Either one can be the right call, but the same dollar can’t go to both. Spare cash that goes to the lenders doesn’t buy back shares, and the other way around. So I look at where the money went, and I use the net debt number, because companies often repay and borrow again in the same year.
Take DXC Technology, which let its cash pile up one year and sent the year’s cash out the next. Its fiscal year ends in March. In fiscal 2025 it spent $14 million on buybacks, repaid no long-term debt, paid $298 million on finance leases and asset financing, and its cash went up $572 million, to $1,796 million, according to its fiscal 2025 annual report.
In fiscal 2026 the year’s cash went out instead, and the pile barely moved, to $1,737 million. The fiscal 2026 annual report shows $1,062 million of long-term debt repaid, which sounds like the lenders came first. But the same statement shows $742 million raised from a new bond. Net, the debt went down about $320 million. Another $188 million went to finance leases and asset financing, and $249 million went to buybacks. DXC’s own free cash flow that year, cash from operations minus all its capital spending, was $713 million.
Buybacks in the earlier years shrank the share count a lot. The average diluted share count was 255 million in fiscal 2022 and 179 million in fiscal 2026, down 30 percent, as the fiscal 2022 and 2026 reports show. That only happens in the years the cash goes to buybacks. In fiscal 2025 it mostly stayed in the bank.
Where this breaks
The repayment line alone overstates how much went to the lenders, and the share count alone doesn’t tell you what was paid for the shares. You need both, and you need a few years of them before you call either one a habit.