Intranasal insulin is a putative neuroprotective therapy after cardiac arrest, but safety in humans at doses extrapolated from animal models is unknown. This phase I, open‐label adaptive dose‐escalation study explores the maximum tolerated dose of intranasal insulin in healthy human participants. Placebo or insulin at doses from 0 to 1000 units was given to healthy participants intranasally on repeated weekly visits. Serum glucose, insulin, and C‐peptide levels were measured serially at 0, 15, 30, 60, 120, 180, and 240 min after administration. Twenty‐four participants (12 female, median age 53, IQR 35–61) were enrolled. There was minimal change in average serum glucose after administration of intranasal insulin. Average serum insulin increased slightly in a dose‐dependent manner, reaching maximum concentrations at 15 min. C‐peptide decreased over time from administration in all groups. One participant had severe hypoglycemia (24 mg/dL at 45 min) and a different participant had mild hypoglycemia (51 mg/dL at 30 min), both after receiving 600 U intranasal insulin. Hypoglycemic episodes were associated with increases in serum insulin. Both participants continued in the study without hypoglycemia after additional doses. High‐dose intranasal insulin up to 1000 U was generally well tolerated, with minimal measurable systemic absorption and without significant aggregate changes in mean glucose. Idiosyncratic episodic systemic absorption and hypoglycemia require further study and additional caution in potential clinical application. Further study of its target engagement and efficacy as a neuroprotective therapy after cardiac arrest at these doses is warranted.