Keeping global mean temperature rise well below 2 °C requires deep emission reductions in all industrial sectors, but several barriers inhibit such transitions. A special type of barrier is carbon lock-in, defined as a process whereby various forms of increasing returns to adoption inhibit innovation and the competitiveness of low-carbon alternatives, resulting in further path dependency. Here, we explore potential carbon lock-in in the Dutch chemical industry via semi-structured interviews with eleven key actors. We find that carbon lock-in may be the result of (i) technological incompatibility between deep emission reduction options over time, (ii) system integration in chemical clusters, (iii) increasing sunk costs as firms continue to invest in incremental improvements in incumbent installations, (iv) governmental policy inconsistency between targets for energy efficiency and deep emission reductions, and (v) existing safety routines and standards. We also identify barriers that do not have the self-reinforcing character of lock-in, but do inhibit deep emission reductions. Examples include high operating costs of low-carbon options and low risk acceptance by capital providers and shareholders. Rooted in the Dutch policy setting, we discuss policy responses for avoiding carbon lock-in and overcoming barriers based on the interviews, such as transition plans for individual industries and infrastructure subsidies.