MIT Technology Review says hyperscalers may need to nearly triple productivity by 2030 just to break even on their trillion-dollar AI infrastructure bet. According to BeInCrypto, Wharton finance professor Jessica Wachter and a coauthor estimated nearly $1.1 trillion in data center spending through 2027 across Alphabet, Microsoft, Amazon, Meta and Oracle, and said hyperscaler productivity must rise 2.7 times to cover costs. Wachter called the buildout “the largest misallocation of capital in history.”