The carbon dioxide (CO2) concentration in the atmosphere has increased dramatically over the last century, and much of this increase is a direct consequence of human activities. Being a greenhouse gas, CO2 significantly contributes to global warming, and there is an international agreement to reduce emission of CO2. One method to reduce CO2 release is carbon dioxide capture and storage (CCS). CCS entails to capture waste CO2 from large point sources, and transport it to a permanently deposit, usually in geological formations. The transported CO2 is likely to contain impurities like water, oxygen, nitrogen dioxide, sulfur dioxide and hydrogen sulfide. These impurities are known to cause corrosion, dust and/or solid formations, and may lead to operational difficulties. The present study examined how water as an impurity, both as a free water phase and water dissolved in dense phase CO2, affects the corrosion of carbon steel. There are few experimental studies that have examined corrosion in CO2 transport pipelines, and these few studies have shown different corrosion rates in experiments where the reported amount of dissolved water in the CO2 was the same. These discrepancies may reflect short-comings in methodology in some of these studies, where formation of free water may have contributed to high corrosion rates. The present study was undertaken to resolve some of the controversy related to corrosion in pure CO2/water systems. For this study we designed experimental equipment that allowed complete control of the water concentration in CO2, without the risk of droplet formation. This provides realistic results that mimic the situation in the operating pipelines. Three corrosion experiments were conducted with water dissolved in CO2 at 25 °C and 95 bar of pressure. Even with water concentrations close to the solubility limit, the general corrosion rates were below 1.5 µm/y and no localized attacks were found. A fourth experiment was conducted at and above the solubility limit of water in CO2. Even though we could confirm at the end of the experiment that water had accumulated in the loop, the corrosion rate was still at 1.0 µm/y, with no localized attacks. A final experiment was performed at 35 °C and 95 bar of pressure, in which no corrosion was detected. Corrosion products were found on the surface in all experiments, but at low quantities, most likely less than of practical significance. However, with water levels above the solubility limits, i.e. in the present of free water phase, the corrosion rate increased by 3 to 4 orders of magnitude (2 mm/y to 16 mm/y, pressure depended). For a CO2 transport pipeline these results suggest that the corrosion will not be a problem as long as the water is kept below the solubility limit. However, it is extremely important to avoid liquid water ingress as this would dramatically reduce the lifetime of the pipeline.