have been curious about the actual numbers behind all the ai hype. pulled the goldman sachs research on generative ai impact from 2023 which projected a significant global gdp boost over ten years. then found the 2026 predictions from cio marco argenti about personal agents and the gigawatt ceiling.
interesting that goldman economists recently walked back near term impact saying ai contributed basically zero to us economic growth in 2025 despite the massive capital expenditure wave. but they still maintain the long run forecast of boosting productivity growth by one and a half percentage points annually over a decade. the task level impacts are around thirty percent median gain for customer support and software coding specifically.
the infrastructure projections are wild too with cumulative capital expenditure estimates in the trillions globally between twenty twenty six and twenty thirty one. the power consumption angle is becoming a real constraint with data center demand way up by twenty thirty. feels like we are in the trough of disillusionment but the decade long structural bet is still there.