Tesla reports third-quarter production and deliveries on Friday, 2 October 2026, and the banks covering it cannot agree on what the quarter looked like. The published estimates run from 421,758 to 482,000 cars — a spread of about 60,000, or roughly a seventh of the quarter.
What the banks actually filed
| Firm | Q3 2026 estimate | Note |
|---|---|---|
| JPMorgan | 482,000 | cut from 516,000 |
| Barclays | 475,000 | Dan Levy |
| UBS | 470,000 | — |
| StoneX | 446,500 | — |
| Goldman Sachs | 435,000 | cut from 490,000 |
| Cantor Fitzgerald | 421,758 | lowest published call |
The median of those six is about 458,000. The aggregated consensus figures do not agree either: Visible Alpha sits at 449,000 to 454,000 depending on when it is sampled, while Bloomberg's compilation is 466,000. Goldman's cut was the subject of its own downgrade earlier this month, when the bank also took 40,000 off its fourth-quarter number.
Why nobody should be confident
The reason to treat all of this carefully is what happened three months ago. Tesla delivered 480,126 cars in the second quarter against a company-compiled consensus of 406,024 — the Street was short by roughly 74,000 units, a bigger error than the entire spread of today's estimates. Analysts modelling the same company with the same public data missed by about 18%.
That is the structural problem with quarterly delivery forecasting at Tesla. There is no monthly reporting to anchor on, regional registration data arrives late and covers only part of the world, and end-of-quarter push has repeatedly moved volume across the reporting line.
Down on last year, and up in Europe
Tesla delivered 497,099 cars in the third quarter of 2025. Even JPMorgan's 482,000 — the most optimistic published figure — would be about 3% below that. The mid-point calls imply a decline closer to 8%.
Set against that, the European picture is the opposite. ACEA data for January to August has Tesla at 142,165 EU registrations, up 65.9% year on year, in a quarter where battery-electric cars outsold petrol across the bloc for the first time. Europe is not where a weak global number would be coming from, which makes the regional split inside Friday's release more informative than the headline.
What to watch on Friday
Three things decide how the number reads. First, the split between the Model 3 and Model Y line and the "other models" bucket, which now carries Cybertruck, Model S, Model X and the Semi. Second, the gap between production and deliveries, which shows whether inventory is building. Third, energy storage deployment, reported in the same release and increasingly the part of the business that moves the stock.
A figure anywhere inside a 60,000-car band was always going to be called a beat or a miss depending on which consensus a commentator reaches for. The Q3 2025 comparison is the harder test, and every published estimate fails it.