Only 10 US Car and Light-Truck Steering Recalls Were Filed in 2026, the Fewest Since 2020 — and Three of Them Cover 98% of the Vehicles

Key facts

• Load-bearing count: 10 car and light-truck steering recall campaigns filed Jan 1 to Oct 8, 2026 (strict definition: component label begins "STEERING"). Eleven under the broad definition, which adds one Lincoln driver-assist campaign.
• This is the lowest, not the highest. Ten is below every year back to 2020 (2020: 12, 2024 peak: 25). The finding survives on every basis we tested, including the unrestricted one.
• The unrestricted number, printed beside it so nothing is hidden: counting every vehicle type in NHTSA's file, including motorcycles, heavy trucks, buses and fire apparatus, 2026 returns 20 campaigns and 1,662,335 vehicles (21 campaigns and 1,663,348 vehicles on the broad label test). Restricting to cars and light trucks nearly halves the count and barely touches the units.
• Vehicle census, an attributed total and never a ranking: 1,584,424 vehicles across the 10 car campaigns. These are manufacturers' launch-time estimates of units potentially affected.
• Three campaigns are 98% of those vehicles: Jeep 26V363000 (1,076,999), Ram 26V621000 (265,512), VW/Audi 26V590000 (208,724 US). The other seven campaigns add 33,189 between them.
• Jeep 26V363000 (filed 2026-06-04): 2021-2025 Wrangler, Wrangler 4xe, Gladiator. Filed under a hydraulic power-steering pump label; the consequence NHTSA records is a vehicle fire, and its remedy text tells owners to park outside. Followed investigation PE24024. Owner letters expected 2026-07-09.
• Ram 26V621000 (filed 2026-09-29): 2022-2026 Ram ProMaster vans plus the 2024-2026 Ram 3500 EV. Water can enter the electric power-steering wiring harness and disable assist. Interim owner letters expected 2026-11-17; VINs searchable at NHTSA from 2026-10-06.
• VW/Audi 26V590000 (filed 2026-09-11): 2018 VW Tiguan, 2018-2019 VW Atlas, 2019-2021 Audi Q3. A corroded right-side steering rack mounting bolt can break, risking loss of steering control. Owner letters expected 2026-11-10.

How many cars were recalled for steering problems in 2026, and is it really a "wave"?

Ten, on the measure that actually tracks how often defects get filed. We counted 10 car and light-truck steering-defect recall campaigns through October 8, 2026, or 11 under the broader label test. That is the lowest January-to-October total of the last seven years. The headlines read like a wave because a few of 2026's campaigns are enormous, not because more of them are being filed.

Here is the year-over-year comparison, run on identical terms each year, on both bases:

Computed from 245,765 records in NHTSA's FLAT_RCL_POST_2010 recalls flat file, retrieved 2026-10-08 (newest report-received date in the file: 2026-10-05). Counts are distinct recall campaigns (RCLTYPECD='V'), not rows. "Strict" means the component label begins "STEERING"; "broad" means it merely contains the word. The per-campaign car classification ships in the frozen CSV below.

The restriction matters, and here is exactly how much. On the unrestricted basis, 10 of 2026's 21 steering-labelled campaigns are not cars or light trucks at all: two Honda motorcycle recalls and one Harley-Davidson, Kenworth and Peterbilt, Freightliner and Mack and Volvo heavy trucks, a Pierce fire-apparatus campaign, and two commercial step-van builders. They hold 77,911 vehicles, 4.7% of the total, and 72,720 of those 77,911 are motorcycles. So the campaign count nearly halves under the restriction while the vehicle total barely moves. Two summary statistics from one filter degraded in opposite directions, which is why both are printed here rather than one.

The single difference between strict and broad in 2026 is campaign 26V377000, a 2019 Lincoln Nautilus filed under "LANE DEPARTURE: LANE KEEP: STEERING ASSIST" (1,013 vehicles). NHTSA's recorded consequence is a failure to deliver hands-off-the-wheel warnings, not a mechanical steering failure. It contains the word "steering" and is not a steering defect, which is why we report both counts instead of blurring them.

By the numbers: where the 1,584,424 vehicles come from

Three oversized campaigns. Those three alone account for 1,551,235 vehicles, or 97.9% of the car and light-truck census. The remaining seven campaigns add 33,189 between them, and several are tiny: BMW's 26V524000 covers 13 vehicles and Land Rover's 26V527000 covers 14. Treat the total as a census, not a ranking: these are manufacturers' launch-time estimates of units potentially affected, and a car can be recalled without ever developing the defect.

⭐ The detail that matters most to a used buyer, and it is not the headline count: only three of the ten campaigns reach model year 2021 or older. Six cover model-year 2026 or 2025-2026 vehicles that are still effectively new, and the BMW X3 campaign starts at 2024. The three that reach genuinely used stock are Jeep 26V363000 (2021-2025), VW/Audi 26V590000 (2018-2021) and Land Rover 26V527000 (2016-2017), and between them they hold 1,285,737 vehicles, 81.1% of the 2026 car and light-truck total. So a used-car shopper's exposure to this year's steering recalls is concentrated almost entirely in two nameplate families: the Wrangler and Gladiator, and the Tiguan, Atlas and Q3.

