1: \begin{abstract}

2:   We prove a bordism version of Gromov's linearity conjecture over a large family of acyclic groups.

3:   Since all groups embed into these acyclic groups, it follows that the linear bordism conjecture is true if one allows to enlarge a given group.

4:   Our result holds in both PL and smooth categories, and for both oriented and unoriented cases.

5:   The method of the proof hinges on quantitative algebraic and geometric techniques over infinite complexes with unbounded local (combinatorial) geometry, which seem interesting on their own.

6:   As an application, we prove that there is a universal linear bound for the Cheeger-Gromov $L^2$ $\rho$-invariants of PL $(4k-1)$-manifolds associated with arbitrary regular covers.

7: \end{abstract}

8: