What brands have the most recalls?

The 25 firms named on the most distinct recalls, with the affected-unit total where the record carries one. A recall that names more than one firm is counted for each firm it lists. The list mixes retailers — CPSC frequently records the recalling retailer for store-brand goods — with powersports and outdoor-equipment makers. Only the 5,419 recalls that name a firm (54.9% of the total) feed this leaderboard.

The largest firms’ share of all 8,116 brand-recall appearances. The single most-recalled firm accounts for 1%; the top 25 for 8.4%. Unlike the vehicle-recall record — where a handful of automakers dominate — the consumer-product record is a very long tail: 5,646 distinct firms each carry at least one recall.

The most-cited hazard categories, using CPSC’s own coded labels. Counts are of distinct recalls citing each category, over the 1,588recalls that carry a coded hazard (free-text hazard descriptions are not counted here). Choking and fire lead by a wide margin — a reflection of how many recalls involve children’s products and electrical goods.

The 20 most-recalled product categories, over the 6,844 recalls that carry a category label (a legacy field, present 1973–2019). Children’s products — clothing, toys, cribs, costume jewelry, pacifiers — dominate the top of the list. CPSC’s category vocabulary changed over the years, so two near-identical bicycle labels appear; both are shown as published, not merged.

Recalls by the year CPSC dated them, from 2000 onward (the full series back to 1973 is in the download). Annual volume peaked at 452 in 2007 — the year of the China-made toy lead-paint wave — and the affected-unit total is unreported before 1999. The 2026 row is partial; the snapshot was taken on 2026-06-24.

• Recall records, not safety. A recall is the firm and CPSC acting to fix a hazard. A high count reflects how often a firm is named on the public record — frequently as the recalling retailer — not a measure of product quality. This study assigns no score and ranks no brand on safety.
• Distinct recalls. The unit is the CPSC recall number. A recall naming more than one firm is counted for each firm it lists, so brand counts sum to more than the recall total.
• Firms as published. Firm names are taken verbatim from the CPSC record, with only the trailing location annotation stripped so the same firm is not split. The study names no individual.
• Uneven coverage. Only 54.9% of recalls name a firm; the product-category field is absent after 2019; and coded hazard categories cover 1,588 recalls. Each cut states its own coverage.

The study aggregates the signed public.cpsc_recalls table in the fonteum-platform warehouse, ingested from CPSC’s public saferproducts.gov Recalls feed (snapshot 2026-06-24). CPSC publishes one record per recall, so the counting unit is the distinct recall_number. The firm, hazard, and product-category fields are arrays and are unnested; a recall’s affected-unit total is one value and is read once per recall.

Brand, hazard, category, and year cuts are plain GROUP BYcounts over the distinct recall set. CPSC annotates each firm name with a location (“, of Chilton, Wis.”); that suffix is stripped — the only transformation applied — so the same firm is not split into fragments. A recall spanning more than one firm is counted for each firm it lists, so the brand counts sum to 8,116brand-recall appearances; concentration shares use that appearances total as the denominator. Hazard cuts read only CPSC’s short coded labels, excluding free-text hazard sentences. Every published figure is re-derivable from the SQL below, whose expected-result comments match the committed JSON snapshot exactly, and every snapshot is content-hashed and witness co-signed.

Re-derive every figure on this page from the published artifacts:

• Reproducible SQL — the exact GROUP BY counts, with expected-result comments.
• Download JSON · Download CSV — the committed recall-by-brand snapshot.

Every figure here traces to a signed source snapshot, not our word for it. The CPSC pull is content-hashed and chained; you can re-hash the published bytes against the attestation yourself.

Re-check a snapshot → — re-hash any Fonteum snapshot and confirm the bytes match the chained attestation.

Fonteum (2026). The Most-Recalled Product Brands: The U.S. Consumer-Product Recall Record Counted by Firm. Derived from the CPSC Recalls feed (snapshot 2026-06-24; recalls dated 1973–2026). https://fonteum.com/gov/research/most-recalled-product-brands-2026

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

• This counts recalls on the snapshot date, not a settled historical series; a recall’s affected-unit count can be amended after it is first published.
• Firm coverage is uneven and has fallen over time, so the all-time leaderboard skews toward long-established retailers; it is not a like-for-like comparison across firms of different size or era, and is not a safety measure.
• Firm attribution follows the CPSC record as published, including multi-firm recalls and recalls where the named firm is the retailer rather than the manufacturer; it is not deduplicated to a parent corporate group.
• Affected-unit totals are reported as published and use mixed units of measure (mostly product counts, occasionally pounds or volume), so unit sums are indicative, not exact product counts.
• A recall is an administrative safety action on the record as of the date queried, not a judgment about any brand or product. Confirm any specific recall’s status at CPSC.

One published source: the CPSC Recalls feed, a U.S. Government public record issued by the Consumer Product Safety Commission. Every figure is sourced to the snapshot date shown below.

Source: U.S. Consumer Product Safety Commission — Recalls (saferproducts.gov RestServices feed), snapshot 2026-06-24. Public domain (U.S. Government Works). Confirm any specific recall at cpsc.gov/Recalls. Confirm current status at SAM.gov →