📌 Also worth recording: the two largest campaigns were both pushed by NHTSA, not volunteered. The flat file marks 26V363000 (Jeep) and 26V621000 (Ram) as ODI-influenced, while all eight smaller car campaigns are marked MFR, manufacturer-initiated. The two campaigns that account for 85% of the vehicles are the two the agency had to press for.

Two cautions the component labels hide. First, the largest campaign, Jeep 26V363000, is filed under a steering component but the consequence NHTSA records is a fire: the power-steering pump wiring can overheat and ignite, which is why its remedy text tells owners to park outside and away from structures. If you are cross-shopping, see how to check if a used car has a park outside fire recall. Second, the Ram campaign is about power-steering assist disappearing, the same "the car suddenly drives differently" failure mode we catalogue in how to check if a used car has a power loss recall, and it covers ProMaster vans plus a Ram 3500 EV, not the heavy-duty pickups some early coverage implied.

On the VW/Audi campaign, use only the US figure. The 208,724 count is the NHTSA campaign. Larger figures circulating in international coverage are worldwide totals and must not be read as a US count.

Which carmaker filed the most steering-recall campaigns in 2026?

Nobody, and the flatness is itself the finding. On a car and light-truck basis, two groups filed two campaigns each and six filers filed one each.

⚠️ One ranking error this basis removes, and it is worth stating because the unrestricted count invites it. Counting every vehicle type, Honda appears tied at the top with two independent steering campaigns. Both are motorcycle recalls (26V146000 and 26V500000, the Rebel CMX300 and CMX500 and the SCL500), so on a used-car comparison Honda filed none. Chevrolet and GMC also look like two filers each on a make-level count, when 26V345000 and 26V399000 are the same two General Motors campaigns listing both brands. Group company-level figures on the Part 573 filer, not the make name.

We rank by campaign count only. We do not rank filers by vehicle count, because vehicle totals are launch-time estimates and one oversized campaign would dominate a leaderboard meant to measure how often defects get filed.

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

A VIN check is a screening and history tool, not a repair tracker. It will flag that a steering recall is attached to a given vehicle, the same recall universe as NHTSA's free lookup, and beyond that show the accident, salvage and ownership history NHTSA's tool does not. What it cannot do is confirm the recall was fixed on that specific car, and it will not surface open NHTSA investigations such as PE26002, the ProMaster steering probe that preceded the Ram recall.

If you are weighing a used Jeep Wrangler, a Ram ProMaster or a 2018-2021 VW or Audi, the practical move is two steps: confirm open-versus-fixed recall status for that exact VIN, free, at NHTSA's recall lookup, and separately check whether the same car was ever crashed, flooded or title-troubled, because a recalled car with a salvage-auction record is a very different purchase. For the broader "is it safe to buy at all" question we walk through how to check if a used car has a do not drive recall, and because FCA models carry overlapping campaigns it is worth also running how to check if a used car has an open airbag recall on any Jeep or Ram you are considering.

A Zilocar VIN check rolls recall screening together with accident and airbag-deployment records, salvage- and junk-auction history, odometer and theft checks, ownership and sales-listing history, specs, safety ratings and valuation, which is the history a recall notice alone never shows. For whether a specific steering recall was completed, and for any open investigation, rely on NHTSA and the dealer. Those we cede.

Method

Source file. NHTSA's Office of Defects Investigation recalls flat file, FLAT_RCL_POST_2010.txt, inside FLAT_RCL_POST_2010.zip. Downloaded 2026-10-08 at 07:30 UTC; the server reported Last-Modified: Wed, 07 Oct 2026 12:12:16 GMT; the archive's CRC was verified intact before parsing. The file carries 245,765 rows resolving to 15,297 distinct recall campaigns, and its newest report-received date is 2026-10-05.

Universe and filter. The flat file stores one row per campaign-and-model combination, while POTAFF (units potentially affected) and RCDATE (report received) are campaign-level fields repeated on every row. We therefore collapse to one record per CAMPNO first; counting rows would multiply-count every campaign by its model count. We keep RCLTYPECD='V' (vehicle recalls), filter RCDATE to January 1 through October 8 of each year from 2020 to 2026, and select campaigns whose component label COMPNAME · begins with · STEERING (strict, N = 10 for 2026) or merely contains it (broad, N = 11). Counts are distinct CAMPNO values.

