Every Active 'Park Outside' Fire Recall, Enumerated: 68 Campaigns, 10.35M Vehicles, and How to VIN-Check One

Key facts

• Census (our computation): 68 active park-outside fire-recall campaigns, 10,346,189 cars and trucks, 15 manufacturers; report dates 2015-06-01 to 2026-07-08.
• Concentration: Kia 4,290,329 vehicles (13 campaigns) and Hyundai 3,542,042 (15) together are 7,832,371 — about 75.7%, roughly 3 in 4, of the total. Stellantis/Chrysler is next at 1,921,053 (11).
• 2026 surge: 11 park-outside campaigns adding 1,668,804 vehicles in the first seven months of 2026, already the second-biggest year on record. NHTSA's literal "Park Outside:" subject-line prefix is essentially a 2026 labelling change (16 campaigns, 1,670,864 units, all in 2026), so the phrasing is newer than the practice.
• The campaign behind the spike — 26V363000: 2021-2025 Jeep Wrangler, Wrangler 4xe, and Gladiator, 1,076,999 vehicles; an electric hydraulic power-steering-pump wiring circuit can overheat and cause a fire even parked and off; remedy inspects/replaces the wiring and pump, free; owner letters mailed around 2026-07-09.
• Largest single campaign — 23V652000 (Kia): 1,730,192 vehicles, 2010-2019 models; HECU engine-compartment fire; free HECU-fuse replacement.
• Fire mechanism pattern: most defects sit in a circuit that stays live with the ignition off (ABS/HECU units, power-seat motors, tow-hitch wiring, a power-steering pump, or an HV battery), which is why the fire can start while parked.
• What this file cannot say: dataset 6axg-epim has no completion/remedied column, so an unrepaired count cannot be computed from it. The 10.35M is vehicles ever placed under a park-outside order, not vehicles still unfixed.

What is a "park outside" fire recall, and how many cars are under one?

A park-outside recall is a safety recall whose defect can start a fire while the vehicle is parked and switched off, so the manufacturer and NHTSA tell owners to park outdoors and away from structures and other vehicles until a free repair is completed. Filtering NHTSA's ODI Recalls file on 2026-08-19 for vehicle campaigns that carry the fire_risk_when_parked flag and an explicit park-outside instruction returns 68 active campaigns covering 10,346,189 cars and trucks from 15 manufacturers.

The flag on its own is broader than the advisory, and the gap matters. All 99 flagged vehicle campaigns cover 12,260,259 vehicles, but 22 of them — 1,911,244 vehicles, including Nissan's ABS-pump campaigns and Subaru's ground-bolt recall — describe a fire that can start while parked without ever telling owners to park outside. Those are real fire recalls; they are not park-outside orders, so they sit outside our headline number. We also exclude 9 campaigns from motorhome, transit-bus and specialty-body makers (Winnebago, Tiffin, New Flyer and similar), which together account for only 2,826 units but would otherwise inflate the manufacturer count from 15 to 23.

That population accumulated over roughly a decade: the earliest park-outside record dates to 2015-06-01. It is not a snapshot of cars burning today — it is the running total of vehicles that have, at some point, been placed under a park-outside order and whose campaigns remain open until each VIN is repaired. Because the file cannot distinguish a repaired car from an unrepaired one, treat 10.35M as "ever ordered to park outside," not "currently unfixed."

Which cars and model years are under a park-outside order right now?

The roster is dominated by Hyundai and Kia engine-compartment fire recalls and, in 2026, by a large Jeep campaign. The table below is the ranked top of the census by potentially affected population; the full manufacturer-level breakdown and the marquee campaigns are in the frozen CSV linked under Sources.

Two pairs here re-recall the same populations — 24V407000 and its 2026 re-recall 26V430000 on the 2020-2024 Kia Telluride (identical 462,869 counts), and 20V519000 / 21V331000 on the same Kia HECU pool (identical 440,370) — so the raw campaign sum modestly double-counts the same vehicles; the 68 campaigns are not 68 disjoint pools. Hyundai's 584,784-vehicle tow-hitch fire recall 23V181000 is deliberately absent: NHTSA flags it as a parked-fire risk but never issued a park-outside instruction for it. If you are shopping a Telluride specifically, our used Kia Telluride problems to check before buying walk-through covers the seat-switch fire recall in detail, and the earlier Kia Telluride fire recall check by VIN piece covers the re-recall history. For the Jeep that drove the 2026 spike, see should I buy a used Jeep Wrangler with an open recall.

