Big Tech's AI Funding Shifts From Own Cash to Wall Street Money
Tech giants are moving from using their own profits for AI infrastructure to deploying shareholder and bondholder capital to fund other companies' projects.
The AI investment cycle began with big tech allocating internal profits to build chips and data centers for their own use. A second stage emerged as companies such as Alphabet, Meta, Oracle, Amazon and Intel raised equity and debt, directing shareholder and bondholder capital toward internal expansion. The current phase involves firms like Nvidia gathering a $500 billion pool from Wall Street investors to underwrite OpenAI's Ohio data-center lease and enable startups to rent Nvidia chips via CoreWeave.
Broadcom is using Blackstone funds to help Anthropic acquire chips, while Google channels bond proceeds to support Fluidstack's compute purchases from TeraWulf. This trend reflects Milton Friedman’s warning that spending other people’s money on other people reduces cost vigilance, a risk that analysts say could affect the efficiency of AI financing.
Why it matters
The shift to investor-funded AI projects could alter cost discipline and risk allocation across the tech sector.
In this story