What V2X PQC migration actually costs: a field-notes TCO

The cost of V2X post-quantum migration is the question every OEM architect we work with eventually has to answer for their CFO. The standards work was finalised in 2024. NIST published FIPS 203, 204, and 205 in August. UNECE R155 went mandatory across 64 contracting parties on 1 July. The deployment question is now an engineering question, and the engineering question is now a budgeting question.

A note on what follows. Every engagement Digital North signs is bound by an NDA naming a specific scope of work, and we honour that obligation on our own marketing surface the same way we honour it inside the work. The cost ranges below are composite, synthesised across the engagements we have run rather than reported from any single one. The numbers are representative of the patterns we see consistently, not budgets from a specific deployment. The shapes are real. The specific values have been generalised to ranges rather than any client's exact figures. Where the standards anchor a number (signature sizes from FIPS 204, frame rates from SAE J2735), the number is exact. Everywhere else, what follows is the pattern, not the pilot.

The reader we have in mind is the OEM architect or technical lead who has finished the migration-readiness work and now has to defend the spend internally. The assumption is that you have read our field notes on the migration itself and understand why the engineering matters. This post answers the next question your CFO will ask.

Six cost categories. Four fleet profiles. A 10-year horizon with the peak migration year shown separately. The ranges are wide on purpose. Where your OEM falls in the range is the question this post is trying to help you answer.

1. Four fleet profiles

The cost of this migration is conditional. The same algorithm change, applied to two different fleets at two different OEMs, can produce migration costs that differ by an order of magnitude. The difference is not the algorithm. It is the operational maturity of the platform the algorithm is being deployed into, and the inventory of hardware that platform has to maintain.

Four fleet profiles capture most of what we see. Real OEMs sit between them and shift profile over time. The point of naming the profiles is to give you a place to start the cost conversation, not to assign you a permanent label.

Profile A: Mature OEM, newer fleet.

Mature PKI infrastructure with crypto-agility designed in. OTA pipeline already running for security telemetry, not just feature updates. SOC and PSIRT integrated with V2X-specific monitoring. Fleet skewed to model years 2022 and newer, with HSMs specified for algorithm agility. Migration is mostly software, with hardware refresh confined to the oldest 5 to 10 percent of fleet.

Profile B: Mature OEM, stranded fleet present.

Same operational maturity as Profile A. But carrying a meaningful long-tail of older vehicles (model years 2018 to 2020) with HSMs that cannot accept ML-DSA. Migration plan must include hardware replacement campaign for the stranded fraction. Twenty to thirty percent of fleet may require physical HSM or ECU replacement over the migration window.

Profile C: Less mature OEM, mixed fleet.

PKI infrastructure exists but was not designed for algorithm agility. OTA pipeline is feature-focused, security cert refresh has been a manual or low-cadence process. SOC is generic enterprise tooling, not V2X-specific. Fleet is mixed across model years, with HSM portfolio inventory not yet built. Migration requires platform investment alongside the algorithm change.

Profile D: Less mature OEM, stranded fleet, regulated environment.

All Profile C conditions. Plus operations in markets with active R155 type-approval evidence obligations (EU, UK, Japan, South Korea, Australia, and China-bound vehicles under GB 44495). Plus stranded fleet fraction requiring hardware replacement. Plus parallel-support obligations for the multi-year transition window. This is the highest-cost profile and the one most reader-OEMs are actually in.

Most OEMs we work with sit between Profile B and Profile D. Profile A is rare. Profile D is the one most readers should plan against, even if they aspire to migrate toward Profile B over the course of the engagement.

2. Parallel-support overhead (the headline finding)

In one engagement we ran in 2024, the OEM had budgeted six months of parallel-support overhead following the algorithm cutover. By month four, the engineering team flagged that the parallel-support window had been planned at the wrong order of magnitude. The realistic window, given the stranded fleet fraction, was six years, not six months.

When an OEM cuts over from ECDSA to ML-DSA, the cutover is not the end of the migration. It is the start of a window during which the fleet runs both algorithms in parallel. New vehicles ship ML-DSA. Stranded vehicles continue running ECDSA. Vehicles in the throughput-degraded bucket run ML-DSA but with monitoring overhead for the performance penalty. The PKI infrastructure issues both certificate types. The OTA pipeline refreshes both certificate formats. The SOC monitors both algorithm paths. The PSIRT process handles incidents on both. Compliance evidence covers both. Engineering teams stand by for both.

This is not a temporary cost. It is a multi-year operational state that scales with how much of the fleet cannot be moved off ECDSA. The cost runs until the last ECDSA vehicle is retired or rotated out of service.

The range is wide because the conditions are wide.

For Profile A, parallel-support overhead runs roughly $5 to $15 per vehicle per year. The PKI and OTA infrastructure already supports algorithm agility; the marginal cost of maintaining two algorithm families is operational, not architectural.

For Profile B, the range widens to $10 to $25 per vehicle per year. The infrastructure can carry both algorithms but the stranded fleet adds compliance overhead and replacement-campaign logistics.

For Profile C, the range is $15 to $30 per vehicle per year. Platform investment is required alongside the migration, and parallel-support obligations land on top of that platform work.

For Profile D, the range is $30 to $60+ per vehicle per year. Where two algorithm families, two certificate formats, dual signing paths, dual monitoring logic, parallel incident response, parallel compliance evidence, and engineering standby must all be sustained, the cost compounds.

The parallel-support cost is the line item most likely to be missed in a first-pass migration budget. It does not appear in any standards document. It does not appear in any algorithm comparison. It appears only when an engineering team that has run a migration sits down and counts the sub-costs honestly. We have built this number for clients. It always comes out larger than the first estimate.

