GM’s EV decline exceeds its overall sales decline: how the comparison works

The decline is concentrated, but the explanation has limits

General Motors reported 670,974 US deliveries for the third quarter of 2026, compared with 710,347 in the corresponding quarter of 2025. The detailed table shows a 5.5% decline; the October 1 headline release rounds the decrease to 6%. Both quarters had 77 selling days, so a different number of selling days does not explain this particular comparison.

Adding the electric-vehicle model rows gives 25,473 deliveries against 66,501 a year earlier. The EV decline is 41,028 vehicles, larger than the 39,373-unit decline in the company total. Subtracting those EV totals from all deliveries leaves 645,501 versus 643,846 for the remaining vehicles, an increase of 1,655, or about 0.3%. This reconciliation identifies where the unit change occurred. It does not, by itself, identify why customers changed their purchases. The comparison covers US deliveries, not GM’s worldwide vehicle sales.

The prior year is part of the story

GM’s October 1, 2025 release described the earlier quarter as a record for EV deliveries and linked an industry-wide buying surge to customers seeking the federal tax credit available at that time. That is historical context for an unusually strong comparison period, not a statement about current tax eligibility.

A high base makes a subsequent percentage decline larger than a comparison against a more typical quarter might show. It does not make the lost deliveries imaginary. Investors should retain the actual unit decline while asking how much reflected timing, product availability, competition, pricing or changes in incentives. The sales table cannot allocate the decline among those explanations.

An offset can make a contribution exceed 100%

Dividing the EV decline of 41,028 by the overall decline of 39,373 gives approximately 104.2%. This is a contribution to a net change, not an EV market-share figure. It exceeds 100% because the remaining vehicles provided a positive offset.

A research memo should label both the numerator and denominator. Describing the calculation as an EV share would be incorrect. Share uses total deliveries within a period as its denominator; contribution to change compares changes between periods. Keeping the two measures separate prevents a dramatic but meaningless percentage from becoming the central investment claim.

Reconcile product definitions before using growth rates

GM’s presentation emphasizes its highest-volume vehicles and includes a chart explicitly limited to internal-combustion models. The detailed delivery table separately lists electric versions and different Silverado and Sierra variants. A growth statement about selected variants should not be silently applied to every vehicle carrying the same name.

The distinction also matters when aggregating the table. Adding individual Silverado rows and then adding the TOTAL Silverado row again would double count those vehicles. The EV reconciliation uses the electric model rows once each, with the corresponding prior-year figures. The Bolt row has no prior-year quarterly sales entry; a zero-base comparison does not support a conventional finite percentage-growth claim.

Sales are an input to the earnings question

The remaining question is what the changed sales mix means for profitability and cash flow. Unit data alone do not supply realized transaction prices, manufacturing costs, warranty expense or the timing of revenue recognition. A relatively stable non-EV total can coexist with meaningful shifts among higher- and lower-margin vehicles.

The next financial update should therefore be read for the bridge from sales mix to revenue and operating results. Relevant checks include incentives, production adjustments, inventory and management’s explanation of EV costs. A reduction in EV deliveries is not automatically an equal percentage reduction in EV losses. Fixed costs, production decisions and product mix can change that relationship.

The practical conclusion is that EV deliveries explain more than the net decline in this US comparison, with a small offset elsewhere. The prior-year record provides useful context but does not establish the size of any timing effect. These observations come from original GM documents, all from one issuer, rather than independent confirmation of management’s interpretation.