Amazon Seller Fees in 2026: A Margin-First Guide

If your Amazon margin feels thinner this year even though sales held up, you are not imagining it, and it is probably not your advertising. It is usually fees. The awkward part of 2026 is that no single fee jumped enough to make you stop and re-open the spreadsheet, so nothing forced the issue while the total quietly climbed. That is the trap: the erosion is real, but no one number is alarming enough to trigger a look.

This guide is not a rate card to memorize. Seller Central already lists your exact bands. It reads the 2026 schedule the way an operator does: as a running subtraction from the margin you keep, after Amazon and the ad account both take their cut. The aim is to help you see that subtraction per SKU, and catch the erosion before it shows up as a bad month you cannot explain.

At a high level, your 2026 fees fall into four buckets: a referral fee on every sale (roughly 8% to 15% of price, set by category), an FBA fulfillment fee per unit (a few dollars, set by size and weight), monthly storage on the space your inventory occupies, and a set of surcharges that fire when your stock moves too slowly or too fast. Two changes drive most of the 2026 pressure: FBA fulfillment rose about $0.08 per unit on average in January, and on April 17 Amazon added a 3.5% fuel and inflation surcharge on top of FBA fees, worth roughly $0.15 to $0.35 per unit. None of these looks big on its own. Stacked against a single sale, and multiplied across a catalog, they decide whether a SKU is worth advertising at all.

Quick Answer

Amazon seller fees in 2026 come in four categories: referral fees (8% to 15% of item price, varying by category), FBA fulfillment fees (per-unit, based on size and weight), storage fees (monthly, based on cubic feet and season), and conditional surcharges (low-inventory, aged-inventory, and placement fees). Two 2026 changes matter most for margin: the average FBA fulfillment increase of about $0.08 per unit in January, and the 3.5% fuel and inflation surcharge applied to FBA fees starting April 17, 2026, worth roughly $0.15 to $0.35 per unit. Referral percentages held steady for most categories. The practical effect is that contribution margin, what remains after every variable cost including ad spend, is thinner than most 2025 spreadsheets assume. The fix is not memorizing rates. It is watching the margin per SKU as fees and prices move, so you catch erosion before it shows up in a bad month.

Who This Is For

• Operators running FBA who set their own prices and manage their own advertising, and who suspect their per-unit economics have drifted since last year.
• Brand and marketplace managers who report on profitability and need to explain why revenue held but margin slipped.
• Anyone deciding which SKUs deserve ad spend in 2026 and which are now too thin to defend.

If you want a rate card to bookmark, Seller Central's fee pages and the FBA revenue calculator are the source of record for your exact bands. This guide is for the layer above that: what the rates do to the number you keep.

The 2026 Fee Schedule, Defined Plainly

Every fee below is a variable cost: it scales with units sold or space used, so it comes straight out of contribution margin. Here is how each behaves in 2026.

Referral fee. Amazon's commission on the sale, charged as a percentage of the total price (item plus shipping). Most categories sit between 8% and 15%. Consumer electronics, computers, and cell phone devices run near the low end at 8%; many consumer categories sit at 15%; a few accessory categories run higher still. For 2026, referral percentages held steady for the large majority of categories. This is the single largest fee on most orders, and because it is a percentage, it moves every time your price moves.

FBA fulfillment fee. The per-unit charge for Amazon picking, packing, and shipping the order. It is set by the unit's size tier and weight, not by its price. In January 2026, fulfillment fees rose about $0.08 per unit on average, by Amazon's own 2026 fee update, though the exact change varies by size tier. Small standard items pay a few dollars; large or heavy items pay considerably more. This fee is flat per unit, which is why it hurts cheap products most: $0.08 is nothing on a $60 item and a real bite on an $8 one.

Fuel and inflation surcharge. New pressure for 2026. Effective April 17, 2026, Amazon applied a 3.5% surcharge on FBA fees, adding roughly $0.15 to $0.35 per unit depending on tier. It rides on top of the fulfillment fee, so it compounds the January increase rather than replacing it.

Monthly storage fee. Charged on the average daily cubic-foot volume your inventory occupies. Standard-size storage is cheaper for most of the year and spikes in Q4 (October through December), when warehouse space is scarce. Oversize is charged on its own, lower per-cubic-foot rate. Storage is the fee most sellers underweight, because it is billed to the account rather than to the order, so it never appears on a per-unit sale.

