Follow the Tariff Down the BOM

The Question the Last Article Left Open

A reader asked me something after the last piece that I didn't have a clean answer to: the 44% North American premium you're quoting — how much of that is actually the tariff, and how much would exist anyway because building batteries here costs more than building them in China?

That's the right question. BNEF's 44% figure is a composite of everything — tariffs, structural manufacturing cost differences, lower domestic production volumes, logistics premiums, and the cost of financing capital equipment in higher-rate environments. BNEF doesn't break the gap into its components. The 44% is the output, not the analysis.

So let's do the analysis. This is a bill-of-materials teardown of a Chinese LFP battery pack, with tariff rates applied to each imported line item. If you're offended by looking at a BOM, you came to the wrong publication.

What's Actually in a $84/kWh Pack

A lithium-iron-phosphate battery pack at the Chinese commercial price of $84/kWh breaks down roughly as follows: cells account for about 68% of total pack cost, with pack integration hardware — battery management system, thermal management, structural housing, wiring, and assembly — making up the remaining 32%.

Within the cell cost, the cathode active material (LFP iron phosphate compound) is the largest single line item at roughly 32% of cell cost. Graphite for the anode is about 10%. Separator, electrolyte, current collectors, and cell casing account for most of the rest. Cell manufacturing labor and overhead — the actual cost of running the production line — is approximately 30% of cell cost and carries no tariff exposure when the cell is built domestically.

That last point matters. When an American manufacturer builds cells domestically using Chinese-sourced inputs, they pay tariffs on the materials but not on the labor. When they import completed cells from China, they pay tariffs on the entire cell value — labor included. These are two different cost structures and two different tariff exposures. I'll show both.

Scenario A: Importing Complete Cells

If a North American pack assembler sources complete LFP cells from a Chinese manufacturer — which is still common for OEMs that don't have domestic cell capacity — the tariff exposure is applied to the entire cell value.

Cell cost represents 68% of the $84/kWh pack price, or $57.12/kWh. Section 301 alone applies a 25% tariff to that value: $57.12 × 0.25 = $14.28/kWh, or $1,071 on a 75kWh pack. That is the confirmed floor. Additional reciprocal tariffs enacted in 2025 sit on top of the Section 301 rate, but those are subject to ongoing adjustment and HTS classification disputes — so I'm treating Section 301 as the sourced minimum.

So at minimum: importing the cell content of a 75kWh pack from China costs $1,071 in tariffs before the BMS, housing, or thermal management is even considered. Against the $2,775 total per-vehicle premium, that single line item — cells only, Section 301 floor — accounts for 38.6% of the entire gap.

Scenario B: Domestic Cell Assembly with Chinese Inputs

This is where most North American battery manufacturers actually operate today — or are trying to. You build the cells here, but you source cathode material, graphite, separator, and electrolyte largely from China because the domestic and allied-nation supply chains for those materials either don't exist at scale or cost significantly more. The tariff hits each input separately, at its own rate.

Here is every major line item with its tariff exposure applied to the Chinese cost baseline:

Against the $2,775 confirmed per-vehicle premium between North America and China on a 75kWh pack, the BOM analysis produces this breakdown:

The range is $1,407 to $1,574 of tariff-attributable cost per vehicle — a floor of 50.7% to a ceiling of 56.7% of the total premium. The structural gap — what it would cost more to build here even with zero tariffs — is approximately $1,200 to $1,370 per vehicle.

The rough answer to the reader's question is: it's approximately half and half, with tariffs accounting for slightly more than half at current rates. That's not a comfortable number for either side of the policy debate. It means the tariff is real and large enough to matter. It also means that eliminating every tariff tomorrow would still leave a $1,200+ structural cost gap that doesn't go away without domestic manufacturing at scale.

The Graphite Number Deserves Its Own Section

Pull graphite out of the table for a moment and look at it in isolation. This is a single raw material — a form of carbon used as the anode in lithium-ion cells. It is not a sophisticated component. It is not a finished assembly. It is a processed mineral.

Its pre-tariff cost in a 75kWh pack is approximately $428. Its landed cost under current tariff policy is approximately $1,114. The tariff itself — the pure policy cost — is $685 per vehicle.

That $685 represents 24.7% of the entire $2,775 per-vehicle premium. One material. One tariff decision. A quarter of the gap.