The car and light-truck restriction. · RCLTYPECD='V' means "vehicle", not "car": it carries motorcycles, buses, heavy trucks, fire apparatus and commercial chassis. Because this is a used-car page, the headline basis excludes them. Classification works in two passes: Part 573 filers that build no passenger vehicles are dropped whole (Harley-Davidson, Mack, Volvo Trucks, Paccar, Daimler Trucks, Pierce, Blue Bird, Gillig and the trailer, RV and fire-apparatus builders), and inside filers that build both, individual motorcycle and medium/heavy chassis model strings are dropped (Honda CMX300, CMX500, SCL500; BMW M 1000 R, M 1000 XR; Ford F-650 SD, F-750 SD, F-53, F-59; GM ZEVO; Ram 4500, 5500). The resulting per-campaign call ships as a vehicle_class column in the frozen CSV, so a reader can redo or reject it. Two borderline calls are flagged there and included: the Chevrolet Express and GMC Savana vans, and the Ram ProMaster vans, which are class 2-3 light commercial vehicles.

Aggregation. Units are the sum of POTAFF over distinct campaigns. The make leaderboard is grouped on the Part 573 filer (MFGTXT), not the make name (MAKETXT), because one campaign routinely spans several makes: 26V345000 and 26V399000 each list both Chevrolet and GMC, and a make-level count reports one General Motors campaign as two filers.

Robustness checks, reported either way.

1. The restriction nearly halves the count and barely moves the units: 20 campaigns to 10 (-50%), 1,662,335 vehicles to 1,584,424 (-4.7%). Two statistics from one filter degraded in opposite directions, so both bases are published.
2. The finding survives every basis: 2026 is the minimum of 2020-2026 on car-strict (10), car-broad (11), all-vehicle-strict (20), all-vehicle-broad (21), and on a deliberately generous union of label-matched plus defect-text-matched campaigns (28 for 2026, against 38 in 2025 and 81 in 2024). That last test exists because a label-based count would be confounded if NHTSA's labelling practice had drifted; it has not drifted enough to reverse the result.
3. Label multiplicity is not the explanation. Only 4 of the 21 steering-labelled 2026 campaigns carry more than one component label, and in 3 of those both labels are steering. Only 26V242000 pairs steering with suspension.
4. The make leaderboard reverses under the restriction, and the unrestricted version is the wrong one for this page: Honda's two campaigns are both motorcycle recalls.
5. The window is three days short. The file's newest report-received date is 2026-10-05, not 2026-10-08, while comparison years are complete. At the measured 2026 rate of one car steering campaign every 28 days, three days cannot close the two-campaign gap to 2020 and 2025; it would take two late filings to produce a tie, and none to produce a reversal.

Corroboration. The four load-bearing campaigns (26V363000, 26V621000, 26V590000, 26V377000) were re-queried against api.nhtsa.gov/recalls/campaignNumber and matched the flat file exactly on component label, unit count and report-received date, 4 of 4. Owner-letter dates, remedy text and the October 6 VIN-searchability date are quoted from NHTSA's own Remedy field rather than from press coverage. Investigation PE26002's opening resume was read directly and names 2022-2023 Ram ProMasters, loss of power steering assist from water intrusion into the power steering module connector, 136,666 units estimated, opened 2026-03-02.

Known limits.

• POTAFF is the manufacturer's launch-time estimate of units potentially affected. It is not a count of cars still on the road, still unrepaired, or still in the US. This file can support "how much exposure was opened" and cannot support "how many dangerous cars are out there" or any model-level "worst car" ranking.
• The flat file has no repair-completion field, so nothing here measures how many of these vehicles were actually fixed.
• The component label is NHTSA's categorisation, not a diagnosis. A campaign labelled steering can have a fire as its recorded consequence, as 26V363000 does.
• The car and light-truck classification is ours and is a judgement at the class 2-3 van boundary. The vehicle_class column exists so it can be rejected.

Related checks

• how to check if a used car had an airbag deployed
• how to check if a used car has a fake or aftermarket airbag
• how to check if a used car has counterfeit parts before buying

Sources

• NHTSA ODI recalls flat file (FLAT_RCL_POST_2010.zip) — the dataset we computed over. 245,765 rows, 15,297 distinct campaigns, retrieved 2026-10-08 07:30 UTC, server Last-Modified 2026-10-07 12:12:16 GMT, newest report-received date 2026-10-05.
• NHTSA recalls API, per-campaign corroboration: Ram 26V621000, Jeep 26V363000, VW/Audi 26V590000, Lincoln 26V377000. All four matched the flat file exactly on component label, units and report-received date.
• NHTSA investigation PE26002 opening resume (PDF) — the ProMaster power-steering water-intrusion probe that preceded the Ram recall; 2022-2023 ProMaster, 136,666 units estimated, opened 2026-03-02.
• NHTSA recalls landing page and free VIN lookup — definitional backing for 49 CFR Part 573 and the free consumer tool for open-versus-fixed status.
• Frozen datasets for this story, so any figure above can be rechecked independently:
• all 2026 campaigns with the per-campaign car classification: steering-recalls-2026-seven-year-low-nhtsa-data-vin-check-2026-10-08.csv
• year-by-year counts on all four bases: steering-recalls-2026-seven-year-low-nhtsa-data-vin-check-by-year-2026-10-08.csv

Last verified: 2026-10-08