Currently there is only one paleo‐CO₂ record close enough to the Cretaceous‐Paleogene (K‐Pg) boundary to record a rapid release in atmospheric CO₂, a greenhouse gas. This record is based on the stomatal frequencies of fern fossils at the K‐Pg boundary and Ginkgo fossils before and after the boundary. Unfortunately, due to deficiencies with the method, the CO₂ inferences are only qualitative. Here we look at the same fossils with a proxy based on leaf gas‐exchange principles (i.e. photosynthesis). We first test the proxy with three living species grown at 500 and 1000 ppm CO₂, including the nearest living relative of the K‐Pg fern, and find a comparable accuracy to other quantitative paleo‐CO₂ proxies. The fossils record a modest ~250 ppm increase in CO₂ across the K‐Pg boundary. These estimates are consistent with most temperature records and with carbon cycle modeling of Deccan volcanism and the meteorite impact. This article appeared on the Geophysical Research Letters website at https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2018GL081215