Low-inventory-level fee. A per-unit surcharge that fires when you hold fewer than about 28 days of supply relative to demand. In 2026 Amazon began assessing it at the individual FNSKU level and extended it to bulky items, so it is easier to trip than it was. The push is toward steady cover: too little stock triggers this fee.

Aged-inventory surcharge. The other side of the same pressure. Units sitting long-term (the surcharge scales up as inventory ages past roughly six months and again past a year) pay an escalating per-cubic-foot charge on top of base storage. Hold too long and this compounds with normal storage.

Adjacent 2026 cost movers. Two more items belong in any 2026 margin model even though they are not line-item selling fees. Amazon Warehousing and Distribution (AWD) storage and transport rose roughly 19% to 22% in 2026, which raises the landed cost of upstream inventory. And on March 12, 2026, seller payouts moved to seven days after delivery under Amazon's Delivery Date Based Reserve policy, which does not change the fee but stretches working capital and changes how quickly you can reorder.

2025 Baseline vs. 2026 Reality: Where the Money Moved

That storage row is the unflattering one, and it is worth sitting with: it is the fee most likely to be missing from your per-SKU math entirely, because it never lands on an order. A tidy per-unit P&L can look healthy while the storage bill quietly eats the quarter. In 2026, no single fee moved enough to notice, and that is exactly the problem.

How Fees Erode Contribution Margin Below the Ad Account

Contribution margin is the number that matters here. Per unit, it is price minus every variable cost that unit incurs: COGS, the referral fee, the FBA fulfillment fee and its surcharge, an allocated share of storage, expected returns, and the ad spend attributed to that sale. What is left contributes toward fixed costs and profit.

Most sellers watch the top of that stack closely. They know their price, their COGS, and their ad account. ACoS is the number in front of them every day, and in 2026 it is not comfortable: average ACoS ran about 32.5% in January 2026, on an average CPC of about $1.18. So the ad line gets managed hard.

The fees, though, sit below the ad account in almost everyone's mental model, and that is where they do their damage unwatched. Here is the sequence:

• 1. A referral fee rides on price, so a price cut to defend rank cuts your margin twice: once on the price and again on the smaller-but-still-percentage fee.
• 2. The FBA increase and the April surcharge take their per-unit bite before ads are even counted.
• 3. Storage and aged-inventory charges land on the account, not the order, so per-SKU dashboards show margin that the storage bill later erases.
• 4. Only after all of that does ad spend come out. By the time ACoS is subtracting, the unit had less margin to give than the ad model assumed.

The result is the pain every operator eventually names: revenue is fine, ACoS looks managed, and margin is still leaking somewhere below the ad account. It leaks because the erosion happened before the ad account ever ran, and nothing in the campaign view shows it. Your ad account can be perfectly efficient against a contribution margin that no longer exists.

Common Mistakes

• Modeling margin on 2025 fees. The January increase and the April surcharge are small individually and easy to leave out of a spreadsheet built last year. Across a catalog they are not small.
• Leaving storage out of per-SKU math. Because storage is billed to the account, sellers compute a clean per-unit margin and never subtract the space cost. Slow movers look profitable right up until the storage and aged-inventory charges arrive.
• Cutting price to hold rank without re-checking the fee. The referral fee is a percentage, so a price drop shrinks the fee in dollars but not in rate. Operators feel like they gave up only the price, when they also reshaped the whole margin equation.
• Managing ACoS in isolation. A target ACoS set against last year's margin can be technically hit while the SKU loses money, because the margin underneath the target shrank.
• Ignoring the payout change. Seven-day-post-delivery payouts (March 2026) tighten cash. Sellers who reorder on old cash-flow assumptions can trip the low-inventory fee, adding cost while trying to save it.
• Treating referral, FBA, storage, and ads as separate reports. They are one subtraction from the same unit. Read apart, they each look fine.

Where a Manual Margin Pass Runs Out

A careful operator can absolutely do this by hand, and should know how, because the manual pass is what teaches you the shape of your own economics. Here is what a good one looks like.

Take one SKU. Pull its current price. Subtract COGS. Look up its category referral percentage and subtract that. Find its size tier in the FBA calculator, subtract the fulfillment fee, then add the 3.5% surcharge on that fee. Estimate its share of monthly storage from its cubic footage and its rate for the current season. Add an allowance for returns. Subtract the ad spend attributed to it. What remains is contribution margin. Do it well and you will know, for that SKU, on that day, exactly where you stand.

