Comcast handles terabytes of telemetry data daily from customer-premises equipment[0]. They could have internal machine learning models for customer retention / churn reduction from the data you're generating, whether bandwidth usage, sites visited, etc. Some model(s) may have inferred that you had a higher likelihood to move and or churn. You can train such models from historical data of people who have moved in the past, and if there's a chance that people who move tend not to care about their subscription anymore, cancel it, or renew it after a long period, there is lost revenue to be minimized. A provider will then engage you, either through email or via a customer support representative if the revenue per customer is high.

This does work and organizations pay close attention to it. If you have a retail bank that makes around $1,000 per customer with 6% churn rate and 10 million customers, this adds up to quite a chunk of change.

It could also be a coincidence due to the current trend of people moving around, and sending emails like this to remove one cognitive process from someone who already has many things to sort out.

[1]: https://www.nctatechnicalpapers.com/Paper/2019/2019-what-can...