Hydrogen (δ2H) and oxygen (δ18O) stable isotope analysis is useful when tracing the origin of water in beverages, but traditional analytical techniques are limited to pure or extracted waters. We measured the isotopic composition of extracted beverage water using both isotope ratio infrared spectroscopy (IRIS; specifically, wavelength-scanned cavity ring-down spectroscopy) and isotope ratio mass spectrometry (IRMS). We also analyzed beer, sodas, juices, and milk ‘as is’ using IRIS. For IRIS analysis, four sequential injections of each sample were measured and data were corrected for sample-to-sample memory using injections (a) 1-4, (b) 2-4, and (c) 3-4. The variation between δ2H and δ18O values calculated using the three correction methods was larger for unextracted (i.e., complex) beverages than for waters. The memory correction was smallest when using injections 3-4. Beverage water δ2H and δ18O values generally fit the Global Meteoric Water Line, with the exception of water from fruit juices. The beverage water stable isotope ratios measured using IRIS agreed well with the IRMS data and fit 1:1 lines, with the exception of sodas and juices (δ2H values) and beers (δ18O values). The δ2H and δ18O values of waters extracted from beer, soda, juice, and milk were correlated with complex beverage δ2H and δ18O values (r1/40.998 and 0.997, respectively) and generally fit 1:1 lines. We conclude that it is possible to analyze complex beverages, without water extraction, using IRIS although caution is needed when analyzing beverages containing sugars, which can clog the syringe and increase memory, or alcohol, a known spectral interference.