Honda Built Its V6 for 27 Years. Then 1.4 Million of Them Started Eating Their Own Bearings.

Honda's J-series V6 is the engine equivalent of a golden retriever: friendly, predictable, and the kind of thing you recommend to your sister without checking Consumer Reports first because you already know what it'll say. Honda has been building the J35 since 1999, dropped it into everything from the Odyssey to the Ridgeline, and accumulated 27 years of precisely the reputation an automaker would sell its factory floor to buy. Then NHTSA opened an investigation into 1,410,806 of them because the connecting rod bearings are eating themselves alive.^[1]

The root cause is almost comically mundane for something that destroys engines. Somewhere in Honda's manufacturing process, the equipment that grinds crank pin journals drifted out of specification, producing pins with a convex crown shape instead of a true cylinder. That means the connecting rod bearing sits on a high point instead of distributing load across its full surface, metal contacts metal where oil film should separate them, bearings wear until they seize, and an engine dies, possibly with a connecting rod punching through the block at 70 mph on I-95 while your kids argue about Minecraft in the back seat.^[2]

Honda's original recall, 23V-751, covered 248,999 vehicles: the 2015–2020 Acura TLX, 2016–2020 Acura MDX, select Honda Pilots, Odysseys, and Ridgelines. Honda estimated only 1% of flagged vehicles actually contained the defective crankshaft, and dealers would inspect and replace the engine if needed, which sounded straightforward enough.^[3]

Except it wasn't straightforward at all. After the recall closed, 173 owners filed complaints describing connecting rod bearing failures with symptoms identical to the recalled defect, and Honda denied every single one of them.^[4] Not most, not a majority, but all 173 complaints that NHTSA's own Office of Defects Investigation said "displayed failures with similar characteristics to those addressed in the original campaign." That perfect denial rate is what prompted NHTSA to expand the investigation from a quarter million vehicles to 1.41 million, spanning the 2016–2020 Acura MDX, 2018–2020 Acura TLX, 2016–2020 Honda Pilot, 2018–2020 Honda Odyssey, and 2017–2019 Honda Ridgeline.

By August 2025, NHTSA had accumulated 3,012 total complaints covering 414 reports filed directly with the agency and 2,598 from Honda's own records, alongside seven crashes, no injuries, and no fatalities. "No fatalities yet" is the motto of every ticking mechanical time bomb that hasn't detonated in the right lane at rush hour, and a company batting a thousand on complaint denials is not the kind of company you want holding the timer.

What makes this defect perverse is which vehicles it targets. Run these models through FARS data and you find some of the safest, most conservatively driven vehicles on American roads: the Honda Odyssey carries a fatality rate of 0.93 per 100 million vehicle miles traveled and an impairment rate of just 15.4% among drivers in fatal crashes, which is the lowest of any van in the entire dataset.^[5] The Pilot sits at 0.29 deaths per 100 million VMT, the Ridgeline at 0.24, the Acura MDX at 0.30, and the TLX at 0.11. These aren't Challengers being wrapped around telephone poles by drunk 22-year-olds; they're the vehicles that parents buy specifically because they want the opposite of that.

Sober families running highway miles with children secured in LATCH-compatible car seats, behind a driver who chose the Pilot because their mechanic said "Honda V6, can't go wrong," represent the demographic least prepared for a sudden engine seizure at highway speed that turns their vehicle into a stationary obstacle in a 65-mph traffic stream.

Twenty-Seven Years of Trust, Three Years of Silence

Honda received its first U.S. complaint about J35 bearing failures in February 2020, and the recall didn't arrive until November 2023, which means three years and nine months of complaints accumulated while 1.4 million vehicles logged an estimated 63 billion miles on potentially defective crankshafts. The arithmetic behind that number: 1.41 million vehicles, roughly 12,000 miles per year on average, across 3.75 years of exposure.

For context, the J-series architecture launched in 1998 with the third-generation Accord, which means 27 years of production and a quarter century of warranty data on an engine Honda knew backward, forward, and in cross-section. This wasn't a first-generation growing pain or an experimental architecture learning its limits; it was a quality control failure in the most mature V6 platform Honda has ever built, one where the grinding machine that produced crankshaft journals was a machine Honda owned, in a factory Honda operated, producing parts for engines Honda designed.

The specific variant is the J35Y6, a single-overhead-cam design with VTEC on intake valves only, which shouldn't be confused with the J35Y8 that moved to dual overhead cams and variable timing control. But the distinction hardly matters when the failure mode is purely mechanical: a machine ground crankshaft journals out of spec, and no amount of variable valve timing sophistication can compensate for a bearing surface that shouldn't exist.

The Teardown

On June 17, 2026, the YouTube channel I Do Cars published a teardown of a J35 V6 pulled from a 2020 Honda Odyssey with approximately 85,000 miles on the odometer. The engine was catastrophically failed, with scored bearing surfaces, metal contamination throughout the oil system, and a connecting rod bearing that had effectively welded itself to the crankshaft journal.^[6] Eighty-five thousand miles is less than seven years of average driving on an engine architecture that regularly runs past 200,000 in proper working order, which means the J35's reputation for longevity is surviving on the back of units that happened to receive properly ground crankshafts.

What This Means for You

If you own a 2016–2020 Honda Pilot, 2018–2020 Odyssey, 2017–2019 Ridgeline, 2016–2020 Acura MDX, or 2018–2020 Acura TLX with the 3.5-liter V6, check your VIN at nhtsa.gov/recalls. If your vehicle was part of the original 23V-751 recall and you received the repair, you're likely covered by the engine replacement remedy. If you weren't included in the original recall scope and you hear a new knocking noise, low oil pressure warnings, or rough idle at operating temperature, document everything with dates, mileage, and audio recordings if possible, then file a complaint directly with NHTSA at nhtsa.gov/report-a-problem. Honda's 100% denial rate on post-recall complaints means the dealer may tell you it's not covered, but NHTSA apparently disagrees with that assessment, and your complaint becomes part of the data that determines whether a mandatory recall follows.

Limitations

FARS data covers fatal crashes only, not all crashes or mechanical failure incidents, so the fatality rates cited reflect crash outcomes rather than engine reliability metrics. Honda's 1% estimate for defective crankshafts in the original recall population may be accurate for that specific production window, but the expanded investigation covers a broader population where the actual defect rate remains unknown. The 63-billion-mile exposure window calculation uses average VMT assumptions from FHWA data, not actual odometer readings, which introduces uncertainty of roughly ±15% for fleet-level estimates. The seven reported crashes attributed to engine failure may undercount incidents where bearing seizure was a contributing factor but not the coded primary cause in the crash report.

The Strongest Case for Honda

Honda identified the problem, initiated a voluntary recall before any injuries were reported, and offered free engine replacements to affected owners. Zero fatalities and zero injuries across 3,012 complaints is a genuinely good outcome for a defect affecting over a million vehicles, and the J35 remains, in aggregate, one of the most reliable V6 engines ever mass-produced. The overwhelming majority of the 1.4 million vehicles under investigation will never experience a bearing failure because the manufacturing drift affected a subset of production, not the entire run. Honda's denial of post-recall complaints could reflect a legitimate engineering determination that those specific engines didn't carry the crankshaft defect rather than a blanket corporate stonewall.

All of that can be true, and it can simultaneously be true that denying 173 out of 173 complaints exhibiting "similar characteristics" is a number that strains credulity, and that three years and nine months between first complaint and first recall is a long time to wait when the failure mode is "engine stops working on the highway while a family of five is inside it."