Section 4865's 95% Threshold Moves With a Cost Method Nobody Has Written

The January 1, 2028 deadline for Section 4865 to apply to new acquisition programs, across all covered battery types, is 484 days out. DFARS Case 2024-D011, which would tell suppliers how to demonstrate compliance with the 95% functional-component-cost threshold, missed its June 16, 2026 statutory deadline. A Federal Register search returns zero documents. The DPCAP class-deviation register contains no §4865 entry. No proposed text, interim guidance, or cost-method selection has appeared.

Supplier activity has continued underneath the unresolved test: Sila's conditional $1.4B loan commitment from the Office of Strategic Capital, Packet Digital's ~$37M in Navy LRIP orders, DIBC's closed-but-unawarded 3 GWh solicitation, and LG Energy Solution's reported early-stage defense inquiries. None of these involves a functional-component-cost submission, because no submission format exists.

"The Cost Method Nobody Wrote" in Issue #12 established that §4865's 95% threshold lacks a valuation convention, a measurement date, and a refresh frequency. The consequence, working from cell cost structure, is that the ratio is price-exposed. A cell whose bill of materials never changes can sit on either side of 95% depending on which convention the rule eventually adopts and what lithium is doing on the day the file is built.

The measurement problem comes before the supplier count

Standard mandate-gap math counts qualified domestic suppliers against known programs and finds the number small. For §4865-covered batteries it is low single digits, and none of them has filed a functional-component-cost representation. That math still holds. But for the 2028 date, something upstream of it binds first: the representation has no defined structure to file into.

The statute enumerates the component set — cathode and anode active materials, separator, electrolyte, anode foil, additives, solvents, internal safety devices — and requires that more than 95% of their cost originate from non-FEOC sources. It does not define cost. It fixes no accounting period, prescribes no allocation method, and incorporates no existing federal cost framework by reference. Several frameworks are available, and they produce different numbers for the same physical cell.

FAR permits several conventions, and they disagree

FAR Part 31 governs cost analysis for contract pricing and allows actual purchase cost, standard cost adjusted for variances, moving-average cost, weighted-average cost, and any consistently applied equitable method for material drawn from stores (FAR 31.001, 31.205-26). Spoilage is included in material cost; scrap credits reduce it. CAS layers on further variation, since coverage turns on contract type, dollar thresholds, and contractor size (48 CFR Part 9903). A large cell maker and a small domestic supplier on the same program can therefore face different cost-accounting obligations, and where CAS 407 applies it imposes its own discipline on how material costs are established and revised. The closest textual analogue to what §4865 needs is the Buy American Act's component-cost definition, acquisition cost plus inbound transportation, but it lives in FAR Part 25 and §4865 does not reach for it.

Any of these could end up in the DFARS rule.

Cathode dominates the denominator, and carries the volatility

Two peer-reviewed cell cost models bound the range. A 2170 NMC811 model (40 GWh plant, 2021 material prices, $107.88/kWh total) puts cathode materials at 54.8% of cell material cost. A 4680 NMC model (10 GWh plant, non-Chinese production, $89.96/kWh) allocates 62.7% to cathode coating. Anode materials come second at 11–21%, separator at 9.5–12.9%, electrolyte at 6.1–7.4%. Neither model isolates every §4865-enumerated category; additives, solvents, and internal safety devices sit inside broader line items or go unreported.

The IEA reports lithium prices at the start of 2026 were more than double their level a year earlier while still running roughly 70% below the 2022 peak. Inventory method, hedging, and contract structure all sit between spot lithium and a cell maker's cathode invoice, so the pass-through is damped. It is not eliminated, and that is the swing the 95% calculation has to absorb.

Take a cost file with 95.2 units of non-FEOC functional-component cost and 4.8 units of FEOC cost: compliant at 95.2%. Assume the non-FEOC cathode line is 60 of those 95.2 units, consistent with the models above. Drop that line by 20%, whether because lithium fell or because the convention shifted from actual purchase cost to a lower standard cost, and non-FEOC total goes to 83.2, denominator to 88.0, ratio to roughly 94.5%. The cell is physically unchanged — every gram traces to the same facilities — but the accounting moved it across the threshold.

Three other variables move the same number by different routes.

Yield losses. The 4680 study reports early-stage scrap at 15–30%, settling near 10% after several years. FAR 31.205-26 includes reasonable spoilage in material cost and credits scrap recoveries. Whether §4865 allocates yield losses by component origin, spreads them across output, or excludes them is unspecified, and the spread between those treatments is widest on new lines, which is where scrap is highest and where most compliant domestic capacity will be in 2028.

Transfer pricing. A vertically integrated cell maker buying cathode precursor from a non-FEOC affiliate can book that input at incurred cost or at the affiliate's commercial price; FAR permits both under stated conditions. Korean makers pursuing defense qualification sit squarely in this case. LGES and Samsung SDI source precursor through affiliate networks with varying upstream FEOC exposure, and the transfer-price convention moves both the numerator and the denominator.

Measurement timing. The same cell costed in Q1 on rising lithium and in Q3 after a correction produces two different ratios. The statute names no measurement date and no refresh frequency.

Supplier signals accumulating against an undefined test

Sila, Packet Digital, DIBC, and the LGES inquiries are all building toward a test that has no arithmetic yet. The LGES item in particular is reported early-stage interest, not active qualification and not an award; treat it as positioning until a program or volume attaches to it. Each of these suppliers could source every functional component from non-FEOC origins and still be unable to produce the number, because producing it requires deciding which costs count, on which convention, measured when, with yield and intercompany transfers treated how.

For 2028, the constraint worth tracking is the cost file rather than the production line. Capacity is the visible problem and the easier one to forecast.

Other deadlines

January 1, 2029 — standard batteries. Last significant signal: the June 16, 2026 expiration of the DFARS revision deadline (Case 2024-D011) without proposed text. Nothing since. Same cost-method gap, twelve additional months of runway, and no additional clarity available absent the rule. Gap assessment: static.

January 30, 2031 — existing programs. Same June 2026 expiration, same absence. Programs on this date will need to start qualification and cost-evidence work years ahead of it and will inherit the same unresolved method on a longer timeline. Longer is not the same as more certain. Gap assessment: static.