Investigation of gene regulatory programs underlying corneal epithelial cell specification and homeostasis is essential for understanding how the cornea maintains vision. Here, we describe the use of true single-cell resolution spatial transcriptomics (ST), enriched with full-tissue single-cell RNAseq (SC), to improve spatial resolution and enhance cell cluster size up to 65-fold and per-cell transcriptomic depth up to 17-fold. This enabled cell type specification across the full differentiation trajectory from limbal stem cells (LSC) to superficial corneal epithelium and identification of an activated signature ( Atf3, Zfp36, Gsta4 and Dapl1 ) marking differentiation-primed states across multiple cell types, including a major activated intermediate epithelium (AIE) population. Validation using ST data from murine corneas at different postnatal ages and multiple human SC datasets confirms a large AIE population, which spatial localization and transcriptomic profiling suggest is an active intermediate state distinct from quiescent wing cells. Sub-clustering further revealed early ( Sox9, Hes1 ), proliferative ( Mki67, Top2a ) and mature ( Ccdn1, Dapl1 ) transient amplifying cell subpopulations and four LSC subpopulations, including putative active ( Atf3, Socs3, Zfp36 ), quiescent ( Gpha2, Ifitm3, Cd63 ) and Apoe -specific. Direct ST-to-SC comparison revealed enhanced axonal processes and genes ( Sema3f, Sema 4d, Pax6 ) and cell-cell adhesion and cell-matrix markers ( Itgb4, Tns4, Tjp3 ) in ST data, suggesting cell dissociation from tissue in SC masks epithelial innervation, adhesion and barrier functions. Our findings identify and localize key transcriptional programs in situ , prompting a re-evaluation of epithelial states in scRNA-seq data.