The landed graphite costs 2.6 times what it costs at the Chinese factory gate. No domestic substitute currently exists at automotive battery specification. Synthetic graphite produced outside China is available but costs two to three times more than Chinese natural graphite on a pre-tariff basis — meaning the domestic cost disadvantage exists independent of the tariff. The tariff then inflates the Chinese import price, which makes the gap between domestic and Chinese graphite appear narrower than it actually is — masking the underlying competitiveness problem rather than solving it.

What makes graphite different from the cathode material or the separator is the tariff rate. CAM and separators are at 25% Section 301. Graphite is at approximately 160% combined. That asymmetry is a policy choice, and it's not obviously connected to the strategic logic being used to justify it. If the goal is to develop domestic graphite capacity, a 160% tariff with no funded domestic capacity plan is a windfall tax on every EV built in the United States until domestic graphite at specification becomes available — which, on current investment trajectories, is not imminent.

What This Analysis Does and Doesn't Tell You

It tells you the tariff is the larger half of the problem — at the confirmed tariff floor (Section 301 rates on most inputs, plus graphite's 160% combined rate), slightly more than half of the $2,775 premium is policy-attributable. If reciprocal tariffs above the Section 301 floor are fully applied across all components, the tariff share of the premium is likely 55–60%. The structural cost gap — what manufacturing here costs above and beyond Chinese costs — is real, persistent, and roughly $1,200–$1,400 per vehicle at current production volumes.

What it doesn't tell you is the exact dollar figure for any specific OEM's actual pack. Every manufacturer has a different BOM, different supplier mix, different chemistry, and different production scale. The numbers here are derived from the Argonne BatPaC cost model and the BNEF pack price baseline — they represent the industry, not any specific program. Any engineer at Ford, GM, or Stellantis can tell you exactly what their graphite line costs. I'm working from published models.

What the analysis does establish is the order of magnitude and the relative contribution of each component. The cathode active material carries the largest confirmed tariff dollar value by line item — $343 per vehicle at 25%. Graphite, despite being a smaller input by cost, carries a disproportionately higher rate, making it the most policy-sensitive number in the analysis — $685 per vehicle at 160%. The structural cost gap is not going away regardless of what happens to the tariff. And together, these factors tell you that the path to a competitive North American EV battery doesn't have a single lever — it requires tariff rationalization, domestic material investment, and volume growth all moving in the same direction simultaneously.

The Policy Arithmetic

If you rationalize the graphite tariff — not eliminate it, just bring it down to the same Section 301 rate applied to cathode and separator — you recover approximately $500 per vehicle of the premium immediately, without touching a single manufacturing cost, without building a single domestic plant. That's real money. On a two-car household, it's $1,000. On a 200,000-unit production run, it's $100 million in cost structure that disappears from the OEM's BOM overnight.

That doesn't happen in isolation. The tariff revenue disappears too, and whatever political logic underwrote the 160% rate has to be revisited. But that's the trade — and it's a knowable, quantifiable trade. This is what it costs. This is what you get back.

The structural $1,200–$1,400 gap is harder. That gap closes through volume — as domestic battery production scales, per-unit fixed costs fall — and through supply chain development, as domestic cathode and separator manufacturing matures. The Inflation Reduction Act (IRA)'s investment tax credits for domestic battery manufacturing were designed to accelerate exactly that curve. Gutting them while maintaining a 160% graphite tariff is not a coherent industrial policy. It removes the mechanism for closing the structural gap while preserving the policy cost that makes the tariff gap worse.

An automotive engineer would call this a system with two failure modes being introduced simultaneously. You don't do that to a production line, and you shouldn't do it to a supply chain either.

The Bottom Line

The $2,775 per-vehicle North American battery premium is approximately half tariff and half structural cost. The tariff half responds to policy — it can be changed, rationalized, or phased. The structural half responds to investment, volume, and time.

Graphite alone accounts for $685 of the premium — a quarter of the entire gap — at a tariff rate that has no peer among battery inputs and no funded domestic substitution program to justify it.

The two-car household paying $5,550 in hidden battery premium is paying roughly $2,800 in tariff-attributable cost and $2,750 in the structural cost of an industry that hasn't been given the conditions to scale. Both of those numbers are a policy failure. They just require different solutions.

The BOM doesn't lie. It just requires someone to read it.