Intonation phrase boundaries (IPBs) were hypothesized to be especially difficult to process in the presence of an amplitude modulated noise masker because of a potential rhythmic competition. In an event-related potential study, IPBs were presented in silence, stationary, and amplitude modulated noise. We elicited centro-parietal Closure Positive Shifts (CPS) in 23 young adults with normal hearing at IPBs in all acoustic conditions, albeit with some differences. CPS peak amplitudes were highest in stationary noise, followed by modulated noise, and lowest in silence. Both noise types elicited CPS delays, slightly more so in stationary compared to amplitude modulated noise. These data suggest that amplitude modulation is not tantamount to a rhythmic competitor for prosodic phrasing but rather supports an assumed speech perception benefit due to local release from masking. The duration of CPS time windows was, however, not only longer in noise compared to silence, but also longer for amplitude modulated compared to stationary noise. This is interpreted as support for additional processing load associated with amplitude modulation for the CPS component. Taken together, processing prosodic phrasing of sentences in amplitude modulated noise seems to involve the same issues that have been observed for the perception and processing of segmental information that are related to lexical items presented in noise: a benefit from local release from masking, even for prosodic cues, and a detrimental additional processing load that is associated with either stream segregation or signal reconstruction.