells whose locations are not known. The Bureau of Land Management puts the average cost to plug one well and reclaim the surface at $71,000.

The minimum bond an operator had to post against an individual federal lease was $10,000, an amount set in 1960. A 2024 rule raised that floor to $150,000. On 2026-06-24 the BLM proposed returning the minimum bond amounts to their pre-2024 levels.

Congress appropriated $4.677 billion under the Infrastructure Investment and Jobs Act to plug wells that operators left behind. Through fiscal year 2025 about $1.85 billion of it had been distributed, and the program reports over 10,000 wells plugged through 2025-06-30.

Whose problem is this?

The operator makes the decision to stop maintaining a well, and the bond it forfeits is the whole of what it pays. Everything above that amount lands on the parties in the last row, none of whom were at the table when the bond was set.

Where does this problem end?

The boundary matters because the count itself is unsettled, and a well that no regulator has recorded cannot be plugged with money that is handed out by inventory.

What is the state now, and what should it be?

Now

Needs a new measurementThe target state: none of the documents read here states a national target for how many orphaned wells will be plugged, by when, or what share of the inventory the appropriation is meant to cover. The federal report says plainly that a comprehensive national inventory has not been established and that there is no standard definition of an orphaned well.

How big is it?

The scale is stated in wells, not in people. Documented wells number 141,959, and the estimated undocumented range runs from 250,000 to 740,000, so the plausible inventory is roughly 392,000 to 882,000 wells.

At the BLM average of $71,000 per well, our own arithmetic on those two published figures gives a cost of about $10.1 billion for the documented wells alone, and $27.8 billion to $62.6 billion across the full estimated inventory. The $4.677 billion appropriated is therefore enough for about 65,900 wells at that unit cost, or under half the documented count. Wells plugged so far total 10,518, about 7 percent of the documented inventory.

Needs a new measurementThe affected population cannot be derived. No source read in this round reports how many people live near an orphaned well, and the largest part of the inventory has no recorded location at all. Multiplying a well count by an invented proximity factor would manufacture a number rather than find one.

Under what conditions does it arise?

Four conditions hold the state in place.

1. The bond is smaller than the work. A floor of $10,000 against an average cost of $71,000 means forfeiting the bond is the cheaper exit. 2. The obligation outlives the balance sheet. A well can produce for decades and change hands several times; the operator holding it at the end is often the smallest and least able to pay, and the bond travels with the lease rather than with the original profit. 3. Nobody has to find the undocumented wells. Between 250,000 and 740,000 wells are estimated to exist without records. Counting them enlarges the recognized liability, so the incentive to count runs backwards. 4. The public purse is a working substitute. Once an appropriation exists to plug wells that operators abandoned, the shortfall stops presenting as a crisis and starts presenting as a program.

What has been tried?

Both directions were tried at once. The money that pays for past abandonment kept flowing while the bond floor that would price future abandonment was raised, then delayed, then proposed for reversal.

What was found?

Why is it still unsolved?

Cost structure — the price of abandoning a well was set below the price of cleaning it up, so abandonment is the rational choice and the difference is paid by someone else.

The party that decides whether to plug a well is not the party that pays when it is left unplugged. An operator that stops maintaining a well loses its bond and nothing more. At the federal minimum in force until 2024 that was $10,000 against an average cost of $71,000, a gap of $61,000 per well that has to land somewhere. It landed on the appropriation. Congress put up $4.677 billion, which by our arithmetic at the BLM unit cost buys about 65,900 wells against 141,959 documented ones, before any of the estimated 250,000 to 740,000 undocumented wells are counted.

The corrective was attempted and is being reversed. The 2024 rule raised the lease minimum to $150,000, which the BLM justified by noting that its own average cost is $71,000 and that a statewide bond covers a median of seven wells, so $500,000 covers seven wells at that average. In December 2025 the compliance deadline for statewide bonds moved out a year. In June 2026 the BLM proposed returning the minimum bond amounts to those in force before the 2024 rule. If that proposal is finalized, the arithmetic that produced the current inventory returns unchanged.

There is a second reason the pressure stays low, and it is about counting. The documented total rose 54 percent in three years mostly because states got better at looking, which means the recognized liability is a function of survey effort. A regulator that counts more wells enlarges the bill it has to explain, and the wells that were never recorded generate no obligation for anyone. The federal report concedes that no comprehensive national inventory exists and that jurisdictions do not share a definition of the word orphaned.

