Cushing, Oklahoma: Why the WTI Delivery Point Matters

Ask where the price of oil is set and most people picture a trading floor or a screen. The physical answer is less glamorous: a cluster of tank farms around Cushing, Oklahoma, a town in Payne County in the middle of the state. Cushing is the delivery point for the NYMEX West Texas Intermediate futures contract, which means the paper market for the world benchmark ultimately settles into real barrels that have to exist, in specific tanks, in three Oklahoma counties. That single design choice explains a surprising amount of oil-market behaviour: why one weekly inventory line gets its own headline, why the front of the WTI curve can detach from the rest of the market, and how the benchmark printed a negative price in April 2020 while oil elsewhere still cost money. This guide covers what the hub physically is, how delivery against the contract works, how much the tanks can actually hold, how the stock level talks to the futures curve, and how to read the EIA weekly Cushing figure without over-reading it. Live inventory context lives on our analysis page and the benchmark itself on the WTI price page; this page explains the plumbing both of them depend on.

1. What Cushing actually is

Start with the statistical definition, because it draws the map. In the footnotes to the EIA Weekly Petroleum Status Report, the Cushing stocks line covers domestic and foreign crude oil held in tank farms in Lincoln, Payne, and Creek counties in Oklahoma, and the same footnote names Cushing as the designated delivery point for NYMEX crude oil futures contracts. So "Cushing" in the data is not one facility and not quite the town limits either. It is a survey boundary around a hub of interconnected tank farms, and the weekly number is the sum of the crude sitting in those tanks on the Friday the survey week ends.

Physically, the hub grew up where pipelines had to cross. Crude produced in the Midcontinent, the Rockies, the Bakken, Western Canada, and later the Permian Basin moves through Cushing on its way to Midcontinent refineries or south to the Gulf Coast refining and export complex. Market coverage of the hub commonly describes nearly two dozen inbound and outbound pipelines with combined capacity above 6.5 million barrels per day, and operator-focused coverage counts 16 major storage terminals at the hub. Treat both figures as reported descriptions of the hub's scale rather than engineering constants: pipelines are reversed, expanded, and idled, and terminal counts shift with ownership. The durable fact is the shape of the network. Many lines in, many lines out, and a large tank farm in the middle that lets barrels change direction, change owner, and wait.

That waiting function is the part newcomers miss. Cushing is not primarily a strategic reserve and it is not where America keeps oil for a rainy day. The Strategic Petroleum Reserve sits in salt caverns on the Gulf Coast and is reported on its own line. Cushing stocks are commercial working barrels: oil owned by producers, refiners, merchants, and trading firms, positioned at the crossroads because that is where it can be redirected fastest. A barrel at Cushing can go to a nearby refinery, batch south to the Gulf Coast, or sit in a tank backing a futures delivery. The hub's value is optionality, and the stock level measures how much of that optionality is currently occupied.

The hub also has a grade function. WTI is a light, sweet crude specification, and the streams that can be delivered against the contract have to meet the contract's quality terms. Tankage at Cushing lets operators segregate streams, blend within specification, and assemble deliverable parcels of the contract size. Storage at a generic location cannot do that job for this contract. Only barrels in the right tanks, accessible to the designated delivery facilities, count when a short has to perform in the delivery month.

2. The contract: why the delivery point is the market

NYMEX WTI futures are physically delivered contracts. The core terms, from the CME Group contract specification, are simple to state and strict in effect. One contract is 1,000 barrels. The price is quoted in US dollars and cents per barrel. Delivery is made free-on-board at any pipeline or storage facility in Cushing, Oklahoma with pipeline access to the designated Cushing storage systems of Enterprise, Enbridge, or Plains. At the buyer's option, delivery can be completed by interfacility transfer (a pumpover into a designated facility), by in-line or in-system transfer of title, or, where the seller agrees and the facility allows it, by in-tank transfer of title without moving the oil. Delivery takes place between the first and last calendar day of the delivery month, and the short has to have its crude available to begin flowing ratably at Cushing by the first day of that month under normal pipeline scheduling practice.

