Switch lost a little ground to the DS this month, but that is all it really has to do. It just needs to basically keep pace or fall just a little behind. The DS was a monster at this point and it's going to last for about 1.5 more years. After that DS sales plummet dramatically (like the worst plummet ever). Switch just needs to keep a steady pace. Even if it pulls a little behind each month, it just needs to be able to go the distance.
Given how long game development times are nowadays, Switch is bound to get seven years minimum before its successor launches. Given that it's now Nintendo's only console, eight years or more are also possible. When the 3DS could have six hardware SKUs, Switch should see at least five over its lifespan.
The DS had six years for itself (with year 6 still managing 20m units) and that will be the decisive difference in this race. Switch doesn't need to reach the same heights or maintain as high of a plateau as the DS because it will have more time.
Switch was at ~49m by the end of 2019, so a timeline like this would be sufficient:
End of 2020: ~75m
End of 2021: ~100m (new model)
End of 2022: ~122m
End of 2023: ~140m (new model)
End of 2024: ~153m (Switch 2 launches in November 2024)
End of 2025: ~160m
End of 2026: ~163m
End of life: ~164m
This example gives Switch a cushion of 10m to beat the DS and it's not especially optimistic. If Nintendo properly supports this console through 2023, then the above sales progression looks fine. 2024 onwards is when the year over year drops get sharper because the big hitters will go to Switch's successor. What the above sales curve assumes is a transitional period that gets handled similar to how Nintendo handled 3DS to Switch where the old system kept getting low profile games for a couple of years.