A general-purpose AI assistant can speed the arithmetic. Hand it the rate card and the inputs and it will compute the stack faster than you will, and it will not fat-finger the surcharge. That is real help.

Here is where both run out. The manual pass and the general-purpose helper are both snapshots. They are true for one SKU, at one price, under one week's storage season, against one period's ACoS. Amazon's schedule is not a snapshot. The surcharge landed in April. Storage reprices by season. Aged-inventory charges escalate the longer a unit sits. Your prices move to defend rank, and every move rewrites the referral fee in dollars. Your ad efficiency drifts with CPC. To keep a manual pass current, you would have to re-run it for every SKU every time any one of those inputs moved, which for a real catalog is every day, forever.

That is the break point. The problem is not that the math is hard. It is that the math is never done. A number that is correct on Monday and unwatched by Friday is how the leak stays hidden: everyone did the analysis once, and no one did it again. The hard part of margin is not calculating it, it is that it changes faster than anyone re-checks it.

Where Trellis' Qinetix Dynamic Pricing Fits

Once the break point is clear, the requirement is clear too: the margin math has to stay current on its own, and price has to be allowed to move within limits you set, not limits Amazon's fee schedule sets for you by default.

That is the level Trellis' Qinetix Dynamic Pricing works at. It is not a system that reprices your catalog on your behalf and hopes you approve. It is operator-set guardrails plus execution inside them. You set the condition, for example an inventory-cover threshold below which a SKU should not chase price down, or a floor price that already accounts for the 2026 fee stack including the surcharge. The system then executes price moves inside those limits, and it holds the line you drew rather than one it inferred. The mechanism is your judgment, made continuous. When the referral fee reshapes at a new price, or storage flips to its Q4 rate, the guardrail is still reading the current fee stack instead of last quarter's.

Pricing and advertising run as parallel mechanisms here, with shared visibility and no coordination between them. The pricing guardrail is not secretly steering your bids, and your bids are not moving your prices. They see the same margin picture so you can read them together, but each stays in its own lane. That is deliberate: coordinating them automatically would hide exactly the trade-off you most need to see.

For the repeated read underneath, the analysis that answers "which SKUs are leaking, and where," Qore is the companion. Qore is judgment and analysis codified: the margin pass you would run by hand, written down once and run on a schedule, so the snapshot becomes a standing view. Its actions are approval-gated, so it surfaces the leak and proposes the move, and you decide.

The honest limit: seeing the margin leak is not the same as closing it. A guardrail keeps price from drifting past a line you set, and a scheduled analysis keeps the leak visible, but neither renegotiates your COGS, reverses a surcharge, or decides for you that a thin SKU should be discontinued. Those are still operator calls. As of July 2026, Qore is on a public waitlist, and its actions default to manual approval. The tools keep the number in front of you and defensible. You still run the business.

A Concrete Pass

• Input: A standard-size SKU priced at $19.99, COGS $6.40, category referral 15%, an FBA fulfillment fee around $4 plus the 3.5% surcharge, a small per-unit storage allocation, and roughly $1.90 of attributed ad spend at the current ACoS.
• Workflow: The scheduled analysis re-runs the full stack in 2026 terms, including the April surcharge and the current storage season, and flags that contribution margin has fallen under the operator's set floor after the SKU's price was cut $2 last month to hold rank.
• Expected output: A flag that this SKU is now contributing below its guardrail, with the driver named: the price cut plus the surcharge, not the ad account, moved it under the line.
• Business decision: The operator lets the pricing guardrail hold the floor rather than chase rank lower, trims ad spend on the SKU to match its real margin, and decides whether to reorder at a negotiated COGS or let it wind down. The move is theirs; the visibility was standing there when they needed it.

Conclusion

The 2026 Amazon fee schedule did not deliver one dramatic number to react to. It delivered a slightly heavier fulfillment fee, a new surcharge riding on top of it, the same seasonal storage pressure, and steady referral percentages that still move in dollars every time you touch price. Each is small. Together they thin the contribution margin your ad account is spending against, and they do it below the line where most operators are looking.

The durable fix is not a better rate card. It is keeping the margin math current instead of correct-once, holding price inside guardrails you set with the real 2026 fees baked in, and reading pricing and advertising against the same margin picture. Do that and the leak stops being a mystery you discover at month-end. It becomes a number you can see, and a line you chose to hold.