A shopping surge. No meaningful change in the measured fleet.
More packages suggest more vehicles, more pressure and more hurry. Across the four days of Prime Day, the measured traffic volume was almost identical to its June baseline: 842,800 rides against 843,064 expected.
−0.03%
Total driving volume
Prime Day vs matched weekdays
264
Fewer rides
Out of more than 843,000 expected
The rounded difference is three-hundredths of one percent.
842,800
Prime Day rides
843,064
matched-weekday baseline
4 days
June 23–26, 2026, US rides
Total US rides
Same zero-based scale: 0–900,000 rides. The near-identical bar lengths are the finding.
Compare like with like
Each Prime Day weekday was compared with the same weekday in the rest of June.
Small differences. No clear safety signal.
Sudden braking and collision-level impacts changed slightly. The source analysis does not identify either increase as a statistical anomaly.
Sudden braking, events per 1,000 rides
Zero-based scale: 0–60 events per 1,000 rides.
Collision-level impacts, events per 1,000 rides
Zero-based scale: 0–5 events per 1,000 rides. Each metric uses its own labeled scale.
A measured difference is not a proven effect. These are event rates in the observed fleet, not an assessment of every delivery driver.
The trucks we saw held steady, too.
Truck activity showed no obvious surge. The measured trucks made roughly the same number of daily trips, with a slightly longer average journey.
Truck rides per day
Average miles per truck trip
Zero-based scales: 0–18,000 daily rides and 0–80 miles per trip.
Regional and long-haul, mostly
Trips averaging about 72 miles do not describe the short loops of last-mile delivery.
The view is large. The boundary matters.
93% of the observed fleet is taxi and rideshare; the remaining 7% are other observed vehicles. Two explanations fit these numbers, and this cut cannot separate them.
93%
Taxi and rideshare
Share of the observed fleet
~11%
Thursday, June 25
Collision-level impacts above the Thursday baseline
Prepared for the surge
Routes, capacity and drivers can be planned months ahead. A prepared fulfillment network may absorb the demand.
Outside this view
Last-mile vans and cargo bikes are largely outside the sample. Their response may be different.
A narrow, useful conclusion: no meaningful change in the measured fleet. What happened to delivery drivers specifically remains unanswered.
One day to watch, not a pattern yet
Thursday, June 25 was the one day with an upward movement in collision-level impacts, about 11% above the Thursday baseline. The sample is not large enough to establish an effect. Several Prime Days would be needed to see whether Thursday means anything.
What was measured
US rides, June 23–26, 2026. Baseline: the same weekdays in the rest of June. Event rates are normalized per 1,000 rides. 842,800 rides, measured against matched weekdays, put a number on how little changed: −0.03%.
Knowing when the road did not change is as useful as knowing when it did.
Source: Empiric Earth Prime Day data pull, June 2026. US rides, June 23–26, 2026, against the same weekdays in the rest of June; event rates normalized per 1,000 rides.