The calendar is where the pressure builds. Trading in a WTI contract month ends three business days before the 25th calendar day of the month before the delivery month (moved earlier when the 25th is not a business day). A participant still holding a position after that termination point is no longer trading a price. The long is committed to accepting 1,000 barrels per contract at Cushing during the delivery month; the short is committed to supplying them. There is no cash-out at the hub. That is deliberate. Physical settlement forces the futures price and the physical price of deliverable crude at Cushing to converge at expiry, which is what makes the contract a benchmark instead of a betting line. The cost of that discipline is that anyone caught long without storage, pipeline capacity, or a resale arranged has a problem measured in days.

In normal times almost nobody has that problem, because almost nobody goes to delivery. The EIA, in its April 2020 review of the negative-price episode, put the usual share at about 1 percent of futures volume reaching physical delivery; the other 99 percent is closed, offset, or rolled into a later month before termination. Financial participants - funds, banks, trading firms without tankage - plan to be in that 99 percent. Their exit depends on someone else wanting the position more than they fear the delivery obligation behind it. When tank space at Cushing is plentiful and cheap, that fear is trivial and the exit is easy. When the tanks are effectively spoken for, the fear is rational, the exit narrows, and the price of the expiring contract becomes the price of escaping a physical commitment. Section 6 walks through the week that logic ran to its limit.

One comparison clarifies the design. ICE Brent, the other global benchmark, is financially settled against an index of North Sea cargo assessments; no Brent futures holder has to find a tank. WTI's physical design ties its price more tightly to real barrels, and it also imports the constraints of one landlocked hub into a global price. Most of the time those constraints are invisible. The rest of this guide is about the times they are not, and about the weekly data that tells you which regime you are in.

3. Storage hub and pipeline crossroads, in one place

Cushing performs three jobs at once, and the stock number means something slightly different in each. First, it is a transit buffer. Pipelines batch crude in slugs separated by grade and owner, refineries take deliveries on their own schedules, and the tank farm absorbs the mismatch. Stocks can rise or fall by a million barrels in a week on scheduling alone. Second, it is a blending and segregation point, where streams are held apart or combined to meet refinery and contract specifications. Third, it is the delivery warehouse for the futures contract, where deliverable parcels wait for the delivery month. A weekly change in the headline number mixes all three flows, which is why a single build or draw is a weak signal on its own and a level sustained over several weeks is a strong one.

The crossroads role also explains why Cushing stocks can fall for reasons that have little to do with US oil demand. When Gulf Coast export economics are strong, barrels get pulled south through Cushing rather than parked in it, and the hub drains even in a well-supplied market. When a major inbound line is down for maintenance, the hub drains for the opposite reason - supply cannot arrive. When Midcontinent refineries return from turnaround, they draw on the nearest tanks first. Each story prints the same downward arrow in the Wednesday report. The framework table in section 5 and the checklist in section 8 exist to separate these cases before the arrow becomes a conclusion.

Geography caps the story. Cushing is landlocked. Barrels leave by pipeline or not at all, and pipeline capacity out of the hub, while large, is fixed in the short run. A coastal hub can clear a glut into tankers when local tanks fill; Cushing cannot improvise an exit. The same constraint works in reverse at the low end: there is no spot cargo market to refill the hub in a hurry. What arrives is what the inbound pipelines deliver on their schedules. That rigidity is exactly why the stock level at this one hub can move the front of the benchmark curve, while a similar swing spread across Gulf Coast tankage usually cannot.

4. Capacity, working capacity, and tank bottoms

Three capacity words get used interchangeably in headlines, and only one of them is the right denominator for the weekly stock figure. Shell capacity is the design volume of a tank or cavern: what it would hold filled to its engineering limit. Working capacity is the usable band inside that shell. The EIA defines working storage capacity as the difference between the maximum safe fill capacity and the quantity below which pump suction is ineffective - the bottoms. Net available shell capacity, the third term in the EIA storage capacity report's title, adjusts further for space already leased or otherwise unavailable to a new customer. For reading the weekly Cushing line, working capacity is the number to carry: stocks are compared against the band the oil can actually move within, not against the steel's theoretical volume.

