If you run the benchmarks longer (i.e. more than a couple hundred milliseconds - the above benchmarks use validation values from the original benchmarksgame, not the ones used for evaluating performance), you will see a different picture:

https://benchmarksgame-team.pages.debian.net/benchmarksgame/...

https://benchmarksgame-team.pages.debian.net/benchmarksgame/... (and this is, by far, the most favorable comparison for Node.js you can make, where it can trade punches but ultimately ends up losing)

Sure, V8 is impressive, but there is no free lunch especially when your optimization target is a dynamically typed, initially interpreted, language.