← All forecasts · October 19

TSLA / Q3 2026

Tesla, Inc.

Reported: October 21, 2026

Published forecast

2026-10-19T23:34:42.849315+00:00

Opus 5.5 High in Shortcut

Shortcut

0.38

Shortcut normalized EPS

-15.8% vs Street

Off by $0.10 per share

Street

0.45

Street normalized EPS

Frozen October 18 snapshot

Off by $0.03 per share

Real

0.48

Issuer non-GAAP diluted EPS

Street win

Reported result · October 21 ↗ · Issuer non-GAAP diluted EPS was $0.48; GAAP diluted EPS was $0.22. The non-GAAP reconciliation excludes stock compensation, digital-asset and SpaceX investment gains/losses, and specified tax items.

USD per Tesla diluted common share. Percentages use unrounded forecasts.

Model download

Explore the assumptions and calculations behind the forecast.

Open in Shortcut ↗Download XLSX ↓

Shortcut’s reasoning.

The model estimates Q3 revenue of $28,583.537 million from the automotive, services and energy builds. Gross profit of $5,015.186 million less $4,450 million of operating expense gives operating income of $565.186 million.

For normalized earnings, operating income plus $410 million of interest income, less $90 million of interest expense and $259.742 million of assumed foreign-exchange loss, plus $1,220 million of excluded stock compensation, gives normalized pre-tax income of $1,845.444 million.

Applying the assumed 26.5% normalized tax rate and $14 million of noncontrolling interests gives normalized net income of $1,342.401 million. Dividing by 3,544 million diluted shares gives non-GAAP diluted EPS of 0.37878136581147626, versus the frozen Street estimate of $0.45.

The accepted public export preserves all 122 cells in the scored EPS dependency chain. Bitcoin and SpaceX mark adjustments cancel from the normalized EPS calculation; the foreign-exchange loss remains in it. The separate 30% beat probability is a judgment and does not multiply forecast EPS.

On the record.

Reporting basis
Tesla non-GAAP diluted EPS / frozen CapIQ normalized EPS
Forecast completed
2026-10-19T21:48:48.050415+00:00
Published
2026-10-19T23:34:42.849315+00:00
Model assumptions and review notes
  • MEDIUM: Quarter-end market inputs were retrieved after the information cutoff, and a pre-cutoff capture timestamp for the USD/CNY closes is not preserved. Those original values remain fixed. The foreign-exchange loss reduces normalized EPS by about $0.054; a $0.01 change in the closing CNY exchange-rate input changes EPS by roughly $0.0064. No immutable historical capture is claimed.
  • The future diluted-share estimate of 3,544 million is a judgment: 3,240 million basic shares plus 304 million of dilution. The cached Q2 filing documents roughly 304 million CEO options exercised, less 17.5 million shares net-settled, and a service condition through January 19, 2028. Its EPS policy excludes restricted awards from diluted shares until deemed earned. The model does not add the roughly 194 million net common-share increase to its weighted share estimate.
  • Bitcoin and SpaceX carrying-value assumptions affect GAAP earnings and cancel from the current normalized EPS calculation. This does not establish the accuracy of those GAAP marks or make the foreign-exchange assumption irrelevant.
  • The current certain-tax-item adjustment is zero. A latent bridge omission would affect a future scenario that entered a nonzero certain-tax-item adjustment. The fixed forecast and all scored EPS dependencies are unchanged; broad scenario accuracy is not claimed.
  • Organic-growth diagnostic rows use a multiplicative convention although the foreign-exchange adjustment is expressed in dollars. Forecast organic growth is about 1.01975% on that convention versus 1.02701% on an additive convention. These diagnostic rows do not drive the scored EPS calculation.
  • Some historical share-dilution, tax-rate and cash-flow comparisons are not computed even though comparative inputs are available. Q4 segment depreciation memos remain blank; simple full-year-minus-quarter arithmetic would imply about $976 million for Automotive and $93 million for Energy, subject to disclosure precision. Total forecast depreciation is not changed.
  • Annual lease-termination and revenue-per-fleet ratios have an annual scale, unlike quarterly columns; nearby notes disclose the difference. The long-term energy-margin comparison uses a single reported quarter affected by a $240 million warranty item. It does not establish a miss against long-term guidance.
  • The balance sheet uses simplified allocations, including charging all depreciation to property, plant and equipment and routing noncash foreign-exchange and interest adjustments through accrued liabilities. Period-end shares use a disclosed convention. Internal balancing identities are not independent validation.
  • The historical beat prior of 0.78 with eight pseudo-observations and the final 30% probability are heuristics. The exact supplied experimental v2 playbook is distinct from the frozen v3 skill. Six historical quarters contain two strict beats and give a mechanical posterior of about 58.86%, before the separate final judgment.
  • Minor presentation limits remain, including clipped wrapped source descriptions and contextual period headers on reference tabs. Full text is preserved in the workbook. No native Excel visual verification is claimed.
Research sources