AWS CFO Says AI Infrastructure Boom Is Sustainable, AI Server Payback Period Under Three Years

AWS Chief Financial Officer John Felton publicly stated that the current expansion of spending on AI infrastructure and services is sustainable, with his core argument being that AI is driving meaningful innovation rather than simply burning money. This statement directly addressed market concerns about an "AI bubble."

Felton specifically noted that AWS's AI capital expenditures are generating actual returns, with AI server payback periods under three years. This figure is significant for assessing the economic viability of AI infrastructure — it means that large-scale compute investments are not a bottomless pit, but can recover costs within a reasonable timeframe.

The current market backdrop is as follows: as the global AI race intensifies, tech giants' capital expenditure has surged dramatically, debt levels have risen, valuations remain elevated, and circular financing deals have become frequent. Investors and analysts are broadly concerned about whether companies' revenue growth can keep pace with such massive infrastructure investments, and the "AI bubble" narrative continues to gain traction.

Felton positioned AI as a "productivity-enhancing tool" rather than merely a "cost-cutting tool," a distinction in perspective worth noting. If AI is viewed solely as a cost-cutting measure, its value is easily quantified and compressed; whereas viewing it as a productivity tool implies that AI can create new revenue streams and value increments, thereby supporting higher return-on-investment expectations.

For the industry, the optimistic signal from the CFO of AWS, the world's largest cloud service provider, helps stabilize market sentiment. However, the sustainability of AI infrastructure ultimately depends on whether AI applications can truly be deployed and generate commercial value, rather than being driven solely by capital expenditure.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.