FBA vs FBM in 2026: Which One Actually Keeps More Per Unit?

I have shipped orders both ways for a decade, and I still see this argued as though it were a personality test. It is not. FBA versus FBM is a per-SKU arithmetic problem, and the answer flips inside the same catalog — in mine it is close to a 50/50 split, and two products have switched sides this year alone because Amazon and USPS both changed their pricing.

What changed in 2026 (and why old comparisons are wrong)

If you are reading an FBA vs FBM breakdown written before this spring, the numbers in it are stale in four separate ways. This is not a small drift — between them these changes moved my own decisions on two SKUs.

• January 1: Amazon stopped prepping and labeling inventory in US fulfillment centers. Inbound defect fees rose from a few cents to a range of roughly $0.32–$5.72 per unit, so prep is now your cost or your prep partner's, either way a real line item on the FBA side.
• January 15: fulfillment fees were restructured into three price bands — under $10, $10–$50, and over $50. Amazon described the change as averaging $0.08 per unit. The average hides the spread: roughly +$0.12 in the lowest band, +$0.25 in the middle, and +$0.51 for items over $50. Two identical products at different prices now pay different fulfillment fees.
• April 17: a 3.5% fuel and logistics surcharge went onto every FBA fulfillment fee — about $0.16–$0.18 on a typical standard-size unit. It is calculated on the fee, not the sale price, and it does not touch storage or referral fees. Amazon called it temporary and attached no end date.
• July 12: this one is on the FBM side and it is the one nobody has priced in yet. USPS eliminated the 4 oz and 8 oz commercial pricing tiers. Every package under a pound now bills at the old 12–15.99 oz rate. If you self-ship light items, your cost per order went up this month without you touching anything.

That last change matters more than its size suggests, because light self-shipped items were the category where FBM's margin advantage was thinnest to begin with. For a lot of sellers, July quietly moved the break-even line up.

The true cost of FBA

The mistake I made for years was treating the pick-and-pack fee as the FBA cost. It is one of six. The honest FBA cost for a SKU is:

Fulfillment fee + fuel surcharge + storage allocation + inbound freight + prep + an aged-inventory reserve if sell-through is slow.

Here is the 2026 fulfillment rate card for the sizes most sellers actually ship, in the $10–$50 price band. Add 3.5% to every figure for the fuel surcharge, and note that items over $50 pay roughly $0.26–$0.51 more per unit while items under $10 qualify for Low-Price FBA rates averaging about $0.86 less.

Billed weight is where people get hurt. Small standard bills on actual weight, but everything from large standard upward bills on the greater of actual or dimensional weight — length × width × height in inches, divided by 139. A light but boxy product routinely bills two or three bands above what the scale says. If you have never checked a SKU's billed weight against its actual weight, that is the highest-value ten minutes available to you today.

Then storage, which is where the slow movers die. Standard-size runs $0.78 per cubic foot per month from January through September and $2.40 in October through December — a 3x peak multiplier. The aged-inventory surcharge now starts at 181 days, not the 271 or 365 you will still read on older blogs, and past a year it reaches about $6.90 per cubic foot on top of the base rate. Amazon also charges in the opposite direction: drop below roughly 28 days of supply (about 35 for high-velocity SKUs in 2026) and the low-inventory-level fee applies to every outbound unit. You are being steered into a 1–6 month window and penalized on both sides of it.

The true cost of FBM

FBM feels cheaper because Amazon sends you no invoice for it. The costs are all still there, just paid to other people, and the two that sellers systematically leave out are labor and returns.

Two notes on that table. First, postage rose twice this year — a 7.8% commercial increase on January 18 and an additional 8% temporary surcharge from April 26 running through January 2027 — so if your FBM model is built on last year's rates it is optimistic by roughly a sixth. Second, price your own labor honestly. If you pay yourself nothing, you will conclude FBM is free, and you will scale a business that cannot survive hiring your own replacement. I use $18/hour because that is what it would cost me to hire the task out.

Above roughly 200 orders a month, a 3PL usually beats in-house FBM on both cost and sanity. Below that, in-house is defensible if you are already in the room. On rate shopping: I have never found a reason to pay for postage software at low volume — commercial rates through a free platform like Pirate Ship get you most of the way, and the deeper mechanics are in our guide to cutting shipping cost per order.

