It depends what you’re trying to achieve. In my experience there are very few scenarios where you actually need lots of GPIOs in the same physical location. In most situations the things you want to control or monitor are spread over a reasonably large physical distance, which makes wiring back the controls/sensors to a single point where your MCU is located impractical.
It’s often more sensible to have multiple devices in different physical locations sending data to a central location - either the Blynk server, as “master” device or something else such as a Node-Red server acting as a gateway to Blynk.
But, if you really do need more GPIOs than you can “shake a stick at” then you could go down this route…
Pete.