I've had a few games playing an early b3 with dividends, particularly vs. the Sicilian... this one turns out to be a fine example

Anticipating the fianchetto and so looking to immediately close off the diagonal

But now the light squares are weak

And here's the position where I came up with a pawn sacrifice (ok it could be theory, I must ask Greenpawn)

Leaving e4 hanging, but Black should ignore for now and continue development with Be7 looking to castle

? By taking, White gets a positional advantage by force

Protecting the Knight only drops material, so Knight falls back

Now e5 is pinned to the King, so d4 falls

And the long term idea is to take advantage of Black's lack of development and backward pawn structure

Black probably realised that Bb7 would have led to Bb5+ losing the Queen!