Segai-Modang languages, located primarily in East Kalimantan, Indonesia, and directly descended from Proto-Kayanic (PKay), are of a phonological type far removed from Proto-Malayo-Polynesian and most of its daughter languages. Segai-Modang languages are stress-final and have innovated sesquiand monosyllabic canonical word forms with expanded vowel inventories. They share these characteristics with a few other, individual languages of Borneo (Sa'ban, Merap, certain Bidayuh languages, including Hliboi), and with Chamic languages of mainland Southeast Asia. In Borneo, however, Segai-Modang is the only large subgroup in which every known member has undergone these phonological innovations, and thus provides a unique opportunity for reconstructing an Austronesian proto-language (Proto-Segai-Modang [PSM]) whose daughter languages are entirely sesqui- or monosyllabic and which was not influenced through linguistic contact. The present study provides evidence for a hypothesis that PSM was itself sesquisyllabic, that the penultimate syllable was reduced to schwa, and the features of PKay penultimate vowels were transferred to the onsets of the final syllable. This created distinct regular, palatalized, and labialized consonants in final-syllable onset position at the PSM level. These features were later transferred to the final-syllable vowels resulting in diverse reflexes of PSM vowels in the daughter languages. The reconstruction, therefore, posits that final-syllable onsets were complex but that vowels remained phonemically conservative. The vowels *a, *aː *u, and *i are reconstructed to the final syllable, and *ə to the penultimate sesquisyllable. The reconstruction also posits conditioned allophony for many of the PSM final-syllable vowels, which became distinct only after the breakup of PSM.