RR_NDB makes a thought-provoking point. The life-cycle and failure rate of Li batteries is very much determined by environment and pattern of usage. For instance they tend to degrade much faster if stored fully charged and/or at high temperature.
When the technology is relatively new, the application/deployment environment novel and the associated battery charge/discharge cycling pattern possibly rather unusual, it leaves open the possibility of some unrecognised factor in physics/chemistry, as there was when metal fatigue started downing early passenger jets.