According to analysts, the most aggressive cloud computing giants investing in AI infrastructure are facing a tricky problem—cash on hand is rapidly shrinking.

Data speaks volumes. A few years ago, these cloud providers' cash and cash equivalents accounted for 40%-50% of total assets, making their financial position quite solid. But now? That ratio has dropped to around 10%-20%. The seemingly small numerical change actually masks a large-scale consumption of real cash.

The reason is straightforward: infrastructure investment in the AI era has no ceiling. GPU chips, data centers, compute pools... these things burn money at an astonishing rate. Companies continuously pour cash into these projects, and their once-thick cash cushions are becoming increasingly thin.

What to do when this path reaches a dead end? There are only two options: either find ways to make these AI investments generate returns and convert into cash flow quickly; or rely more frequently on external funding methods like issuing bonds or financing to sustain the investment pace. Neither choice is easy.