What observation would mean it is solved?

Candidates — (a) the documented orphaned well count falls year over year (b) the minimum bond meets or exceeds the average cost to plug and reclaim, and holds (c) the rate of newly orphaned wells falls below the rate of plugging.

(a) alone is a trap. A falling count can mean wells were plugged or it can mean the survey stopped looking. The count rose 54 percent in three years for the second kind of reason, so it can fall for the mirror image of that reason.

(b) alone is a trap. A bond floor that is written but phased in over years, then extended, then proposed for reversal, has not changed any operator behavior yet. The number in the rule and the number actually posted are different observations, and only the second one prices anything.

(c) is the one that would settle it, and it is the one nobody publishes. It requires knowing how many wells became orphaned in a given year, which requires the national inventory that the federal report says does not exist.

What is it connected to?

Fills with researchplausible links to methane regulation, to state severance tax and idle well policy, to groundwater contamination, and to the financial health of small marginal-well operators. Relation type and evidence grade were not researched in this round.

What these sources do not say

• How many people are affected. No source read here reports population near orphaned wells. The largest part of the inventory has no recorded location, so the join that would produce the number cannot be made from public data yet.
• What a plugged well actually cost. The federal report gives funds distributed and wells plugged but no cost per well, and dividing one by the other would be wrong because grants are awarded ahead of the work. The $71,000 figure is a BLM planning average for federal wells, not an observed program cost.
• Which documented count is right. Totals of 141,959 and about 157,000 both circulate and none of the sources read here reconciles them. The survey table prints the 2020 documented total as 92,202 where the federal report gives 92,203.
• What the emissions figure rests on. The rate of nearly 63 million grams of methane per hour is presented for the whole documented and undocumented universe with the verb may, and the per-well measurement behind it is not shown.
• Whether the 2026 proposal will be finalized. The comment period closes 2026-08-24 and no final rule had been published at the time of this check.
• What happens to the money that is not yet distributed. About $2.8 billion of the $4.677 billion had not been distributed as of 2025-09-30, and no schedule for the remainder appears in the report beyond a Tribal Phase 3 opportunity estimated for the fourth quarter of fiscal 2026.

See the evidence

The three documents that carry the argument were read as full text rather than through a summary. The Interior report to Congress of November 2025, the BLM bonding factsheet of April 2024 and the Interstate Oil and Gas Compact Commission survey revised in September 2024 were all downloaded as PDFs and their text extracted, so the appropriation split, the well counts, the emissions estimates, the bond minimums and the $71,000 average cost are quoted from the documents themselves. The two Federal Register items were read through the structured document interface, which returned the abstracts word for word.

The remaining rows were read as web pages and are treated as supporting rather than load-bearing. One disagreement survived checking and one dissolved. The documented national count appears as 141,959 in the Interior report and as about 157,000 in the High Country News piece of March 2025, and neither explains the other; the survey table itself prints the 2020 documented total as 92,202 where the Interior report gives 92,203. The undocumented estimate looked like a second disagreement, appearing as 250,000 to 740,000 in the Interior report and as 310,000 to 800,000 on the Department of Energy laboratory page, but opening the underlying survey settled it. Table 2 of that survey carries both ranges as one series at two vintages, 310,000 to 800,000 estimated for 2020 and 250,000 to 740,000 estimated for 2023, and the undated laboratory page gives the earlier pair exactly, so it is quoting an older survey round rather than a rival estimate. The Interior report still supplies the reason spreads of this kind persist, because it states that no comprehensive national inventory exists and that jurisdictions classify wells differently. Figures derived in this dossier by our own arithmetic, namely the $10.1 billion documented-inventory cost, the $27.8 billion to $62.6 billion full-inventory range, the 392,000 to 882,000 plausible inventory, the roughly 65,900 wells the appropriation would buy, the 10,518 total plugged, the 7 percent share plugged, the $61,000 gap between the old bond floor and the average cost, and the roughly $2.8 billion not yet distributed, are sums, multiplications and divisions of the published counts and the published $71,000 unit cost, and are not themselves published anywhere.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below