×

AI hyperscalers upend usual earnings stock swing pattern

By Thomson Reuters Aug 11, 2026 | 11:14 AM

NEW YORK, Aug 11 (Reuters) – Larger-than-usual post-earnings stock swings by AI and hyperscaler companies this earnings season have upended a historical pattern in which smaller companies’ results typically ​drive the larger stock moves, data from options analytics ‌service ORATS showed.

The first few weeks of earnings seasons are typically dominated by major market leaders, including financial giants like JPMorgan Chase and Wells Fargo, early-reporting Dow Jones blue chips, and mega-cap tech powerhouses like Apple, Microsoft, ‌Alphabet, ​and Meta. Smaller companies in the ⁠index take center stage in ⁠later weeks.

Thinner liquidity, smaller floats, less institutional ownership and sparser analyst coverage relative to larger companies combine to make small-cap earnings reactions typically larger in magnitude than those of mega-caps.

Not ​so this time.

“The smaller companies in recent weeks have had more muted earnings moves – a contrast to the early weeks ⁠when AI and hyperscalers’ reports helped ⁠drive larger-than-usual moves,” ORATS founder Matt Amberson said.

Some ​of the largest hyperscalers, including Amazon, Microsoft, Google and Meta, have ​produced big post-earnings stock swings, surpassing their respective average ‌moves in past quarters, as investors aggressively cheered or sold off shares depending on whether AI capital expenditures appeared to be paying off.

For companies reporting in the first week of the second-quarter ⁠earnings season, which kicked off in mid-July, buying options straddles — a strategy combining the purchase of a put and a call — fetched the ⁠largest average gains, with ‌profits shrinking in weeks two through four, ⁠ORATS data showed.

Week one gains averaged 23%, compared ​with ‌an average loss of 2% for the strategy ​in the ⁠first week of earnings over the last 12 quarters, the ORATS analysis showed. In contrast, for the fourth week of results, the strategy produced an average loss of 6% compared with the historical average of a loss of 5%.

(Reporting by Saqib Iqbal Ahmed; Editing ​by Mark Porter)