Why do we need the := symbol in golang?

I do understand the difference between = and := in golang, my question is why do we actually need the :=? Couldn't the compiler figure out on its own that a variable was not yet defined? Is it just about readability?

Because you can shadow outer-scope variables in an inner scope, such as a closure or loop; the compiler cannot infer purely based on whether the variable existed before or not. Having to explicitly define variables also helps to reduce defects which are commonplace in languages like PHP and JavaScript that allow you to just foo = bar without any consideration of whether foo has been defined. That kind of loose compilation paves over common mistakes like typos such that they're no longer caught at compile time, they show up as "weird behavior" at runtime, which is much, much harder to track down and fix.