Short answer: dropshipping is cheaper to test and slower to deliver. Amazon FBA delivers fast and costs more per unit, on a fee schedule that moved twice in 2026 alone. Here is the arithmetic. As of 2026 Amazon charges a referral fee of roughly 8% to 15% depending on category — 15% covers most of them, 20% for jewelry, with a $0.30 minimum per item. On a $30 product that is $4.50 gone before anyone has touched the box. Stack the per-unit fulfillment fee on top, then the 3.5% fuel and logistics surcharge that took effect April 17, 2026 on US and Canada FBA fees, and roughly 25% of revenue is spent before you have paid your supplier a cent. Dropshipping barely registers by comparison on fixed cost — a Shopify Basic plan is $25/month billed monthly, or $19/month billed annually, as of 2026 — but the bill still arrives, in transit time and in customs, because the $800 de minimis exemption that used to wave low-value parcels into the US duty-free no longer applies.
Dropshipping vs Amazon FBA at a glance
| Factor | Dropshipping (own store) | Amazon FBA |
|---|---|---|
| Cash tied up in inventory | $0 — you buy after the customer pays | Full purchase order plus inbound freight, weeks before revenue |
| Fixed monthly cost | $25/month Shopify Basic (or $19/month annual), 2026 rates | $39.99/month Professional plan; Individual plan is $0.99 per item sold |
| Variable fees | Payment processing, plus a 2% Shopify fee on Basic if you use a third-party gateway instead of Shopify Payments | Referral fee (mostly 15%) + fulfillment fee + 3.5% surcharge + storage |
| Typical delivery | 15–30 days on AliExpress standard shipping; 6–12 days from a US-stocked supplier | 1–2 days on Prime |
| Traffic | You buy it or earn it. Nobody browses your store. | Built-in demand, but you compete on the same page as everyone else |
| Returns | Your problem, end to end | Amazon processes them; you eat the fee and the unsellable units |
| Main failure mode | Refund and chargeback spiral from slow shipping | Storage and aged-inventory fees on stock that will not move |
Where each option actually wins
Dropshipping wins when you do not yet know what sells. You can list twenty products, spend $300 on ads, and find out in ten days which two get add-to-carts — with no dead stock at the end. It also wins for bulky, fragile, or seasonal goods where FBA storage would eat you alive, and for anything with a genuinely long sales cycle. If your product is a $90 niche item that sells eleven units a month, FBA’s storage clock is working against you the whole time. One caveat on those test results, though: a 21-day delivery estimate suppresses checkout conversion, so the test tells you reliably which products people want and unreliably what they convert at. Read the ranking, not the percentage.
FBA wins the moment you have proven demand and repeatable margin. Prime badge conversion is the whole argument. If the same product sells at $34.99 with a two-day badge and $29.99 with a three-week estimate, FBA is usually the better business even though the fee load is higher — the extra $5 of price covers a lot of referral fee. It also wins on operational sanity: no packing, no carrier accounts, no 11pm customer service about a tracking number that has not updated in nine days. Two things break that logic. If your gross margin before fees is under about 35%, a 15% referral plus fulfillment plus the surcharge does not leave a business behind. And if the unit tips into an oversize tier, the fulfillment fee jumps enough that a product which pencils out at 12 x 9 x 3 inches does not at 14 x 10 x 5.
Hidden costs nobody quotes you
On the FBA side, the storage schedule is where beginners get surprised. Under the fee schedule effective July 1, 2026, standard-size monthly storage runs $0.87 per cubic foot from January through September and jumps to $2.40 per cubic foot from October through December. Oversize is $0.56 and $1.40 for the same windows. Run that on a real position: 40 cubic feet of stock that stopped selling in August costs $96 in October, $96 in November, and $96 in December — $288 for the privilege of storing something nobody is buying, in the exact months you needed that cash for ads. Then the aged-inventory surcharge kicks in far earlier than most sellers expect. It now triggers once a unit passes about 180 days, at $1.50 per cubic foot per month on top of base storage, and escalates for stock sitting past a year. Note what that does to a normal Q4 plan: inventory you send in during September crosses 180 days in March, so the pallet you over-ordered for the holidays gets charged rent at peak rates and then charged again for being old. Pulling it back out is not free either — removal and disposal orders carry their own per-unit fee.
Two smaller lines belong in the same calculation. Amazon raised fulfillment fees on January 15, 2026 by an average of $0.08 per unit sold, and the 3.5% surcharge stacked on top of fulfillment fees works out to roughly $0.17 per unit on average for US FBA by Amazon’s own estimate. Twenty-five cents. Nobody reprices a SKU over twenty-five cents. At 4,000 units a month that is $1,000 walking out the door, which is most of a full-time hire or a month of ad budget, and it will not show up anywhere in your product-level margin sheet unless you put it there by hand.
