US Oil Reserves Explained: How Long Would They Last?

The United States held 45,951 million barrels of proved crude oil and lease condensate reserves at the end of 2024, which is 9.6 years of that year’s production. The ratio is a bookkeeping quotient, not a countdown: it has hovered between 9.2 and 12.5 years for eleven straight years while production rose by half, because the EIA counts a barrel as a reserve only once it is economic, planned and reasonably certain.

Behind that count sits a much bigger number. EIA models 310.0 billion barrels of technically recoverable US crude, of which proved reserves are about one part in seven. Which of the two a headline is quoting changes the answer from under ten years to about 65, and neither is a forecast.

Three numbers called reserves

Searches for the size of US oil reserves surface three different stocks, and a flow that is used to size them. Only the first stock is what the EIA means when it uses the word.

The proved figure and the TRR figure describe the same resource at different stages of knowledge. At January 1, 2024 the EIA counted 46.4 billion barrels as proved and 310.0 billion as technically recoverable in total, so proved reserves were 15.0% of the larger number. The SPR belongs to a different category entirely, because it is inventory, not resource; the SPR levels guide tracks it week by week.

What the word proved requires

The EIA does not invent the definition. It follows the one in the Securities and Exchange Commission’s Regulation S-X, rule 4-10(a), which companies use in their annual filings. Five phrases in that rule do most of the work, and each one explains something about why the reserves numbers behave as they do.

The practical consequence of the price rule shows up in the year-by-year numbers later in this guide. The EIA’s report cites WTI at Cushing at $77.58 for 2023 and $76.63 for 2024, a 1 percent decline, which match the annual averages of the EIA spot series. The five-year rule explains why the reserves report distinguishes between proved producing and proved nonproducing volumes, which is where the largest hidden judgement sits.

Eleven years of the ledger

Reserves accounting is an inventory ledger. Start with last year’s stock, add the year’s extensions and discoveries, add the net of acquisitions and divestitures, add or subtract revisions, subtract production, and read the new stock. The EIA publishes every line since 2014, so the replacement question can be answered directly.

A replacement ratio above 1.00 means the stock grew. It did in 6 of the eleven years, most strongly in 2017 at 2.99 and 2021 at 2.50. It fell below zero in 2015 (-0.37) and 2020 (-1.16), the two years in which revisions alone removed 4,493 and 8,382 million barrels. In 2023 and 2024 the ratio was 0.60 and 0.90, which is why reserves slipped from a 2022 peak of 48,321 million barrels to 45,951 million even though drilling continued at a high rate.

Production is the part that grew

Production in this table rose from 3,200 million barrels in 2014 to 4,806 million barrels in 2024, a 50% increase. Reserves rose less, from 39,933 to 45,951 million barrels, an increase of 15%. The reserves-to-production ratio fell from 12.5 to 9.6 years because output rose faster than the stock did. That is the accurate reading of the trend: the US is pumping through its booked reserves faster and booking new ones at nearly the same pace.

Why about ten years is the wrong reading

The ratio in the last column is the figure most often quoted as how long US oil will last, and it fails as a forecast for three reasons that the ledger itself shows.

First, the stock is replenished continuously. Reserves are booked in the year a well is planned and drilled, so each year’s production is partly replaced by that year’s extensions and discoveries. Over the eleven years in Table 3, additions averaged 4,846 million barrels a year against average production of 3,989. A ratio calculated from a replenishing stock describes the present book, not a time of exhaustion.

Second, the denominator moves. At the 2014 production rate the 2024 stock would be 14.4 years; at the 2024 rate it is 9.6. A ratio of 9.6 would hold for exactly 9.6 years only if production, prices and drilling plans stayed fixed, and none of them has stayed fixed in any year of the table.

Third, the numerator is a category. Replacing it with the technically recoverable resource gives 65 years at the 2024 rate, but the TRR is the EIA’s own modeled estimate, uncertain by its own description. The assumptions document says that estimates “often change over time” as drilling reveals more about the rock, and adds that cumulative production in the projection can exceed the TRR shown because the model lets technology raise it. So neither ratio is wrong; each answers a narrower question than the one searchers ask.

What moves the number: price, plans and a balancing item

A common explanation is that reserves follow the oil price, since the SEC rule ties economic producibility to a 12-month average price. The data support that for the biggest swings and not for the small ones.

The two largest negative revisions, 8,382 million barrels in 2020 and 4,493 in 2015, fall in two of the three lowest-priced years in the table, $39.16 and $48.66; the third, 2016 at $43.29, saw revisions of only -262 million barrels. The link is loose in the middle of the range: 2017 revisions were +3,464 million barrels at WTI $50.80, while 2019 revisions were -2,976 at the higher $57.00. And the last two years break the pattern: revisions were -3,898 and -3,845 million barrels in 2023 and 2024 with WTI at $77.58 and $76.63, higher than in any year from 2015 to 2021. Because EIA labels the revisions column an annual balancing item, a reader cannot tell from the table alone how much of that was engineering judgement and how much was reclassification, which could include undeveloped locations leaving the book when drilling plans changed. The EIA’s own summary for 2024 shows where the changes landed: reserves fell by 529 million barrels in Texas and 487 million in North Dakota, while New Mexico grew by 496 million.

Where the proved barrels are

Reserves are even more concentrated than production. Two states hold more than half of the US total, and the reserves-to-production ratio varies from seven years offshore to twenty in Alaska.

Texas and New Mexico together held 26,712 million barrels, 58% of the total, which fits the Permian Basin’s role in the production-by-state ranking. The ratio column is a map of capital and time. Federal offshore areas have the shortest ratio on the table, 7.0 years, with 40% of the volume still nonproducing. Alaska’s 20.3 years is a large reserve base divided by a modest production rate. The Permian states sit in the middle, at 9.6 years for both Texas and New Mexico, which is consistent with a play where wells are drilled continuously and each new batch adds booked locations.

