Goldman Sachs expects Amazon, Alphabet, Microsoft, Oracle, and Meta to collectively spend $1.2 trillion on AI infrastructure in 2027. This is over 50 percent higher than the roughly $800 billion projected for this year, according to Bloomberg, citing strategist Ryan Hammond.
The investment would represent the largest infrastructure spending cycle since railroad construction in the 19th century relative to GDP. However, growth is expected to slow, with the pace dropping from nearly 100 percent in 2026 to 54 percent in 2027 and 12 percent in 2028.
To recoup these outlays, the companies would need about $300 billion in annual AI revenue. While current earnings still fall short, cloud revenue growth has surged from 25 percent in 2024 to 48 percent in Q2 2026.
"Spending now exceeds what the companies generate from ongoing operations, which means more debt financing," said Ryan Hammond, according to Bloomberg. This could further strain financial resources as bottlenecks in power, labor, and memory chips may slow progress.
The announcement follows Goldman's June warning that consensus estimates were far too low. The firm also highlighted that AI labs like OpenAI and Anthropic are central to these expectations and financial instruments backing the buildout.
Goldman did not specify how AI revenue growth at these labs will meet investment costs, and it remains unclear whether the pace is sufficient to justify the scale of spending. The firm said the outlook will continue to be closely watched.
Source: thedecoder
