- Get the latest hedge fund news – sign up for the FREE AFI Briefing
- 2026 AFI 500: Mapping the global hedge fund industry
- Join the AFI Hedge Fund Innovation Summit, 8 October 2026
Hedge funds are taking their reassessment of the artificial intelligence trade a step further, shifting attention from whether AI investment can generate sufficient revenues towards how the infrastructure boom will be financed.
The latest positioning data from Hazeltree shows long participation falling across the Magnificent Seven in August as rising debt issuance and financing costs became a greater consideration for investors.
The shift follows July, when Hazeltree found hedge funds rotating away from AI infrastructure towards companies demonstrating clearer paths to monetisation, with managers making targeted adjustments rather than broadly abandoning the theme.
In August, no Magnificent Seven company recorded an increase in the number of hedge funds holding long positions. Amazon, Tesla, Meta, Alphabet and Apple all saw meaningful reductions, while declines for Microsoft and Nvidia were more modest.
Alphabet positioning reverses
The most pronounced shift came at Alphabet, where the ratio of long to short hedge fund holders fell from 1.70 in July to 0.92 in August – the first time this year that more funds have been short the Google parent than long.
Long participation declined by around 25% during the month while the number of short holders jumped approximately 38%. The reversal extends a trend stretching back to May, with long participation falling and short participation increasing in every month since then.
Apple also moved below parity, with its long-to-short fund count ratio falling from 1.44 to 0.81.
The Alphabet figures do not, however, point to a wholesale exit by existing bulls. Net long exposure declined only modestly from 86.8% to 84.7%, suggesting that managers retaining positions continued to hold significant exposure.
AI financing moves into focus
Hazeltree said investor attention is increasingly turning to the financing of the AI infrastructure build-out as companies make greater use of debt to fund escalating capital requirements.
Alphabet returned to the bond market with $25bn of senior notes in August, while Hazeltree said AI-related equities are becoming more sensitive to interest rates and credit conditions as external financing plays a larger role.
Semiconductor positioning also softened. The proportion of 30 major US-traded semiconductor companies exhibiting net long positioning fell from 70% in July to 66.7% in August. Applied Materials remained the most crowded semiconductor long, followed by Nvidia and Lam Research.