Skip to content
Fonteum·Government Procurement EvidenceThe Leakage ReportQuestions

NHTSA ODI recalls - snapshot 2026-06-24

Software is now the most-recalled vehicle component

In the NHTSA recalls table, cars now look like computers. The largest recalled component bucket is not brakes, airbags, engines, or fuel pumps. It is ELECTRICAL SYSTEM:SOFTWARE: 20,508 affected make-model-year-campaign records.

Across 242,041 current NHTSA recall records in the 2026-06-24 snapshot, ELECTRICAL SYSTEM:SOFTWARE is the leading component with 20,508 records (8.5% of the file). It is ahead of TIRES:MARKINGS at 19,460. A recall record here means one affected make, model, model year, and campaign row. It is not a unique campaign count, and the study is aggregate-only.

Key findings

20,508

ELECTRICAL SYSTEM:SOFTWARE is the largest NHTSA recall component by recall-record count in the 2026-06-24 snapshot. It appears on 20,508 make-model-year-campaign rows.

8.5%

Software accounts for 8.5% of all 242,041 current recall records in this snapshot, ahead of every mechanical component-name bucket.

19,460

The next component, TIRES:MARKINGS, has 19,460 recall records. The software lead is 1,048 records.

242,041

The table contains 242,041 recall records. One record means an affected make, model, model year, and recall campaign row - not a unique campaign.

44,209

Mercedes-Benz USA is the largest manufacturer bucket by recall-record count in this snapshot. This is a count of affected-row records, not a finding about unique campaigns or fault.

At a glance

242,041
NHTSA recall records in scope
20,508
Software component recall records
8.5%
Software share of all recall records
44,209
Largest manufacturer bucket

Top recalled components

The ranking is by component_name on current rows of public.nhtsa_recalls. Software has overtaken every mechanical system in this recall-record view: no component-name bucket in the table has more rows.

RankNHTSA componentRecall recordsShare of records
1ELECTRICAL SYSTEM:SOFTWARE20,5088.5%
2TIRES:MARKINGS19,4608%
3COMMUNICATION:AUTO CRASH NOTIFICATION16,6886.9%
4EQUIPMENT9,9934.1%
5FUEL SYSTEM, GASOLINE:DELIVERY:FUEL PUMP8,4043.5%

Top manufacturers by recall-record footprint

Manufacturer counts use the same row unit. A manufacturer bucket can be large because many make-model-year rows sit under one or more campaigns. These are aggregate row counts, not a unique recall-campaign ranking and not a claim about fault.

RankManufacturer bucketRecall recordsShare of records
1Mercedes-Benz USA44,20918.3%
2Prinx Chengshan Tire19,3218%
3Toyota Motor Eng. & Mfg14,1235.8%
4Ford12,5875.2%
5Honda12,0915%

What the row count means

Recall records are affected make x model x model-year x campaign rows under an NHTSA recall campaign; they are not unique campaigns.This is why the page says "recall records" throughout. It does not count unique campaign numbers, affected vehicles, injuries, complaints, or defect narratives.

Reproduce this

The SQL below is the reproducing query served by the SQL download route. It selects only aggregate grouping columns, includes the expected top-five counts as comments, and reads no individual-level material.

  • Download SQL - reproducing query as a committed TypeScript constant.
  • Download JSON / Download CSV - the committed aggregate snapshot.
-- Software Is Now the Most-Recalled Vehicle Component.
-- Source: NHTSA ODI Recalls flat files loaded to public.nhtsa_recalls.
-- Snapshot checked: 2026-06-24.
-- Unit of analysis: recall records are make x model x model-year x campaign rows,
-- not unique recall campaigns and not affected-vehicle counts.
-- Aggregate-only: no VINs, complaints, individuals, or defect narratives are selected.

WITH current_recalls AS (
  SELECT
    record_id,
    campaign_number,
    make,
    model,
    model_year,
    component_name,
    manufacturer
  FROM public.nhtsa_recalls
  WHERE valid_to IS NULL
)
SELECT count(*)::int AS total_recall_records
FROM current_recalls;
-- Expected: 242041

WITH current_recalls AS (
  SELECT component_name
  FROM public.nhtsa_recalls
  WHERE valid_to IS NULL
)
SELECT
  component_name AS recalled_component,
  count(*)::int AS recall_records
FROM current_recalls
WHERE component_name IS NOT NULL
  AND btrim(component_name) <> ''
GROUP BY component_name
ORDER BY recall_records DESC, recalled_component
LIMIT 5;
-- Expected:
-- ELECTRICAL SYSTEM:SOFTWARE                         20508
-- TIRES:MARKINGS                                    19460
-- COMMUNICATION:AUTO CRASH NOTIFICATION             16688
-- EQUIPMENT                                          9993
-- FUEL SYSTEM, GASOLINE:DELIVERY:FUEL PUMP           8404

