Approximately 20 to 30% of cows diagnosed not pregnant 32 d after timed artificial insemination (TAI) lack a corpus luteum (CL), and cows submitted to a resynchronization protocol in the absence of a CL have about 10% fewer pregnancies per AI (P/AI) than cows with a CL. An understanding of luteal dynamics after synchronization of ovulation and TAI may help refine strategies for reinseminating cows failing to conceive. Lactating Holstein cows (n = 141) were synchronized for first TAI using a Double-Ovsynch protocol. Thrice weekly from 4 to 32 d after TAI, blood samples were collected for evaluation of plasma progesterone (P4) concentrations, and CL diameter was measured using transrectal ultrasonography. Pregnancy status was determined using transrectal ultrasonography 32 d after TAI. Nonsynchronized cows (n = 4) were removed from the study. For cows diagnosed pregnant 32 d after TAI (n = 57), P4 increased from 4 to 15 d and then remained constant until 32 d after TAI, whereas CL volume increased from 4 to 11 d and then remained constant until 32 d after TAI. For cows diagnosed not pregnant 32 d after TAI (n = 80), P4 profiles were evaluated using statistical cluster analysis based on the day after TAI that P4 decreased to <1 ng/mL, resulting in 5 clusters: (1) CL regression 15 d after TAI (1.3%), (2) CL regression 18 to 22 d after TAI (55.0%), (3) CL regression 25 to 27 d after TAI (17.5%), (4) CL regression 29 to 32 d after TAI (5.0%), and (5) CL maintained until 32 d after TAI (21.3%). Plasma pregnancy-associated glycoprotein (PAG) levels at 25 and 32 d after TAI differed among clusters and were below the cut-off value of the assay for the classification of cows as not pregnant for cows in clusters 2, 3, and 4, whereas more than half of the cows in cluster 5 had increased plasma PAG levels. We conclude that at least half of the nonpregnant cows that maintained their CL until 32 d after TAI were initially pregnant but underwent early pregnancy loss based on increased plasma PAG levels at 25 and 32 d after TAI.