I've seen this statement several times, but I haven't seen an explanation for it. What about Lasso Regression makes it struggle against multicollinearity?
When I've seen this stated, it's usually in the context vs. Ridge regression or some feature selection method like subset/forward/backward selection, etc...