Three states, line by line

The national table hides how differently each region moves. The same ledger exists for every state, and three of them tell the 2024 story. Each starts with a different stock, loses roughly its production to the pump, and then depends on what is added.

New Mexico replaced its 739 million barrels of production 1.67 times over, mostly through 1,201 million barrels of extensions and discoveries, which are the barrels booked when wells are drilled or planned in new locations of known formations. North Dakota booked 349 million barrels of extensions and 756 of acquisitions, and still lost 892 million barrels to revisions, the largest negative revision of any state except Texas. Texas booked the biggest additions in absolute terms, 2,332 million barrels of extensions and 2,224 of acquisitions, and still fell, because 2,762 million barrels of revisions outweighed them. A reader who knows only that a state's reserves rose or fell cannot tell whether that came from the drill bit, a purchase or a re-estimate; the ledger can.

The acquisitions and divestitures lines also cancel at the national level: US divestitures of 1,100 million barrels and acquisitions of 4,135 leave a net of about 3,036 million barrels, the single line in Table 3 (the published acquisition and divestiture lines differ from it by one million barrels of rounding). Oil that changes hands between companies does not change the country's total except through the estimates of the buyer and the seller. The ledger therefore separates ownership from geology better than any headline can.

How the survey is built

The reserves numbers come from a survey, not a census. The EIA collects them on Form EIA-23L from a sample of operators of oil and natural gas fields and estimates the share held by operators outside the sample. For the 2024 report it received responses from 385 of 398 sampled operators, which it describes as covering about 95 percent of proved oil reserves and 96 percent of proved gas reserves nationally. The EIA redesigned the form for report year 2022: it now collects combined crude oil and lease condensate instead of separate categories, no longer asks respondents to split reserves by reservoir type, and collects state and subdivision tables in place of county data. The redesign matters to anyone comparing a pre-2022 number with a later one, and to anyone looking for a county-level answer, which the current report no longer provides.

Operators estimate their own reserves under the SEC definition summarized in Table 2, so the totals inherit each company's engineering and its planning calendar. That is also why the figures differ from other tallies. Texas, for example, publishes its own oil-well and condensate production figures, and the production-by-state guide shows how large the gap between two official Texas counts can be. The same care applies to reserves: a number with a source, a year and a definition is comparable; a number without them is not.

Proved, but not yet producing

The last two columns of the state table hide the single most important sentence in the reserves data: 37% of US proved oil reserves are nonproducing. Of 45,951 million barrels, 28,843 million were classified as proved producing and 17,108 million as nonproducing, a category made up mostly of undrilled locations the operator plans to drill within five years, plus wells that need further work before they flow.

In the Lower 48, 16,164 million barrels were nonproducing, and 11,180 million of those, 69%, were in shale plays. New Mexico is the starkest case, with 49% of its proved reserves not yet flowing. Reading those barrels as stored oil would be a mistake: they are mostly a statement about a development plan, subject to the five-year rule in Table 2 and to prices. If the plan changes, the barrels leave the book without a drop of oil having been produced.

The shale plays

The EIA reports shale separately. Proved reserves from shale plays reached 27,533 million barrels, 60% of the US total, after rising 5% from 26,229 million at the end of 2023. Production from those plays was 2,994 million barrels in 2024.

The table separates the stories that a single national number blends. The Permian alone added 1,995 million barrels of proved reserves in 2024 while the Bakken lost 237 and the Eagle Ford lost 634. A national figure that falls 1.0% is the net of a large gain in one basin and a decline in others. Those declines are consistent with the dynamics in the rig count and drilled-but-uncompleted wells guides, where activity moved toward the Permian, but the reserves table does not itself assign causes.

The same ledger for natural gas

The EIA publishes the natural gas ledger beside the oil one, in billion cubic feet of wet gas after lease separation. Its ratio is higher and more variable, which shows how much the answer depends on the commodity chosen.

Gas reserves ended 2024 at 583,942 billion cubic feet, a 3.3% decline, which is 13.8 years of that year’s production. The range across the table runs from 11.1 years in 2015 to 17.0 in 2022, wider than oil’s because gas reserves moved with the 64,538 billion cubic foot drop of 2015 and the 152,088 billion cubic foot jump of 2021. For what a decline in gas reserves means for the price of gas, the Henry Hub guide and the storage report guide cover the market side; the reserves figure is a slow-moving book number that rarely drives a daily price. For what reserves leave out, the production versus consumption guide covers imports and exports and the API gravity and sulfur guide covers the quality of the oil.

Data and sources: EIA, U.S. Crude Oil and Natural Gas Proved Reserves, Year-end 2024 (released April 7, 2026; Tables 1, 2, 5, 6, 7 and 12 from ARR_2024_TABLES_ALL.xlsx), checked October 23, 2026. The survey is Form EIA-23L; for 2024 EIA received responses from 385 of 398 sampled operators covering about 95 percent of oil reserves. EIA, Assumptions to the Annual Energy Outlook 2026: Hydrocarbon Supply Module (April 2026, Table 1 technically recoverable resources as of January 1, 2024), checked October 23, 2026. SEC, 17 CFR 210.4-10 (Regulation S-X definitions of proved, probable and possible reserves, reasonable certainty and undeveloped reserves), checked October 23, 2026. EIA WTI Cushing spot price, annual (RWTC), from the EIA petroleum bulk file updated October 22, 2026, checked October 23, 2026. EIA weekly SPR stocks (WCSSTUS1) as reported in the SPR levels guide. Ratios, shares, replacement ratios and totals are this page’s arithmetic from the published tables.