What Defense Demand Instruments Fail to Buy

Three defense battery demand instruments are visible in the public record, and none of them pays a cell manufacturer to keep a qualified configuration unchanged.

The DIBC's BES-26-01 solicitation seeks domestic cylindrical cell production capacity. 10 USC §4817 authorizes purchase commitments of up to ten years for critical materials including batteries. Packet Digital holds roughly $27M and $9.8M in Navy SBIR Phase III orders for pouch cells and advanced battery packs. These are the most concrete demand signals the defense battery market has produced, and they share the same omission.

A cell factory's post-commissioning economics run on continuous process improvement. Manufacturers swap precursor suppliers for cost or availability, adjust coating speeds and mixing parameters and drying profiles, revise electrolyte formulations, and rebalance formation protocols as production data accumulates. Individually these are small changes. Together they are how unit economics improve after the capital is sunk.

Defense qualification requires freezing those variables. The Army's first-article testing guidance is explicit that changes in battery design, materials, manufacturing process, or production facility trigger a new first-article test. A manufacturer holding a defense-qualified baseline is a manufacturer electing, for the duration of the hold, not to improve its cost position on that line. That election has a price, and the price grows with time.

What each instrument actually buys

BES-26-01 buys a prototype production line and a capacity pathway. The RPP asks for 18650 and 21700 lithium-ion cells, initial capacity of at least 50 MWh per year within two years, and a path to 3 GWh. It requires supply-chain reporting — supplier identities, part numbers, relationship descriptions — and sets a greater-than-95% target for electrode active material value from FEOC-compliant sources.

Those provisions generate traceability inputs. They do not define a controlled cell baseline. The RPP and its Q&A do not specify a configuration-management plan or a named cell revision. There is no approved-source-list hold, no engineering-change-notice requirement, and no government approval right over material or process substitutions. The Q&A permits respondents to propose offtake or other financial instruments but discloses no committed quantity, and no configuration for demand to attach to. The solicitation also asks for cross-format production flexibility and a path to COTS cost parity by 2030, both of which presume continued manufacturing development. Nothing in the public documents addresses how an approved defense configuration would be insulated from that development.

Section 4817 supplies statutory authority, not terms. The Secretary may determine quantities, conditions, and advance payments, for periods up to ten years. The statute prescribes no minimum purchase floor or take-or-pay structure. It contains no configuration stability requirement, no change-control protocol, and no requalification terms. Any of those could be negotiated into a specific commitment. None arrives as a default.

The more immediate problem is that no battery-cell or pack purchase commitment under §4817 has surfaced in the public record through September 2026. The FY2026 mandatory allocation plan names $10 million for "Energy Storage and Batteries," directed to analytics, coordinated investments, and data management, and not described as a purchase commitment or a cell procurement. The first §4817 annual report to the congressional defense committees is due October 15, 2026. It may provide visibility; the statute does not require it to be public. A ten-year permitted duration is a ceiling on what could be authorized, and there is no disclosed exercise of that authority for batteries at any meaningful scale.

Packet Digital's Navy orders fund pilot prototypes, raw materials, pilot and low-rate production, multiple pack designs, pack testing, and delivery through 2027–2028. The award summaries identify no controlled cell part number, chemistry, formulation, production line, or approved precursor sources. They disclose no process window, no quantity in MWh, and no annual delivery floor. Configuration controls may well exist in the non-public statements of work. The record a manufacturer can actually read before deciding whether to bid shows pilot-scale funding with no configuration-hold provisions disclosed.

What configuration hold forecloses

The public evidence will not support a single annual percentage for the cost of holding a cell configuration stable. It does support the conclusion that the cost is economically material and that it compounds.

A 2026 electrode-processing study modeled a 10 GWh LFP cylindrical-cell plant and estimated that individual process improvements — mixing time reduction, coating speed increases, continuous mixing, dry-electrode processing — could each cut annual production costs by roughly 0.8% to 2.0%. The largest modeled case was about $22 million per year from mixing-time optimization alone. These are modeled savings from adopting an improved process state rather than demonstrated recurring annual reductions, and they are competing alternatives rather than additive gains. They do establish the scale of what sits inside the electrode-processing cost pool, which the same study puts at more than 20% of total production cost.

