We study how workers' wages respond to TFP-driven innovations in firms' labor productivity. Using unique data with highly reliable firm-level output prices and quantities in the manufacturing sector in Sweden, we are able to derive measures of physical (as opposed to revenue) TFP to instrument labor productivity in the wage equations. We find that the reaction of wages to sectoral labor productivity is almost three times larger than the response to pure idiosyncratic (firm-level) shocks, a result which crucially hinges on the use of physical TFP as an instrument. These results are all robust to a number of empirical specifications, including models accounting for selection on both the demand and supply side through worker-firm (match) fixed effects. Further results suggest that technological progress at the firm level has negligible effects on the firm-level composition of employees.