A new analysis from Bloomberg reveals that despite massive market valuations and widespread adoption of artificial intelligence technology, concrete evidence of AI-driven productivity improvements remains limited. The article examines the disconnect between AI market enthusiasm and real economic impact.
While companies like Nvidia have seen dramatic market capitalizations, overall US worker productivity has grown at just 1.86% annually since the pre-pandemic period. This rate falls well below the 3.3% growth seen during the internet revolution of the mid-1990s to mid-2000s.
Experts present three possible scenarios for AI’s economic impact:
- Goldman Sachs projects a modest 2.3% GDP increase by 2034,
- while McKinsey estimates a more optimistic 5-13% boost by 2040.
- On the pessimistic side, Nobel laureate Daron Acemoglu expects only a 1% GDP contribution within a decade.
Recent research shows unprecedented AI adoption rates, with 40% of US adults using AI within two years of its introduction – far faster than previous technologies like personal computers or the internet. Some studies demonstrate promising results in specific applications, such as programmers using GitHub Copilot completing tasks 56% faster, and customer service agents with AI assistance resolving 14% more issues per hour.
However, historical precedent suggests technological revolutions take time to manifest in productivity data. The personal computer’s impact wasn’t visible in economic statistics until a decade after its introduction, as noted by economist Robert Solow’s famous 1987 observation.
The recent emergence of DeepSeek, a Chinese AI firm claiming to develop advanced models at significantly lower costs, has disrupted market expectations and triggered a selloff in US tech stocks. This development could accelerate AI’s economic impact by reducing implementation barriers.