Argonne's BatPaC sensitivity analysis adds a second data point: in a modeled 50 GWh plant, a 10% change in cell yield moves total pack cost by roughly 7–8%. The sensitivity is steep because rejected finished cells already carry full materials and manufacturing cost. A manufacturer under configuration hold cannot chase yield improvements that require process or material changes breaching the qualified baseline.

Korean cell makers describe ongoing optimization in terms consistent with this evidence, without isolating what it contributes financially on a mature line. SK Innovation's Q2 2026 earnings commentary listed supplier diversification, value engineering, yield and loss improvement, inventory management, and automation as continuing cost-reduction activities. LG Energy Solution's technology roadmap describes iterating on separator coatings, cathode surface treatments, material doping, and electrolyte chemistry across product generations. LGES projects that its planned dry-electrode process could reduce electrode-manufacturing costs by 17–30% after targeted commercialization in 2028.

Map those disclosures onto the Army's requalification triggers and the overlap is close to total. LGES's separator roadmap describes three generations of ceramic coating, each a materials change. SK Innovation's supplier diversification substitutes one precursor source for another, also a materials change. A coating-speed increase that alters dried electrode porosity or thickness distribution is a manufacturing-process change. Each category appears in the FAT guidance as grounds for a new first-article test. A manufacturer holding a defense-qualified baseline either forgoes these changes or absorbs a requalification cycle every time one is ready.

What no reviewed disclosure provides is a defensible count of how many such changes a major cell maker executes per line per year. The record establishes that optimization is continuous and spans multiple FAT-triggering categories; it does not establish frequency. Frequency is the multiplier on the configuration-hold cost, and it is not public.

The exposure compounds. Six months of hold costs a manufacturer one or two rounds of process improvement. Three years of hold means watching commercial lines iterate through several sequential optimizations, each building on the prior quarter's gains, while the defense-qualified configuration sits at its original process state. Total foregone value over a multi-year qualification period is the sum of that accumulating series, not a flat annual rate applied to the hold duration.

Volume and stability are separate exposures

In Issue #11, this publication argued that the absence of recurring volume in defense battery procurement amounts to an unpaid option held by the government: capacity reservation without a purchase commitment. The finding here is adjacent but distinct. Committed volume does not by itself cover the cost of configuration stability.

Take a hypothetical ten-year §4817 commitment for 5 MWh per year of a specific 21700 cell. That guarantees volume. It does not compensate the manufacturer for holding that cell's electrode formulation, precursor sources, coating parameters, and formation protocol unchanged while its commercial lines improve around it. These are two separate economic exposures, and the current instruments address neither.

Issue #12 argued that qualification, provenance, and acceptance records must converge on the same controlled cell revision before a configuration is releasable under Section 842 compliance. That converged revision is precisely what the manufacturer is being asked to hold, which is why the compliance requirement and the optimization penalty are the same problem seen from two directions.

What §4817 permits but no instrument yet provides

The authority to price configuration hold already exists. Section 4817 gives the Secretary broad discretion over quantities, terms, and conditions for commitments of up to ten years. At least three mechanisms fit inside that discretion.

A configuration-hold premium — a periodic payment for maintaining a qualified baseline without change — is the most direct. Its periodic structure matches the cost's behavior: because the foregone value accumulates, a recurring payment tied to hold duration tracks the exposure better than a one-time fee. It requires a defined baseline, a change-control protocol, and a payment schedule, none of which is difficult to draft and none of which currently exists.

A change-fund escrow sets aside money to cover requalification when a change eventually becomes unavoidable. Under current arrangements that cost falls entirely on the manufacturer.

A guaranteed requalification pathway — a government commitment to fund and schedule first-article testing for an updated configuration within a defined window after the manufacturer proposes a change — shortens the hold by giving the manufacturer a predictable way out of the frozen state.

Any of these fits within the statute, which permits the Secretary to set terms and conditions, including advance payments, for periods up to ten years. Breadth of authority is not exercise of it. As of September 2026, none of these mechanisms appears in the public terms of BES-26-01, the Packet Digital awards, or any disclosed §4817 action. What the three instruments buy is a capacity pathway, pilot production, and an authorization that has not been used. The cost of keeping a qualified cell unchanged while the rest of the factory improves around it is still carried by the seller.