The hub's working capacity has grown with the shale era. The EIA reported working crude oil storage capacity at Cushing of 61.9 million barrels as of March 31, 2012, against operating shell capacity of 74.6 million barrels, with utilization of 64 percent at that date. By April 2020 the EIA's storage capacity data put working capacity at 76 million barrels. ICE's benchmark comparison, citing EIA data, puts working capacity at 76.3 million barrels in April 2020 and 78.4 million barrels as of March 2024. The lesson in that sequence is not any single figure. It is that the denominator moves - tanks get built, converted, and taken out of service - so a stock level that meant 85 percent full in one year can mean 75 percent in another. Quote a utilization rate without dating the capacity figure behind it and you have manufactured a number.

Tank bottoms deserve a careful treatment because the phrase does real work and gets stretched in commentary. At the level of one tank, the concept is physical: a heel of oil below the suction line, plus the volume needed in lines and manifolds for pumps and transfers to function. The EIA builds that concept straight into its definition of working capacity. What the EIA does not publish is a single hub-wide bottoms number for Cushing, because bottoms belong to individual tanks and systems, not to the survey total. The roughly 20 million barrel figure that appeared across 2026 market coverage - analysts quoted in Reuters-syndicated reporting described operational challenges emerging below about 20 million barrels - is a market rule of thumb for where hub-wide operations feel tight, not an official threshold. It is still useful, held at arm's length: as the survey total approaches the low twenties, a rising share of the remaining barrels sits in individual tank bottoms, in the wrong grade, or in tanks whose outbound line is not the one a buyer needs, so usable, deliverable supply shrinks faster than the headline level suggests.

The symmetry matters. Near the top of working capacity, the scarce asset is empty space, and the uncommitted share of that space is what counts - the EIA made exactly this point in April 2020, noting that some physically unfilled storage may already be leased or committed. Near the bottom, the scarce asset is movable, deliverable oil, and again the headline level overstates what a buyer can actually lift. In the middle of the band, neither constraint binds and the stock number is mostly a transit reading. A useful shorthand: the same 5 million barrel move means different things at 70 million, 45 million, and 24 million barrels. Our own EIA series put Cushing at 24.3 million barrels for the week ending September 25, 2026 - the low end of the band, where the bottoms discussion stops being theoretical.

5. How Cushing stocks talk to the WTI price

The futures curve is the translator between tanks and prices. Our contango vs backwardation explainer covers curve mechanics in full; the Cushing-specific version is shorter. When deliverable stocks at the hub are ample and rising, nobody pays a premium for a barrel today over a barrel next month, and the market pays for storage instead: the front spread sits in contango, roughly covering the cost of tank space, financing, and insurance between the two delivery months. When deliverable stocks are thin and falling, refiners and merchants bid for prompt barrels, the front spread firms into backwardation, and holding oil in the tank carries an opportunity cost instead of earning a carry. Cushing is where that bidding physically happens for WTI, so its stock trend and the WTI front spread usually tell the same story from opposite sides - one in barrels, one in dollars.

The storage trade is the mechanism that keeps the two sides consistent. If the contango on offer exceeds the full cost of storing a barrel at Cushing for the interval, a merchant can buy prompt oil, lease tank space, sell the later month, and lock the difference. That buying fills tanks and supports the prompt price until the spread narrows back toward carrying costs. Run the incentives the other way in backwardation and stocks get drawn down and sold into the prompt premium. Two failure modes break the mechanism, and both are Cushing stories. At the top, the trade needs empty, uncommitted tanks; when none can be leased at workable rates, the spread can blow past any textbook full-carry level because the arbitrage cannot be executed. At the bottom, the trade needs movable barrels; when the remaining stocks sit in tank bottoms or committed parcels, backwardation can sharpen without pulling the supply the price is asking for. The curve is always quoting a storage incentive. The tanks decide whether anyone can collect it.

