FBA vs FBM Unit Economics: An Operator's AI Cost Model for 2026

The FBA vs FBM debate is the single most expensive decision most Amazon brands get wrong. The default reflex — "FBA for everything because Prime badge" — has been mathematically incorrect for a meaningful slice of the typical catalog for at least three years, and the 2024 and 2025 FBA fee changes have widened the gap further. Brands that route every SKU through FBA without modeling the unit economics are leaving 4 to 9 percent of gross margin on the floor across their slow-velocity and oversize SKUs, and they almost never see it because the fee leakage hides inside aggregate reports.

This is the operator's cost model we use on managed accounts to make the FBA-vs-FBM call SKU by SKU. We will walk every line in the unit-economics breakdown, surface the 2026 FBA fee changes that matter, explain when the hybrid posture wins, and show how AI compresses what used to be a quarterly spreadsheet exercise into a continuous decision loop.

The Full FBA Cost Stack

Most brands think about FBA fees as "the FBA fee" — one line item. The reality is that FBA carries six distinct cost categories, and the 2026 fee schedule has shifted the weighting between them.

1. Referral Fee

Standard 15% on most categories, with category-specific exceptions. Same for FBA and FBM — this is platform access, not fulfillment. Often forgotten when operators are mentally pricing a "shift to FBM" because the referral fee does not change.

2. FBA Fulfillment Fee

Per-unit fee based on size tier and weight. The headline FBA cost. As of the 2024 simplification and 2025 inbound placement changes, this fee runs from roughly $3.06 for the smallest standard size tier up to $14+ for large oversize. The 2026 schedule held most tiers flat from 2025 but increased the apparel-specific surcharges and reintroduced a low-velocity surcharge for SKUs sitting in fulfillment centers without sales for 30+ days.

3. Storage Fees (Two Variants)

Monthly inventory storage fees by cubic foot, with a Q4 surcharge that roughly triples the per-cubic-foot rate from October through December. Plus aged inventory surcharges that escalate at 181, 271, and 365 days of storage. For slow-moving SKUs, storage fees are often the dominant cost line — not the fulfillment fee.

4. Inbound Placement Service Fees

Introduced in 2024 and refined in 2025 to 2026, these fees apply when you ship inventory to a single FBA location and Amazon distributes it across the network. The "minimal shipment splits" option is cheaper but requires you to ship to multiple locations yourself. For most brands without their own distribution capability, this is a meaningful per-unit cost that did not exist three years ago.

5. Returns Processing Fees

FBA charges a returns processing fee for SKUs in apparel and a few other categories with high return rates, applied per unit returned. The fee runs roughly equal to the original FBA fulfillment fee, which means a 25% return rate effectively raises FBA fulfillment cost on those SKUs by 25%.

6. Removal and Disposal Fees

When you decide to pull dead inventory out of FBA, removal fees apply per unit. Most brands underestimate how often they end up paying these — on slow-moving long-tail SKUs, the cumulative removal fee over 18 months can rival the original landed cost of the inventory.

The Full FBM Cost Stack

FBM looks simpler on paper but carries its own hidden cost structure once you account for everything Amazon does for you under FBA.

• 3PL or in-house pick-and-pack cost. Per-unit, often $2.50 to $4.00 for standard size with reasonable volume.
• Outbound shipping cost. Your negotiated carrier rate, which is almost always meaningfully higher than Amazon's network rate — often $5 to $9 for standard 3-5 day, more for 2-day.
• Storage fee at 3PL or warehouse. Per pallet per month, scaling with volume.
• Customer service overhead. Real money. FBA orders go through Amazon CS for most issues. FBM orders mean your team handles "where is my order," "I want a refund," and "the tracking shows delivered but I never got it." On managed accounts we estimate this at $0.50 to $1.50 per order all-in for established teams.
• Returns handling. Per-unit return processing, plus inbound shipping cost on returned units.
• Seller-Fulfilled Prime requirements (if running SFP). SLA penalties, performance metrics, weekend pickup requirements that cost real money.
• Ad performance penalty. The hidden line. FBM SKUs without Prime badge convert 15 to 30 percent worse than the same SKU under FBA at the same price point. This shows up as required CPC increases on ads to maintain the same ROAS, or as outright lost organic conversion.