WITH current_recalls AS (
  SELECT manufacturer
  FROM public.nhtsa_recalls
  WHERE valid_to IS NULL
)
SELECT
  manufacturer,
  count(*)::int AS recall_records
FROM current_recalls
WHERE manufacturer IS NOT NULL
  AND btrim(manufacturer) <> ''
GROUP BY manufacturer
ORDER BY recall_records DESC, manufacturer
LIMIT 5;
-- Expected:
-- Mercedes-Benz USA          44209
-- Prinx Chengshan Tire       19321
-- Toyota Motor Eng. & Mfg    14123
-- Ford                       12587
-- Honda                      12091

Limitations

  • This is a count of recall records, not unique campaigns. A single NHTSA campaign can appear on multiple affected make-model-year rows.
  • The snapshot uses current rows in public.nhtsa_recalls as of 2026-06-24. NHTSA can add, amend, or supersede recall data after that date.
  • Component names are the published NHTSA component buckets. They are not normalized into a custom Fonteum taxonomy for this study.
  • Manufacturer counts are affected-row footprints. They are not unique-campaign counts, vehicle counts, severity measures, or fault findings.

Sources

U.S. federal public records (U.S. Government Works, public domain). The study uses the NHTSA Office of Defects Investigation Recalls flat file, loaded into public.nhtsa_recalls and frozen as a committed aggregate snapshot dated 2026-06-24.

provenance: the source file, snapshot date, SQL, JSON, and CSV stay chained to the same committed aggregate artifact so the counts remain re-checkable against NHTSA's own files.

Source: NHTSA Office of Defects Investigation Recalls flat files, snapshot 2026-06-24. U.S. Government Works. Recall records are make/model/model-year/campaign rows, not unique campaigns; the files are re-checkable at nhtsa.gov.

  • NHTSA datasets and APIs →official source
  • NHTSA homepage →official source
  • ODI recalls flat file →official source

How to cite this

Fonteum (2026). Software Is Now the Most-Recalled Vehicle Component. Derived from NHTSA ODI Recalls flat files (snapshot 2026-06-24). https://fonteum.com/gov/research/software-most-recalled-vehicle-component-2026

Canonical URL: https://fonteum.com/gov/research/software-most-recalled-vehicle-component-2026 · License: U.S. Government Works (public domain; 17 U.S.C. §105)

Reviewed by the Fonteum Vehicle Safety Data Desk. Public-records analysts. This study reports aggregate component and manufacturer counts from official NHTSA recall-record files. It names no individual, VIN, complaint, or vehicle owner and makes no determination about any specific recall.
Published 2026-06-24 - updated 2026-06-24 - methodology nhtsa-software-recalled-component/v1 - Fonteum.

Frequently asked questions

What is a recall record in this study?

Recall records are affected make x model x model-year x campaign rows under an NHTSA recall campaign; they are not unique campaigns. A campaign that applies to many makes, models, or model years can therefore contribute many recall records. This page does not count unique campaign numbers.

Is software really ahead of mechanical components?

Yes, by recall-record count in this committed NHTSA snapshot. ELECTRICAL SYSTEM:SOFTWARE has 20,508 records, ahead of TIRES:MARKINGS at 19,460 records and all other component-name buckets in the top-five ranking.

Does this page name individual drivers or vehicles?

No. The study uses the NHTSA recalls table only and publishes aggregate counts by component and manufacturer. It includes no VINs, complaints, driver names, addresses, narratives, or individual-level records.

How can I reproduce the numbers?

Use the SQL linked on this page against public.nhtsa_recalls filtered to current rows (valid_to IS NULL). The SQL groups by component_name and manufacturer, and the expected top-five counts are included as comments.

Does a high manufacturer count mean that manufacturer had more unique recalls?

Not necessarily. Manufacturer rows are also recall-record counts, so a campaign covering many make-model-year rows can contribute many records. The count is a measure of affected-row footprint in the NHTSA table, not a unique-campaign count or a fault ranking.

Part of the Fonteum government public-records evidence silo.

Fonteum is a public-records evidence platform. This Government Procurement Evidence silo reports exact regulatory facts from federal public records (SAM.gov, USASpending.gov, FAPIIS). It assigns no risk score and makes no determination of wrongdoing; confirm current status at the official source.

Silo home · The Leakage Report · Questions · SAM.gov · USASpending.gov