The worked example below is a labeled hypothetical, built to show the arithmetic a storage desk runs, not a quote of any actual 2026 market rate.

Hypothetical example (illustrative numbers only): Suppose the front WTI contract trades at $78.00 per barrel and the next month at $78.65, a contango of $0.65 for one month. Suppose, hypothetically, that tank space at Cushing can be leased at $0.40 per barrel per month, financing the barrel for a month costs $0.22 at prevailing short-term rates, and insurance and incidentals add $0.03. The all-in carry is $0.65 - exactly the spread on offer. The storage trade earns nothing before fees and operational risk, so a rational desk passes, and you would expect Cushing stocks to drift with transit flows rather than build on storage demand. Now move only the spread: at $0.90 of contango against the same $0.65 carry, the trade locks roughly $0.25 per barrel before costs, tanks fill, and the EIA line should confirm it within a few weeks. Flip the curve instead - prompt at $79.10, next month at $78.65, a backwardation of $0.45 - and the same arithmetic pays the tank owner to empty: every barrel held a month forgoes the $0.45 premium and still pays storage. Real desks run this with actual lease quotes, credit terms, and pipeline tariffs; the point of the example is the comparison, spread versus carry, and the stock movement it predicts. When the predicted movement fails to appear, the constraint is physical - no space at the top, no movable barrels at the bottom - and that failure is itself the signal.

One more channel, quieter than the first two: the WTI-Brent spread. Because WTI settles into a landlocked hub and Brent settles financially against waterborne cargoes, a Cushing-specific surplus or shortage shows up partly as WTI moving against Brent rather than as the whole oil complex moving. Our WTI vs Brent spread explainer works through that relationship with the site's own price history. The reading discipline is the same: if Cushing stocks swing and WTI-Brent swings with them while the Brent curve holds steady, you are looking at Oklahoma plumbing. If both curves move together, the story is bigger than one hub.

Read the table as a diagnostic, not a trading rule. It tells you which question to ask next; it never tells you what the price will do. The next two sections give the table its evidence - first the week the top of the band broke the contract, then the quieter problem at the bottom.

6. The full-tank episode: April 2020

The sequence is documented in the EIA analysis published April 27, 2020 (Barnett and Barron), and the numbers below follow that account. Pandemic restrictions cut consumption sharply. US refinery runs fell to 12.8 million barrels per day in the week ending April 17, down 4.1 million barrels per day, or 24 percent, from a year earlier. Crude that refineries were not taking flowed into storage. At Cushing, the EIA's storage capacity data put working capacity at 76 million barrels. Inventories for the week ending April 17 totalled 60 million barrels, of which about 2 million barrels were in transit by pipeline, water, or rail, leaving 58 million barrels in the tank farms - 76 percent of working capacity. Six weeks earlier that would have read as comfortable. The catch was commitment: part of the physically unfilled space was already leased or otherwise spoken for, so the uncommitted room available to a futures holder without pre-arranged storage was smaller than the headline gap, and priced accordingly.

The May 2020 contract was set to stop trading on April 21. Holders who were long the contract and had no tank, no pipeline arrangement, and no buyer for the physical barrels faced taking delivery at Cushing in May into that storage squeeze. On Monday, April 20, the contract traded in negative dollars per barrel for the first time since WTI futures began trading in 1983, reaching an intraday low of negative $40.32 per barrel in the EIA's account and settling at negative $37.63 - a figure our curve explainer also carries. In plain terms, sellers paid counterparties to take the obligation off their hands. The EIA account stresses the boundary of the event: longer-dated WTI contracts stayed positive, the June Brent contract closed at $19.33 on April 21, and few physical sellers actually paid to sell oil. The negative print was a price for escaping a delivery commitment at a nearly full hub, on one expiry date - which is precisely why the delivery point matters. Change the settlement design to cash, as Brent uses, and that specific failure mode cannot occur. Keep physical settlement, and the tank farm is part of the contract whether traders think about it or not.