3. HSM and hardware refresh

In one engagement we ran in 2024, the HSM inventory exercise referenced in our field notes on the migration itself ended with twenty-five to thirty percent of the fleet sitting in the stranded bucket. The OEM had budgeted the migration as a software project. The inventory turned it into a hardware-replacement campaign.

For newer or partially upgradeable fleets (Profiles A, B), HSM and hardware refresh runs roughly $8 to $35 per vehicle per year amortised over the migration window. This range covers firmware update tooling, throughput recovery for the degraded bucket, and selective hardware replacement for the genuinely stranded subset.

For genuinely stranded fleets (Profile C and D, with significant pre-2020 model-year exposure), the range moves above $35 and realistically lands in $50 to $75+ per vehicle per year amortised. This number absorbs physical ECU or HSM replacement, dealer campaign logistics, recertification, warranty exposure, and parts availability for hardware that may be at the end of its commercial supply curve.

The upper bound is open-ended because warranty terms and dealer campaign costs vary by market and by OEM. We have seen the number exceed $75 in fleets with complex warranty obligations and dealer-network costs.

4. OTA bandwidth

In one engagement we ran in 2024, the OTA cost line item in the migration budget was set at "negligible, increases by less than 10 percent." The actual increase was 45 times. Not 45 percent. 45 times.

The math is in our earlier post. ECDSA pseudonym certificate pools at 24 KB per vehicle become ML-DSA hybrid pools at 1.1 to 1.2 MB per vehicle. Refresh cadence is the same. Bytes per refresh grow 45 times.

The dollar cost depends on cellular pricing and refresh cadence, both of which are commercial choices not engineering ones. Across Profiles A through D, OTA bandwidth attributable to PQC migration runs roughly $1.50 to $4.50 per vehicle per year. This is the smallest cost line in the entire migration. It is also the one most likely to be invisible until the cellular bill lands.

5. Engineering time

In one engagement we ran in 2024, the OEM had planned 18 months of dedicated migration engineering. The work took 32 months. The variance was not in the algorithm work. The algorithm work took roughly the time planned. The variance was in everything else.

Migration planning, HSM inventory building, OTA pipeline retrofit, certificate format design, SOC tooling updates, PSIRT process changes, compliance evidence preparation, dealer-campaign coordination for stranded fleets. The algorithm change is roughly 15 to 25 percent of the engineering time. Everything else is 75 to 85 percent.

Amortised across the fleet over the migration window, engineering time runs roughly $4 to $12 per vehicle per year for Profiles A and B, and $8 to $20 per vehicle per year for Profiles C and D. The range is wide because the absolute cost depends on team size, salary geography, and migration duration, and the per-vehicle amortisation depends on fleet size.

6. Type-approval evidence updates

In one engagement we ran in 2024, the OEM's R155 evidence package had been built around ECDSA-specific test vectors and security claims. The PQC migration required regenerating the evidence package for each market the OEM type-approved in. The exercise took longer than the algorithm migration itself in two of the three markets.

R155 documentation refresh, audit preparation, regulator interactions per market: $2.00 to $6.00 per vehicle per year for Profiles A and B, $4.00 to $10.00 per vehicle per year for Profiles C and D. The range scales with the number of markets the OEM operates in and the regulatory rigour in each market.

7. Monitoring and SOC tooling

In one engagement we ran in 2024, the verifier-queue-depth alert that surfaced the throughput collapse in our field notes article did not exist before the migration. It had to be built, instrumented, and integrated with the existing SOC workflow. The cost was small. The fact that nobody had budgeted for it at all is the point.

Verifier queue alerts, PSIRT pipeline updates, dashboard changes, dual-algorithm monitoring logic: $0.80 to $2.50 per vehicle per year for Profiles A and B, $2.00 to $5.00 per vehicle per year for Profiles C and D.

8. The TCO summary

Across the six categories, summed, the per-vehicle per-year cost falls into ranges by fleet profile.

Profile A OEMs absorb migration cost in the $14 to $50 per vehicle per year range over a 10-year horizon, with peak migration years closer to the upper bound and tail years closer to the lower.

Profile B widens to $25 to $80, primarily driven by hardware-refresh costs for the stranded fraction.

Profile C runs $35 to $110, with parallel-support overhead and platform investment as the dominant lines.

Profile D, the profile most reader-OEMs are actually in, runs $120 to $150 per vehicle per year in peak migration years and tapers to $40 to $90 in tail years.

For a fleet of 250,000 vehicles, Profile D's peak-year migration cost runs $30 million to $37.5 million. For a fleet of 1 million vehicles, the peak-year cost lands in the $120 million to $150 million range.

9. The cost surprise

The cost categories most often missed in first-pass budgets are not the ones the algorithm conversation prepares you for. Parallel-support overhead is the largest single line item and the one least likely to appear in a vendor pitch. Type-approval evidence updates scale with markets and rarely show up in first-pass estimates. Monitoring and SOC tooling are tiny by dollar value but nearly always absent from first-pass plans.

The first-pass migration budget in most engagements we have run captures roughly 60 to 70 percent of the realistic 10-year TCO. The 30 to 40 percent gap is where the surprises live.

10. Close

The migration is not the migration. The cost is not the algorithm. The work is everything that touches the algorithm, and the cost is everything that touches the work.

If you are working on this migration and any of the above is recognisable, the contact form routes engagement enquiries directly to us. Pick "Explore a business engagement."