[Bernstein Breaks Down AI Infrastructure Costs: Up to $39.5 Billion Investment Per Gigawatt] A Bernstein research report shows that building a 1GW data center with different AI accelerator architectures requires capital expenditures of approximately $34.6 billion to $39.5 billion. Among them, the construction cost of NVIDIA's Vera Rubin architecture is the highest, while OpenAI's self-developed ASIC architecture, Jalapeno, is relatively lower. Bernstein has significantly revised its cost estimate for NVIDIA Rubin NVL72 single cabinet from the previous $9.1 million to $7.52 million, a reduction of about 17%, primarily reflecting adjustments in expectations for HBM prices and NAND storage capacity. The report also points out that the primary economic burden of AI data centers is not electricity costs but the massive capital expenditures and the resulting depreciation.