Two lessons survived the episode and both are built into the checklist in section 8. First, utilization is about uncommitted space, not gross space: a hub at 76 percent of working capacity can be functionally full for the marginal holder. Second, expiry concentrates risk. The stress appeared in the expiring contract in its final sessions, not across the curve. A reader who watched the whole curve, the Brent comparison, and the storage data together got a plumbing story with a known fix - production shut-ins, refinery recovery, and time. A reader who watched only the front-month print got a headline about worthless oil that the rest of the market did not support.

7. The low-tank problem: when the constraint runs the other way

Full tanks break the contract's exit. Empty tanks strain its operation. The mechanics are less dramatic and easier to misread, because a low Cushing print usually arrives dressed as bullish news: stocks down, premium up, benchmark firm. Sometimes that reading is right - strong refinery demand drawing the hub down is the market working as designed, and the crack spread explainer shows the refinery margins that drive that pull. But the level deserves the same zone discipline as the level in April 2020 got, for three physical reasons.

First, blending and batching need working stock. A hub that segregates grades and assembles deliverable parcels cannot run its tanks to zero any more than a warehouse can ship pallets it does not have. As individual tanks approach their own bottoms, transfers between tanks slow, grades strand in the wrong place, and outbound nominations get harder to fill on schedule. None of that appears in the survey total; all of it appears, eventually, in wider location and quality differentials and a jumpier front month. Second, the futures delivery mechanism itself needs deliverable barrels behind it. Open interest heading into a delivery month against thin deliverable stocks is the mirror of April 2020: this time it is the short who must source physical oil at the hub on a deadline, and the squeeze expresses itself as a spiking prompt premium rather than a negative print. Third, a drained hub has no shock absorber. The transit buffer that normally soaks up a pipeline outage or a refinery upset is the stock itself. At 45 million barrels an operational hiccup is a scheduling problem. At 22 million it is a price event.

The 2026 experience put the hub through that low zone in real time. Our EIA series - the same data behind the weekly report guide - shows Cushing stocks working lower through the year, and market coverage through the summer and fall of 2026 reported prints down to 18.6 million barrels in late July and a recovery range in the low twenties into September, with our archive recording 24.3 million barrels for the week ending September 25, 2026. The drivers reported at the time were a heavy export pull on US crude toward the Gulf Coast and strong refinery demand during a period of disrupted global flows - draws caused by barrels moving through the hub, not by US production failing to arrive. That distinction is the whole reading skill. A draw caused by throughput says the crossroads is busy. A draw caused by inbound supply failure says the crossroads is starved. The Wednesday tables separate the two if you read past the Cushing line: throughput draws come with firm refinery inputs and product supplied, while supply failure comes with falling PADD 2 receipts and production problems upstream. The rig counts page and our rig count explainer cover the upstream side of that check.

A note on precision before moving on. You will see exact capacity figures for Cushing quoted to the decimal in market tools - 76 million, 78.4 million, and others. Each was a real EIA figure for a specific date, and each goes stale as tankage changes. The same caution applies to any single tank-bottoms threshold. Build your reading on the level's direction, its distance from the zone where the last episode bit, and the confirming flows in the same report. That combination stays true even when the exact denominator has moved since the figure you memorized.

8. Reading the EIA weekly Cushing line: a checklist

The EIA publishes Cushing stocks every Wednesday in the Weekly Petroleum Status Report, as part of the crude stock table by PAD District: the hub sits inside PADD 2 (Midwest) and is broken out on its own line because of the delivery-point role covered above. Our guide to reading the weekly report walks the full release table by table. The routine below is the Cushing-specific cut, in the order that prevents the common errors.

