You can get the first eigenvector through power iteration, just take some random vector then repeatedly normalize it and multiply by the adjacency matrix. So no

You can get the first eigenvector through power iteration, just take some random vector then repeatedly normalize it and multiply by the adjacency matrix. So no giant matrix inversion needed.