By the numbers

• 68 active park-outside fire-recall campaigns (cars and trucks); 10,346,189 vehicles; 15 manufacturers.
• 75.7% of that population is Hyundai + Kia (7,832,371 vehicles) — the claim we can defend is "Hyundai and Kia account for roughly 3 in 4 vehicles ever placed under a park-outside fire order."
• 1,668,804 vehicles added across 11 park-outside campaigns in Jan–Jul 2026, already the second-biggest year on record with five months still to run.
• 4,156,721 vehicles in 2023 alone — the single biggest year in the file — versus 223 in 2015, the first year one appears.
• 1,730,192 — the largest single active campaign, Kia 23V652000.
• 1 campaign / 367 vehicles is the only park-outside campaign that also carries a Do Not Drive order — the overlap is negligible.

Based on 30,236 NHTSA ODI Recalls records (dataset 6axg-epim), retrieved 2026-08-19, filtered to recall_type=Vehicle with fire_risk_when_parked='Yes' and an explicit park-outside instruction, excluding motorhome/bus/specialty-body makers. Frozen roster CSV linked in Sources.

How robust is the 68-campaign count?

This number moves a lot depending on how "park outside" is matched, and we would rather show that than hide it. Three regimes on the same file:

The consequential step is the second one. Taking NHTSA's flag at face value overstates the park-outside population by 22 campaigns and 1,911,244 vehicles, because the flag marks any fire that can start while parked, whether or not owners were told to park outdoors. The largest campaigns it wrongly sweeps in are Hyundai's 584,784-vehicle tow-hitch recall 23V181000, Nissan's two ABS-pump recalls (606,603 combined) and Subaru's 271,694-vehicle ground-bolt recall. Adding do not park to the text match changes nothing — no campaign uses that phrasing without also saying "park outside." The third step is cosmetic for units (2,826 vehicles) but matters for the make count, cutting 23 manufacturers to 15 once RV, transit-bus and refuse-body makers come out.

So the honest headline is 68 campaigns and 10.35M vehicles, not the 99 and 12.26M the raw flag would license. The concentration finding survives the correction and in fact strengthens: Hyundai and Kia rise from 68.8% of the flagged population to 75.7% of the park-outside population.

Where the widely cited "3.2 million unrepaired" figure comes from

Separately from our census, the figure in the August 17-18, 2026 coverage — 3.2 million vehicles under an unrepaired park-outside recall — is Carfax's proprietary estimate, published in a Carfax release dated 2026-06-18 and republished by the Washington Post ("They were told to park outside. Then their cars burst into flames," Todd C. Frankel, 2026-08-17) and echoed by Yahoo Autos. Carfax attributed a near-50% jump to the roughly 1-million-vehicle Jeep campaign 26V363000.

We keep that number out of our headline for one reason: it is an unrepaired estimate, and it is not reproducible from the ODI flat file, which carries no completion or remedied-count column. Our reproducible number is the population under the order (10.35M ever placed; 1.67M added in 2026), not an unrepaired count. Confirming how many remain unfixed would require NHTSA's separate quarterly recall-completion reports, not this file. The Washington Post's human example — a 2022 Jeep Wrangler 4xe plug-in hybrid that burned — maps to Chrysler's high-voltage-battery campaign 25V741000.

What a VIN check can and can't tell you here

A VIN check answers the exposure question, not the repair question. Entering a car's 17-character VIN — at NHTSA's free tool or in a history report — tells you whether that specific vehicle falls within one or more open recall campaigns, including these park-outside fire recalls. It does not tell you whether the free fix was actually performed on that car. Repair status lives with the manufacturer and is surfaced by NHTSA's VIN lookup or a franchised dealer, not by a presence-only screen.

The practical order is: screen the VIN for the recall's presence, then confirm the fix with NHTSA or the dealer. For the broader principle, see is it safe to buy a used car with an open recall, which shows that the biggest recalls are statistically the least likely to already be fixed — directly relevant, because the campaigns on this roster are among the largest ever filed.

Buying one anyway: what to weigh

Buying a car on this roster is not inherently a bad decision — the repair is free, transfers with the car, and does not depend on you being the original owner. The two real costs are the interim inconvenience of parking outdoors until the fix, and any wait for parts on newer campaigns like the Jeep 26V363000 where the remedy involves replacing wiring and a pump rather than swapping a fuse. The older Hyundai/Kia engine-fire campaigns that dominate this list (2010-2019 model years) are cheap, high-volume used cars, and their remedies are typically a fuse or relay.

On exactly those cheap, high-mileage cars, the recall is rarely the biggest risk — history is. A 2013 Sorento or 2014 Sonata has had a decade of owners, and its salvage-auction, accident, odometer, and prior-listing record tells you more about whether to buy it than the recall does. That history layer is where a paid VIN report earns its place alongside the free recall screen.