1. Take the level and the change together. Note the stock level in millions of barrels and the weekly build or draw, then place the level in its zone against a dated working-capacity figure (section 4). A 1 million barrel draw at 24 million is a different event from the same draw at 45 million.
2. Smooth before you conclude. One week mixes pipeline batching, tank transfers, and survey timing. Line up the last four prints and the four-week average change, the same smoothing discipline the rig-count reading uses. A trend is three or four prints walking the same way.
3. Split Cushing from PADD 2 and from the nation. If Cushing falls while total PADD 2 stocks rise, barrels moved within the Midwest - a logistics story. If Cushing and US commercial stocks fall together, the draw is systemic. The national crude line and the SPR line sit in the same table; read all three before summarizing any one of them.
4. Name the flow that moved it. Check refinery crude inputs and utilization (demand pull), crude production and imports (supply push), and exports (Gulf Coast pull) in the same release. A draw with rising refinery runs is demand doing its job. A draw with flat runs and surging exports is throughput draining the crossroads. A build with falling runs is April 2020 in miniature, whatever the level.
5. Ask the curve whether it agrees. Compare the front spread's direction with the stock trend using the carry test from section 5. Stocks and curve confirming each other is the signal. Stocks drawing while the front spread softens means the prompt scarcity sits somewhere else - or the move reverses next week.
6. Check the calendar. Within two weeks of a contract's trading termination, front-month behaviour increasingly reflects position exits and delivery logistics rather than fresh views on scarcity. Weight the second and third months more heavily, and treat any expiring-contract spike as a question about storage and deliverable supply before treating it as a price opinion.
7. Write the sentence you can defend. The output of the checklist is one sentence naming the level, the zone, the likely flow, and the curve's verdict - for example: "Cushing drew for a third week to the low twenties on firm refinery runs, and the prompt spread firmed with it." If you cannot fill in all four slots from the release, the correct output is what is missing, not a stronger adjective.

Steps 1 to 3 take the data as published. Steps 4 and 5 force the interpretation to survive contact with flows and prices. Step 6 protects you from the expiry calendar, and step 7 is the audit. Run in this order, the Cushing line stops being a weekly mood swing and becomes what it actually is: the inventory gauge on the benchmark's own warehouse.

9. Five common misreads

Misread 1: "Cushing fell, so the US is running out of oil." Cushing is one hub's working stock, not the national cupboard. Total US commercial crude stocks run in the hundreds of millions of barrels and the SPR adds a separate buffer in the same EIA table. A Cushing draw usually means barrels moved - to refineries, to the Gulf Coast, into export programmes - which is the crossroads working. National scarcity is a claim about the national lines and production, not about three counties in Oklahoma.

Misread 2: "Storage is 70 percent full, so 30 percent is available." Three gaps hide in that sentence. Working capacity, not shell capacity, is the denominator. Bottoms sit inside the reported stock figure and cannot be pumped. And uncommitted space is what a new storage customer, or a futures holder facing delivery, can actually lease - the EIA flagged leased and committed space explicitly in April 2020, at 76 percent utilization. Availability is a commercial fact about specific tanks, not an arithmetic remainder.

Misread 3: "The exact capacity is X, so the threshold is Y." Capacity figures carry dates (61.9 million barrels of working capacity in March 2012, 76 million in April 2020, 78.4 million as of March 2024 in the EIA figures ICE cites), and hub-wide bottoms have no official figure at all - the roughly 20 million barrel stress zone is a market rule of thumb from analyst commentary. Precision borrowed from a stale denominator is false precision. Read zone, direction, and flows; let the exact threshold stay approximate, because in the tanks it is.

Misread 4: "April 2020 proved oil can be worthless." It proved that an expiring, physically delivered contract can price the cost of escaping delivery into a full hub. The same week, later WTI months and Brent stayed firmly positive, and the EIA account notes few physical barrels changed hands at negative prices. The episode belongs in the settlement-design file, not the commodity-value file - which is also why it is the single best argument for understanding this page's subject.

Misread 5: "A big Cushing move must mean supply or demand just changed." Start with the plumbing. A reversed pipeline, a refinery turnaround ending, an export programme pulling south, or a batch arriving a day after the survey cut-off all print as million-barrel weekly moves with no change in underlying scarcity. The four-week trend, the PADD 2 split, and the curve check exist to catch exactly this. A move that survives all three checks has earned a fundamental reading. One that fails them has earned a shrug and a calendar note to re-check next Wednesday.

