Goldman Sachs economists said on Tuesday that the global bill for adopting artificial intelligence agents - beyond just chips and data center hardware - could top $1 trillion in the coming years. Their analysis emphasizes that spending will extend into data infrastructure, software, and organizational changes required to integrate AI into business operations.
The bank noted that U.S. firms are already incurring heavy labor-related costs as part of the AI shift, estimating current annual spending tied to the transition at about $150 billion. Goldman’s Global Economics Analyst report points to company disclosures and accelerating top-line growth at data-intelligence and cloud-service providers as indicators that these broader investments are gaining pace.
Organizational and workforce costs
Goldman quantifies organizational capital needs in two ways. First, the bank uses executive time allocation to infer a recurring organizational capital investment, estimating roughly $40 billion per year. Second, the bank projects the cost of workforce reorganization across the full AI adoption cycle at between $800 billion and $900 billion, based on labor restructuring expenses observed so far.
Combined, these and related complementary expenditures support the bank’s estimate that non-hardware investment in AI will reach about $1 trillion globally. Goldman frames these expenditures as intangible capital - investments in things like organizational processes and software management rather than physical plant and equipment.
Intangibles and productivity measurement
Using EU KLEMS data, Goldman notes that intangible capital has grown to roughly the same scale as traditional capital expenditure in G10 economies. Much of the rise in intangible investment over the last 20 years, the report says, has been driven by increased spending on organizational capital and software management.
The bank adds that this type of intangible investment typically follows a productivity curve in which firms initially direct resources inward, building the organizational and technical foundations needed for new technology adoption. Because many of those internal investments are not captured well by gross domestic product measures, Goldman argues that recent acceleration in U.S. productivity growth could be understated.
Corporate outcomes and market implications
Goldman’s analysis indicates that companies that deploy intangible capital more effectively have historically captured larger revenue shares, achieved higher productivity, and realized better returns on investment - in part through reduced labor costs. The bank’s equity analysts have previously concluded that firms concentrating on data structure and AI deployment will be critical to unlocking AI’s economic value. They also suggest that increased market concentration and lower labor costs could raise valuations for companies that invest most effectively in AI agents.
Summary of key numbers and claims:
- Estimated global non-hardware AI-related investments could exceed $1 trillion.
- U.S. companies are estimated to be spending about $150 billion per year on labor costs tied to the AI transition.
- Executive time allocations imply roughly $40 billion per year in organizational capital investment.
- Workforce reorganization costs across the full AI adoption cycle are projected at $800 billion to $900 billion.