The 2026 FBA Fee Changes Worth Flagging

Three changes in the 2026 fee structure shift the FBA-vs-FBM calculus meaningfully and are worth explicitly modeling:

Low-velocity surcharge expansion. SKUs that sit in FBA without sales for 30+ days now incur a small per-unit surcharge that compounds with aged inventory fees. For brands carrying a long tail of slow SKUs, this surcharge alone can flip the FBA-vs-FBM math on those SKUs.

Q4 storage surcharge widening. The October-through-December storage premium has been extended slightly and the multiplier has crept up. For oversize SKUs, Q4 FBA storage on slow-moving inventory has reached the point where many operators pay more in Q4 storage than the unit margin.

Apparel returns surcharge increase. Apparel categories saw the returns processing fee structure tighten, which materially raises effective FBA cost on apparel SKUs with above-category-average return rates.

Where AI Models the Break-Even

The break-even calculation is genuinely difficult because the costs interact non-linearly. AI models that run the full unit-economics decomposition per SKU, projected over a 12-month volume forecast, surface break-even ASINs that spreadsheet models miss. The variables that matter:

• SKU velocity (units per month).
• Size tier and weight band.
• Return rate.
• Conversion rate differential between FBA and FBM (category-dependent).
• Seasonality — SKUs that spike in Q4 may justify FBA carry-cost in Q1 to Q3 to capture Q4 demand.
• Substitution risk — if competitors are FBA on the same keyword cluster, FBM penalty is steeper.

The output of a properly built model is a per-SKU profitability comparison, with explicit break-even thresholds. We see the same general pattern emerge across managed accounts: roughly 60 to 70 percent of SKUs are clearly FBA-positive, 15 to 25 percent are clearly FBM-positive, and the remaining slice sits in a margin-meaningful gray zone that the model needs to evaluate continuously. Our deeper coverage of FBA fee structure lives in FBA fee optimization with AI, and the broader margin context is in how AI improves Amazon profit margins.

The Hybrid Posture: When It Wins

The right answer for most multi-SKU brands is not "FBA for everything" or "FBM for everything" — it is a SKU-level hybrid. The pattern that consistently outperforms on managed accounts:

• Top 20 to 30 SKUs by velocity: FBA. Prime badge, fast shipping, conversion advantage all justify the fee stack at high velocity.
• Oversize SKUs with low velocity: FBM. The combination of high fulfillment fees and Q4 storage surcharges on oversize destroys margin on slow-velocity oversize SKUs. FBM almost always wins here.
• High-return-rate SKUs (apparel, certain consumer electronics): often FBM, especially after the 2026 returns surcharge changes. The math gets close once you fold in returns processing.
• Long-tail SKUs at the bottom of the catalog: FBM, almost always. Velocity is too low to amortize FBA fixed costs, and the low-velocity surcharge punishes them further.
• New product launches: FBA for the launch period, with re-evaluation at month 4 once velocity is established.

The Operating Cadence That Captures This

The hybrid posture only works if you re-evaluate it continuously. We run a quarterly review on managed accounts where every SKU goes through the model, and roughly 5 to 10 percent of SKUs change fulfillment posture each quarter based on velocity shifts, fee schedule updates, or seasonal positioning. Brands that set their FBA-vs-FBM posture once at launch and never revisit it are systematically running the wrong fulfillment mix as their velocity distribution drifts.

The closely related decision — whether to use Amazon Multi-Channel Fulfillment to handle off-Amazon orders from your FBA inventory — is covered in our piece on Amazon MCF when to use it. The two analyses share variables, and brands considering a fulfillment overhaul should run them together rather than separately.

The bigger point: FBA-vs-FBM is not an ideological question, and treating it as one costs real money. It is an empirical question with a SKU-specific answer that changes as fees, velocity, and competitive posture shift. Brands that build the operating discipline to answer it continuously capture the margin that everyone else leaks.