Side by side on the same product

Abstractions do not decide anything. Here is the same $24.99 kitchen item — my own category — modeled both ways, at 1 lb billed weight shipping to Zone 4, at 40 units a month.

FBA wins by $5.24 a unit here and it is not close — and that is before any Prime conversion effect. This is the shape of the answer for most light, steady-selling products, and it is why "FBA fees are killing me" is usually a pricing or packaging problem rather than a fulfillment-method problem. Change the inputs to a 4 lb item that sells 8 units a month and the table inverts just as decisively. Run yours in the Amazon FBA calculator rather than trusting mine — your COGS, weight and zone mix are not my COGS, weight and zone mix.

The Prime premium, and what SFP now costs you

Prime eligibility is the variable FBM math consistently undercounts. On a competitive listing the badge is worth a meaningful conversion lift — I model 1.35x for non-commodity items and less where buyers are purely price-driven — and the lift compounds: better conversion improves organic rank, better rank lowers the ad spend needed to hold visibility, lower ad spend raises net. Any comparison that treats conversion as constant across the two methods is answering a question nobody asked.

Seller Fulfilled Prime is the way to keep the badge while shipping yourself, and in 2026 the bar went up. You need on-time delivery of 93.5% or better, valid tracking at 99%, an order defect rate under 1%, a pre-fulfillment cancel rate under 0.5%, and a late shipment rate under 4%. Then, as of July 6, 2026, standard-size items must display a one-day delivery date on 40% of Prime page views (up from 30%) and two-day on 75%. Read that as a geography requirement rather than a performance one: hitting one-day coverage on 40% of the country from a single warehouse is not realistic, so SFP now effectively means multiple fulfillment locations. For most one-warehouse sellers, SFP is off the table and the honest choice is FBA-with-badge or FBM-without.

When FBM wins

FBM is the right answer when the item is heavy or oversize, where FBA's tiers escalate past what a regional carrier charges for the same box. It wins on slow sell-through, where storage and the 181-day surcharge do the damage rather than the fulfillment fee. It wins when you have a 3PL at sub-$3 all-in pick costs with tight SLAs, and it is the only option for handmade or made-to-order goods, for hazmat Amazon restricts, and for oversize dimensional outliers now facing the $17–$25 Overmax surcharge. It is also worth a look on high-ASP products, where the Prime lift is smaller in proportion and the buyer is deliberating rather than impulse-buying.

When FBA wins

FBA wins on small, light, fast-moving SKUs — fulfillment fee under about $4, 30+ units a month so storage per unit stays trivial, in a category competitive enough that Prime visibility moves rank. The July USPS change widened this zone: sub-pound items that used to ship cheaply now bill at the 12–15.99 oz rate, and Amazon's bundled fee looks better by comparison than it did in June.

FBA is rarely the cheapest fulfillment. It is often the highest-net.

The other half of FBA's case never shows up in a spreadsheet: customer service, returns processing and SLA management all transfer to Amazon. FBM's real hidden cost is operational complexity, and it compounds with volume in a way that per-unit math does not capture until the week you are packing boxes at 11pm instead of sourcing your next product.

The hybrid model most experienced sellers actually run

The split I use, and the one I see in most mature catalogs: FBA for the top-selling, lightweight, high-margin SKUs where the Prime badge earns its fee, and FBM — usually via a 3PL — for heavy, bulky, or low-velocity inventory that would sit in a fulfillment center collecting surcharges. It captures the conversion premium exactly where it pays and avoids the storage penalties exactly where they bite. There is no prize for consistency across a catalog.

Whichever way a SKU lands, revisit it. Amazon has repriced fulfillment twice this year and USPS three times. A decision that was correct in January is not automatically correct now, which is the whole reason I keep the calculator open next to Seller Central.

Fee figures reviewed July 2026 against Amazon's published 2026 US rate card and USPS commercial pricing. Rates change and vary by size tier, price band, category and zone — confirm your own SKU in Seller Central's Fee Preview report and your carrier's current rate table before pricing. I am a seller, not an accountant.