AI’s trillion-dollar buildout faces a demanding economic test
AI hyperscalers are committing unprecedented sums to data centers, raising questions about whether future revenues and productivity gains can justify the spending. Jessica Wachter of the University of Pennsylvania’s Wharton School and a collaborator estimate that expenditures could approach $1.1 trillion by 2027. Their accounting-based analysis says the companies would need to increase their productivity 2.7-fold to break even by 2030, including financing costs, a 15% return and asset depreciation.
The companies—Alphabet, Microsoft, Amazon, Meta and Oracle—are expected to spend about $750 billion on data centers this year, while AI revenues may reach only $150 billion to $200 billion, according to Gary Gensler of MIT’s Sloan School. Some projections put their combined AI capital investment above $5 trillion over the following four years. Alphabet’s latest quarter included nearly $120 billion in revenue but a free-cash-flow deficit of about $5.9 billion after infrastructure spending.
Analysts say success depends on rapidly growing demand, economy-wide productivity improvements and continued competitiveness of costly frontier models against cheaper alternatives. Borrowing has also spread exposure to banks, funds, insurers and pension investors. If demand weakens or productivity gains fail to appear, data centers could become stranded assets and the investment could damage companies and the broader US economy.
Written by Sakhanda Wire AI from reporting by MIT Tech Review. Facts, quotes and full context are in the original article below.