The role of the ocean has been pivotal in modulating rising atmospheric CO<sub>2</sub> levels since the industrial revolution, sequestering nearly half of all fossil-fuel derived CO<sub>2</sub> emissions. Net oceanic uptake of CO<sub>2</sub> has roughly doubled between the 1960s (~1 Pg C yr<sup>−1</sup>) and 2000s (~2 Pg C yr<sup>−1</sup>), with expectations that it will continue to absorb even more CO<sub>2</sub> with rising future atmospheric CO<sub>2</sub> levels. However, recent CO<sub>2</sub> observational analyses along with numerous model predictions suggest the rate of oceanic CO<sub>2</sub> uptake is already slowing, largely as a result of a natural decadal-scale outgassing signal. This recent CO<sub>2</sub> outgassing signal represents a significant shift in our understanding of the oceans role in modulating atmospheric CO<sub>2</sub>. Current tracer-based estimates for the ocean storage of anthropogenic CO<sub>2</sub> assume the ocean circulation and biology is in steady state, thereby missing the new and potentially important "non-steady state" CO<sub>2</sub> outgassing signal. By combining data-based techniques that assume the ocean is in a steady state, with techniques that constrain the net oceanic CO<sub>2</sub> uptake signal, we show how to extract the non-steady state CO<sub>2</sub> signal from observations. Over the entire industrial era, the non-steady state CO<sub>2</sub> outgassing signal (~13 ± 10 Pg C) is estimated to represent about 9% of the total net CO<sub>2</sub> inventory change (~142 Pg C). However, between 1989 and 2007, the non-steady state CO<sub>2</sub> outgassing signal (~6.3 Pg C) has likely increased to be ~18% of net oceanic CO<sub>2</sub> storage over that period (~36 Pg C). The present uncertainty of our data-based techniques for oceanic CO<sub>2</sub> uptake limit our capacity to quantify the non-steady state CO<sub>2</sub> signal, however with more data and better certainty estimates across a range of diverse methods, this important and growing CO<sub>2</sub> signal could be better constrained in the future.