Related checks

• Should I buy a used Jeep Wrangler with an open recall — the 26V363000 park-outside campaign behind the 2026 spike, in depth.
• How to check a used BMW for open recalls before buying — another park-outside fire campaign (26V441) on the roster.
• Used electric car battery recall check before buying — HV-battery fire recalls, the same class as the Jeep 4xe 25V741000.

Method

Source file. NHTSA Office of Defects Investigation Recalls Data, Socrata dataset 6axg-epim on datahub.transportation.gov, retrieved via the JSON API on 2026-08-19. This is the public flat file behind NHTSA's recall lookup. The universe is N = 30,236 recall records covering every recall type (Vehicle, Equipment, Tire, Child Seat).

Filter, in three stated steps. Each step is published above with its own campaign and vehicle counts so any of them can be reproduced or disputed independently:

1. recall_type = 'Vehicle' and fire_risk_when_parked = 'Yes' — NHTSA's own structured flag for a fire that can start while the vehicle is parked. Yields 99 campaigns / 12,260,259 units.
2. Require an explicit park-outside instruction: case-insensitive park outside in either the subject or the corrective_action field. Yields 77 campaigns / 10,349,015 units. We also tested do not park as an alternate phrasing; it adds nothing, because every campaign using it also says "park outside."
3. Exclude manufacturers whose products are not cars or light/medium trucks — Winnebago, Tiffin Motorhomes, Grand Design RV, REV Recreation Group (motorhomes), New Flyer (transit buses), Crestline Coach (ambulances), Textron Specialized Vehicles and The Heil Co. (refuse bodies). Yields the published 68 campaigns / 10,346,189 units across 15 manufacturers, report dates 2015-06-01 to 2026-07-08.

Aggregation. Campaigns are counted with count(*); vehicle totals are the sum of NHTSA's potentially_affected column. Manufacturer shares are that sum grouped by the manufacturer string as NHTSA records it, so Kia America and Hyundai Motor America are separate rows even though the two are corporate affiliates; the article states their combined share explicitly rather than merging the rows.

No dedup was applied, and that is a limit, not an oversight. · potentially_affected is a per-campaign population, and some campaigns re-recall vehicles already covered by an earlier one. Two pairs in the top ten do exactly this: Kia 24V407000 / 26V430000 (both 462,869 Tellurides) and Kia 20V519000 / 21V331000 (both 440,370 HECU vehicles). The ODI flat file carries no VIN-level roster, so the overlap cannot be removed from this source. 10,346,189 is therefore a count of vehicle-campaign placements, not of distinct vehicles, and the distinct-vehicle figure is lower by at least the ~903,000 units in those two pairs.

Other known limits.

• The file has no completion or remedied column, so nothing here is an estimate of cars still unrepaired. Every figure is "ever placed under the order." Carfax's widely quoted 3.2 million unrepaired figure comes from its own VIN database and is not reproducible from this file.
• A campaign has no closed date in this file, so "active" means the campaign exists and remains open until each individual VIN is remedied, not that every listed vehicle is currently unfixed.
• Counts are as-filed by manufacturers and are revised from time to time; a re-pull on a later date will not match exactly.
• The year-by-year series is by report_received_date, which is when NHTSA received the report, not when owners were notified.

Reproduce it. The frozen CSV linked below carries all 68 campaigns individually (id, date, manufacturer, subject, component, units, corrective action), the 15 manufacturer subtotals, and the three-step robustness ladder as its own rows.

Sources

• NHTSA — Check for Recalls by VIN (free VIN lookup; authoritative repair/remedy status)
• NHTSA ODI Recalls Data, dataset 6axg-epim (Socrata JSON API) — the file we processed; carries the fire_risk_when_parked flag
• NHTSA ODI Recalls Data landing page (6axg-epim)
• NHTSA Recalls API — campaign 26V363000 (confirms parkOutSide:true)
• NHTSA campaign lookup — Jeep 26V363000
• NHTSA consumer alert — Kia & Hyundai "park outside" fire recalls
• Zilocar frozen roster CSV — park-outside fire-recall census, retrieved 2026-08-19
• Carfax release — "Park Outside recalls jump nearly 50%" (origin of the 3.2M figure, 2026-06-18)
• Washington Post — "They were told to park outside. Then their cars burst into flames." (Todd C. Frankel, 2026-08-17)
• Yahoo Autos — "3.2 Million Vehicles Carry an Outstanding Park Outside Recall"

Last verified: 2026-08-19.