When the record goes quiet

FMCSA retired its legacy registration stack on May 14, 2026 and cut over to Motus, its new USDOT Registration System. The public data split: the old Socrata datasets froze that day, and a new family of Motus-native datasets began publishing daily. The catch, which surfaced on July 8, 2026, is that the new feeds can silently omit entire carriers that exist inside Motus. Not lagging behind, absent, with no error and no flag. FMCSA's own QCMobile API shows the same gap. So a carrier with active, verifiable insurance can read as uninsured across every public data surface at once, including the tools brokers vet with. This is a transition artifact, dated and reproducible, and it corrects itself once FMCSA's exports catch up. Until then, absence of a record is not evidence that the record does not exist.

What changed on May 14

Motus is FMCSA's replacement for decades-old registration plumbing: the parts of the agency's systems that handle operating authority, insurance filings, and new entity registration. The main phase finished on May 14, 2026 at 8:00 PM ET, when the legacy stack was retired.

FMCSA did not migrate its public open data in place. It forked it. The legacy Socrata datasets on data.transportation.gov kept publishing on their daily schedule, but stopped receiving any new licensing or insurance activity after the cutover. Alongside them, FMCSA stood up a parallel family of Motus-native datasets that carry every authority and insurance action since May 14 and grow daily. The insurance file in that new family (Motus Insur, c5y8-a4uz on data.transportation.gov) is where a filing made through Motus is supposed to appear.

The correct response to that fork is to read both families and combine them. That much was understood. What was not visible until early July is that the new family itself can be incomplete.

The case that proved it

A small interstate carrier had its insurer file a liability policy (a BMC-91, the standard federal filing) through Motus in early June 2026, effective June 2, at the $750,000 minimum for general freight. The filing was received and confirmed, and the carrier could see it listed as active the entire time in its own Motus account at motus.dot.gov.

Five weeks later, that filing was absent from every public FMCSA data surface. Not in the frozen legacy datasets, which was expected. Not in the live Motus-native datasets, which had updated that same morning and together held 15,021 insurance filings for June 2026. Not returned by FMCSA's public QCMobile API, which reported no liability insurance on file. The carrier's Motus entity had existed since mid-May. The filing existed. Neither propagated to the exports.

The consequence is the part that matters for brokers. Every tool that reads FMCSA's public data, fleetfax and its competitors alike, showed an insured carrier as carrying no insurance. There was no error to catch. Each tool did exactly what it was built to do with the data it was given. The data was missing a carrier.

The data was missing a carrier.

One detail confirms which side is correct. FMCSA never initiated a revocation. When required insurance genuinely lapses and no replacement is filed, authority is revoked on a roughly eight to nine week clock. No such clock was running here, because inside FMCSA's own system of record the carrier was insured the whole time. The public exports were the only place the coverage did not exist.

Why it is whole carriers, not stray fields

The omission is entity-level. The affected carrier is missing from the entire Motus-native dataset family at once, not just the insurance file but the master carrier file that lists the cohort of entities processed through Motus. A single dropped insurance record would be a narrow bug. A carrier absent from every table in the new family at the same time points to something shared upstream.

The working hypothesis, and it is a hypothesis, is that FMCSA's export pipeline draws the Motus-native datasets from a common extract, and that some entities are not selected into it. Entities created or migrated around the cutover window are the likeliest candidates to fall out. That is a reasonable read of the pattern, not a proven mechanism, and it is FMCSA's pipeline to diagnose. What is established is the observable fact: specific carriers that are active inside Motus do not appear in the Motus-native public exports, and the boundary of the omission is the whole carrier.

Both directions of the failure

The proven case runs one way: a carrier that is insured but reads as uninsured. That is unfair to the carrier and wastes a broker's time, but it is the safer of the two directions, because a cautious broker double-checks a bad-looking result before walking away.

The same gap can run the other way, and that direction is worse. If a carrier's insurance is cancelled and the cancellation is recorded only inside Motus, and it does not propagate to the public exports, then every public-data tool keeps showing the carrier as insured after its coverage is gone. We have not observed a confirmed instance of this direction. It is a logical consequence of the same mechanism, and it is the one worth watching, because in that direction absence in the feed would hide a real cancellation rather than a real filing.

How big the exposed set is

In fleetfax's ingest of the public FMCSA data, about 9,850 carriers with active authority currently look insurance-lapsed: a recent cancellation and no replacement filing visible in any public feed. That population is not all victims of the gap. A real share of it is genuinely uninsured, which is a documented pattern in its own right, and the public data reads those carriers correctly.

To separate the two, we verified the set carrier by carrier against Motus, all 9,849 of them. 541 (5.5%) were confirmed to hold active insurance inside Motus despite reading as lapsed in public data. Those are the carriers the exports left out. The rest had no coverage in Motus either, and the public data was right about them. The live check sorts one from the other; the public feed alone cannot.

What fleetfax did

fleetfax vets carriers off FMCSA public data, so this gap hit its results the same way it hit every other tool reading that data. The response was to stop treating public-feed absence as ground truth for this class of carrier. For each carrier that looks lapsed, fleetfax now verifies insurance status live against Motus, the system of record, and where Motus shows active coverage the exports are missing, it corrects the carrier's page and labels the correction with its provenance. Those corrections retire themselves automatically once FMCSA's exports catch up, so nothing lingers as a stale override. It is a bridge across a transition gap, not a permanent second source.

What brokers should do

Absence of a record in bulk data is not absence of the fact, and during a system transition it is especially not. A few practical steps:

• When a carrier disputes an insurance result, check the source of record before concluding. The carrier can show its own Motus account at motus.dot.gov, which reflects filings the public exports may not yet carry. FMCSA's registration help line is 1-800-832-5660.
• Treat a fresh "no insurance on file" on an otherwise clean, active carrier as a question, not a verdict. In this window it can mean the filing is stuck upstream rather than that coverage is gone. Verify rather than assume.
• Remember the reverse can happen too. A carrier still shown as insured may have a cancellation recorded only inside Motus, so during the transition a public "insured" reading is not on its own a guarantee if something else about the carrier prompts caution.

What affected carriers should do

If your coverage is active in your Motus account but a lookup shows you as uninsured, the fix lives upstream. Ask your insurer to confirm the filing was accepted in Motus, and push FMCSA to propagate your record into the public exports (registration help line 1-800-832-5660). Fixing the source fixes every downstream application at once, which is faster and more durable than asking each tool to patch you individually.