The DB’s full report is yet to come, but the cancelled trains don’t make a significant difference in that statistic. What should be said is that a train is counted as delayed when the delay is at least 6 minutes, which is less than some other countries use as a metric. Also, this is purely accounting for long distance trains. If you add local trains, the punctuality jumps significantly to just under 90%.

What they do have is the traveller’s punctuality of long distance customers. Essentially the question is: how many passengers actually arrived to their destination within 15 minutes of their scheduled arrival time. That includes cancelled trains and also missed connections due to delays (e.g. if your train has 10 minutes delay but that causes you to miss a connection and have to take the next train which leaves an hour later, than your arrival delay is an hour). The DB actually likes to focus on this statistic since this is actually what matters, but unfortunately the media prefer to focus on the other statistics. Not that the traveller’s punctuality is great, though - the punctuality there was 67.4% in 2024.