On the dropshipping side, the big 2026 cost is customs. The $800 de minimis exemption ended for China and Hong Kong on May 2, 2025 and for all other countries on August 29, 2025, and CBP indefinitely suspended it for non-postal shipments on June 24, 2026. It is scheduled to be statutorily eliminated on July 1, 2027. In practice: every commercial parcel now needs a formal or informal entry and full duty payment regardless of value. If your $12 landed cost from Shenzhen was built on duty-free entry, that model is priced wrong today. Re-quote it with duty and brokerage included before you run another ad. And if a supplier quotes you DDP to make the problem disappear, ask in writing who is named as importer of record — if the answer is you, the duty and any penalty are yours no matter what the quote says.
The mistake that ends accounts
The single most common way people blow up a dropshipping operation on Amazon is the packing slip. Amazon’s dropshipping policy requires you to be the seller of record and to appear as the seller on every invoice, packing slip, and piece of external packaging — and it requires that no other retailer’s or supplier’s branding is visible to the customer. Buying from another retail site and having them ship to your Amazon customer is not allowed, full stop.
How to avoid it: before you list, order one unit from your supplier to your own address and open the box. If a supplier packing slip, a promotional insert, or a retailer-branded label falls out, that supplier is unusable for Amazon until they agree in writing to white-label. Check the shipping label too, not just the contents — the return address and any scannable code on the outer carton are customer-visible. Get that agreement before you scale, not after your first suspension notice, and buy a second test unit in November, because the warehouse that stripped its inserts in June often puts the holiday flyer back in when volume spikes. And do not solve any of this by editing PDFs — doctored documentation is exactly what account-health review looks for.
Which one fits your situation
Three checks, in order of what actually kills people. Cash first: if you cannot put $3,000–$5,000 into inventory and freight and then wait 6–10 weeks to see it back, dropship until you can — running FBA on money you need next month is how sellers end up liquidating good inventory at a loss. Velocity second: if a SKU cannot plausibly clear its stock within 180 days, FBA’s aged-inventory clock will erase your margin, so keep it out of the warehouse and fulfill it yourself. Plan choice third: Amazon’s Individual plan at $0.99 per item beats the $39.99/month Professional plan below about 40 units a month and loses above it, so there is no reason to pay a subscription during validation — with one string attached, which is that the Individual plan is not eligible for the Featured Offer, so confirm what that costs you in your category before you save the $39.
The honest answer for most people reading this is sequential, not either-or: dropship to find the winner, then move that single winner — not the catalog — into FBA.
A checklist you can run today
- Pull your exact fulfillment fee for your top SKU from Amazon’s FBA Revenue Calculator, not from a blog table. Size tier and price band both change the number, and a half-inch of packaging can move you a tier.
- Add 3.5% to that fulfillment fee. That is the current surcharge, and older calculators and third-party spreadsheets do not show it.
- Re-quote your landed cost with duty included. Ask your freight forwarder or customs broker for the HTS code and duty rate on your product, and get the code in writing — guessing it wrong is a repricing problem at best and a penalty at worst.
- Order one unit from your supplier to yourself. Photograph everything in the box and the outer label. Confirm nothing identifies the supplier.
- Check your inventory age report and flag anything past 150 days. That gives you roughly a month to discount, bundle, or remove it before the aged-inventory surcharge starts.
- Confirm current card processing rates for your plan and region on the Shopify pricing page. They vary by country, and if Shopify Payments is not supported where you sell, the 2% third-party gateway fee on Basic applies on top — enough to move your margin by 1–2 points on its own.
- Write down the unit economics on one line: sale price minus product cost minus duty minus platform fees minus ad cost. If that line is under 15%, fix pricing before fixing fulfillment. No warehouse decision rescues a product that is priced wrong.
Summary
Dropshipping and FBA are not competing philosophies, they are different risk positions. Dropshipping converts inventory risk into delivery risk and customer-service load. FBA converts delivery risk into inventory risk and a fee schedule that changes at least once a year — twice in 2026 alone, in January and April, with a new storage schedule in July on top. That is the part worth internalizing: the model you priced in March may not be the model you are running in October, and nobody emails you when the spread closes. Verify every fee against the current Seller Central schedule and your own broker before you commit capital, because the numbers above are accurate as of 2026 and will not stay that way.