10. Frequently asked questions

What is Cushing, Oklahoma in oil markets?

Cushing is a crude oil storage and pipeline hub in central Oklahoma and the designated delivery point for the NYMEX WTI crude oil futures contract. The EIA defines the Cushing stock figure as domestic and foreign crude held in tank farms in Lincoln, Payne, and Creek counties. Because WTI futures settle into physical barrels at that hub, the weekly Cushing stock level is watched separately from total US crude stocks.

How does WTI futures delivery at Cushing work?

Each NYMEX WTI contract covers 1,000 barrels. Trading in a contract month ends three business days before the 25th calendar day of the month before the delivery month, and delivery takes place between the first and last calendar day of the delivery month. Delivery is made free-on-board at a pipeline or storage facility in Cushing with access to the designated Enterprise, Enbridge, or Plains storage systems, by interfacility transfer, in-line transfer, or in-tank transfer where the facility allows it. Most positions are closed or rolled before expiry; the EIA notes only about 1 percent of futures volume typically goes to physical delivery.

What is the difference between working capacity and shell capacity at Cushing?

Shell capacity is the design volume of the tanks. Working capacity is smaller: the EIA defines it as the volume between the maximum safe fill level and the quantity below which pump suction stops being effective, the tank bottoms. The gap between the two is not usable storage. The EIA put Cushing working capacity at 76 million barrels in April 2020, and later EIA figures summarized by ICE put it at 78.4 million barrels as of March 2024. Capacity changes as tanks are built or retired, so check the date on any capacity figure before using it.

What are tank bottoms and why do traders watch them?

Tank bottoms are the layer of oil that has to stay in a tank and its lines for pumps, suction, and transfers to keep working. The EIA uses the concept to define working capacity, but it does not publish a single hub-wide bottoms number for Cushing. The roughly 20 million barrel level quoted in 2026 market coverage is a trader rule of thumb for where hub-wide operations get tight, not an official EIA threshold. Stocks can print above zero and still be hard to move once individual tanks approach their own minimums, especially when the remaining oil is the wrong grade in the wrong tank.

Why did WTI trade negative in April 2020?

On April 20, 2020, the May WTI contract settled negative for the first time since WTI trading began in 1983, with an intraday low of negative $40.32 per barrel reported by the EIA. Demand had collapsed, refiners had cut runs, and Cushing stocks had reached 60 million barrels against 76 million barrels of working capacity, with much of the remaining space already leased or committed. Holders who could not take or store physical barrels at Cushing paid to exit before expiry. Longer-dated WTI and Brent stayed positive, which is why the episode is read as a delivery-point and expiry problem, not a verdict that oil itself had no value.

How should I read the EIA weekly Cushing number?

Read the level, the weekly change, and the position against working capacity together, then check whether the front of the WTI curve agrees. A build toward capacity with a steepening contango, or a draw toward the low twenties with a firming prompt spread, is a confirming signal. A Cushing move that the curve ignores usually means the barrels moved for logistics reasons - a pipeline batch, a refinery turnaround, an export pull to the Gulf Coast - rather than a change in scarcity. Section 8 of this guide gives the step-by-step checklist, and the analysis page tracks the current EIA series.

11. Related reading

• How to Read the EIA Weekly Petroleum Status Report - the Wednesday release this guide's checklist is cut from: every major table, in reading order, with the Cushing line in context.
• Contango vs Backwardation, Explained - the futures curve mechanics behind section 5: carry, curve shape, and the storage trade that links tanks to spreads.
• WTI vs Brent Spread Explained - how a landlocked delivery point shows up in the spread between the two benchmarks, with the site's own price history.
• WTI Crude Oil Price Today - the benchmark whose delivery warehouse this guide describes, with the current front-month picture.
• Analysis: Energy Data and Market Trends - the running EIA inventory series, including the Cushing line, updated from the weekly report.
• Crude Oil Prices: WTI, Brent, Henry Hub - front-month benchmarks side by side, the starting point for the curve comparison in the checklist.