
The Race for AI Infrastructure Faces a Big Challenge: Pressure from Expensive Loan Costs
The global race to develop artificial intelligence (AI) is triggering an unprecedented wave of infrastructure investment, ranging from giant data centers to server systems and specialized computing chips. However, according to VTV24's reportage, as more and more businesses have to rely on loans to finance this race, escalating interest rates are making AI infrastructure investment costs more expensive than ever.
To meet the computing needs for AI models, leading technology corporations such as Microsoft, Google, Amazon and Meta have all skyrocketed capital spending this year and are expected to continue expanding. According to estimates from JPMorgan Chase bank, more than $4,000 billion in AI-related loans may be mobilized from now to 2030 as businesses massively expand capacity. Notably, the yield on 10-year US government bonds has increased to nearly 5.17% - the highest level since 2007. When benchmark interest rates increase, businesses that want to raise capital are forced to offer higher interest rates to attract investors, pushing the financial costs of infrastructure projects to an alarming level.
This pressure is the heaviest on the group of businesses specializing in providing cloud computing infrastructure. Typically CoreWeave, the company said that every time interest rates increase by 1 percentage point, their annual interest costs could increase by about 30 million USD. Meanwhile, SoftBank has just mobilized more than 11 billion USD through high-risk bond issuances, with a 7-year yield of up to 9.75%. Oracle Corporation even had to send a notice to postpone payment obligations for a data center project in New Mexico when costs increased dramatically. Many experts began to warn about the risks of overinvestment, similar to the 2001 telecommunications bubble.
Despite this, market demand for AI products remains very high. Commitments to use long-term computing capacity from big names like OpenAI or Anthropic continue to create a fulcrum for infrastructure units to seek funding. However, when the period of cheap capital officially ends, the current survival problem for businesses is not only how much money they can raise, but the ability to control interest costs and optimize actual efficiency on each server cluster.
Reference video source: A.I race and pressure on infrastructure investment costs | VTV24
Need a reliable technology partner for your business? See DUDI's corporate website design.