Let M k , Q k k∈N be independent copies of an R 2 -valued random vector. It is known that if Y n := Q 1 + M 1 Q 2 + . . . + M 1 · . . . · M n−1 Q n converges a.s. to a random variable Y , then the law of Y satisfies the stochastic fixed-point equationIn the present paper we consider the situation when |Y n | diverges to ∞ in probability because |Q 1 | takes large values with high probability, whereas the multiplicative random walk with steps M k 's tends to zero a.s. Under a regular variation assumption we show that log |Y n |, properly scaled and normalized, converge weakly in the Skorokhod space equipped with the J 1 -topology to an extremal process. A similar result also holds for the corresponding Markov chains. Proofs rely upon a deterministic result which establishes the J 1 -convergence of certain sums to a maximal function and subsequent use of the Skorokhod representation theorem.