The Landed Cost of an Imported Robot Battery Pack Beyond the Unit Price

The landed cost of an imported custom lithium-ion pack is the sum of seven categories — unit price, Dangerous Goods packaging, DG freight, customs brokerage and entry fees, duty, inventory carrying cost, and currency and payment-term drag — and it only stays predictable when those categories are modeled as assumptions rather than filled in with freight quotes. Quotes expire. Categories don't. The number you carry into a budget review has to survive a year of rate resets, a fuel surcharge change, a currency move, and at least one revision to the robot's connector or BMS protocol, and the only version that survives all of that is a category-level model with a live quote laid on top as a sanity check, not as the foundation.

Three quotes, three different numbers, none of them the cost

I keep three quotes open on a second monitor when I do this. One pack house quotes FOB. One supplier quotes CIF. One contract manufacturer quotes DDP. The first thing I do, every time, is throw the totals away and rebuild all three into the same shape, because those totals aren't measuring the same thing. The FOB number doesn't include the ocean leg. The CIF number includes freight but not the customs entry. The DDP number includes everything and buries it in a line I can't see or audit. Rank those three numbers against each other and I've learned which Incoterm each supplier prefers. Nothing about which supplier is cheaper.

That's why the model comes before the quote, not after. A quote is a photograph of one lane, one weight, one mode, one week. The categories underneath it are what's still true when the photograph is stale.

The stack, in the order it hits you

Packaging comes first, and it's the one people accidentally buy twice. A custom pack for an AMR or a cobot arm isn't a commodity cell in a retail blister. The box gets designed around the geometry, and if the geometry is awkward you're paying for tooled foam and possibly a custom outer that satisfies the transport spec. The UN 38.3 test summary is a document, but the packaging that carries the pack legally is a physical cost the supplier bears and passes through. When a quote is ex-works and never names the packaging standard, I find that cost later — as a re-pack line on the forwarder's invoice, after the goods are already at the dock and I have no leverage left.

Then freight, where the pack's own design choices set the price.

The state of charge decision

Lithium-ion packs move as Class 9 dangerous goods, and the cheapest lever in the entire stack is decided at the factory, by QC, on the day it ships: state of charge. Cells and batteries moving by air have to be at 30 percent SOC or less — that's a condition of carriage, not a preference — and if the supplier's outgoing test ends at full charge because that's how they ship by ground, then an air shipment means a discharge cycle, a re-test, a re-pack, and a delay. Or it means the forwarder refuses the tender outright. Air also carries tighter quantity limits and a different rule set for cargo aircraft versus passenger aircraft, and it costs multiples of ocean per kilo for Class 9. Ocean is cheaper and slower and puts the pack in a container for weeks. Ground is cheapest and the only mode that's genuinely fast, which is why a domestic build carries a structural freight advantage that has nothing to do with unit price.

Brokerage and duty

Brokerage is a bundle: the entry filing, the customs bond, the importer security filing if it's moving ocean, and the fee lines that ride along with a formal entry. Duty is driven by classification and country of origin. Origin is the one that bites on a custom pack, because a pack assembled in one country from cells manufactured in another may not carry the origin you assumed. That's a question I want answered in writing, with the bill of materials attached, early — not argued about after the entry is filed. I don't try to settle it myself. I make sure somebody qualified has the BOM before the first shipment, because a reclassification after the fact is a cost with no offset.

Where the carrying cost actually lives

Inventory carrying cost is where the arithmetic stops being obvious, and where I've seen the widest gap between a model that looks right and a model that is right. It is not a warehouse rent line. It's the capital tied up in buffer stock, the insurance on that stock, the floor space, the shrink, and the expected loss when a revision makes a pack unusable.

For industrial robot packs, that last piece is not a rounding error. The fleet program will change — a connector, a BMS protocol, a can dimension — at least once before it stabilizes. Every pack sitting in the buffer that doesn't match the new revision is either reworked on your dime or scrapped on your books. A ten-week ocean pipeline means you own ten weeks of that risk, times the buffer quantity, times the probability the revision lands inside the window. That's a real number, and it's the one most landed-cost write-ups leave out, because it doesn't come from a rate sheet.

The hinge: freight premium versus carrying cost

Most of the landed-cost decisions that matter collapse into a single trade. Pay more per kilo to move it fast, or pay less freight and finance the pipeline. Air against ocean is the loud version. Domestic assembly against overseas is the quiet one: local build cuts transit weeks and buffer, but adds labor, overhead, tooling amortization, and often a smaller supplier base with less leverage on cells.

Neither answer is right in general. What I do is put both into the same model and find the crossing point — the buffer length at which the domestic premium stops being a premium. Sometimes it's three weeks. Sometimes there is no crossing point anywhere in the range of volumes I'll ever buy, and then the decision isn't close and I stop pretending it is.

Currency, terms, and the day the number was born

If the PO is denominated in the supplier's currency, I'm exposed from the day I issue it to the day the balance clears. Deposit-and-balance terms mean I'm financing their production run. A letter of credit moves some of that risk and adds bank fees plus a working capital line tied up against it. Quoting in USD pushes the exposure back to the supplier, who prices it in — the hedge isn't free, it's just invisible.

None of that changes the pack. All of it changes what the pack costs by the time it's on my floor. Every landed-cost number has a birth date on it, and a number sitting on a stale currency assumption is already wrong; you just haven't been told yet.

What the model is actually for

I don't ask suppliers or forwarders for rates when I build this. I ask for category-level assumptions and I hold them in bands — low, expected, high — because a single-point estimate hides exactly the thing I need to see. Then I look at the swing. If duty and carrying cost together swing more than freight does, then negotiating freight is theater, and I should spend that meeting on origin documentation and buffer sizing instead. If currency swing dwarfs everything else, the fix is a term in the contract, not a cheaper quote.

The model isn't there to predict the invoice. It's there to tell me which category to attack, and to give me a baseline that catches the charge nobody quoted. I keep a tab for assumptions and a tab for what actually hit the ledger, and I reconcile them the first week of every month. The gap between those two tabs, run over a few quarters, is the most useful document in the program — better than any quote, because it's the only one that includes the lines that showed up after the goods were already on the water.