Summary Two main types of reservoirs are considered for geological storage of carbon dioxide (CO2): deep saline formations and depleted oil and gas (O&G) reservoirs. The former offers very large potential capacity and a more-even distribution at the expense of high uncertainty because of the very poor characterization of their properties to date, including their sealing capacity. The latter offers smaller overall capacity, but with a reduced risk because of better reservoir knowledge. Gas reservoirs have also provided a proven seal to pressurized gas. However, reusing depleted O&G reservoirs presents challenges that must be considered in the evaluation of performance factors and the associated risks. Depletion can cause pore collapse in the reservoir, with an associated loss of capacity and injectivity and can weaken caprock and bounding faults or even well completions, leading to possible containment losses because of mechanical failure. O&G reservoirs are also intersected by many wells, and it is likely that stricter regulatory requirements on well integrity and the quality of zonal isolation will force operators to recomplete or work over wells that will be exposed to CO2, with an obvious impact on cost. Low reservoir pressure may also mean that injection of CO2 in a dense phase would result in reservoir fracturing and very strong thermal effects that may lead to injectivity problems. In the reservoir, chemical and physical differences in behavior between CO2 and methane may adversely affect geological containment and injectivity. Analyzing the benefits and challenges with respect to all performance factors (capacity, injectivity, containment) shows that depleted O&G reservoirs and deep saline reservoirs both offer potentially attractive targets for geological storage of CO2, mostly for complementary reasons. Uncertainty on capacity and injectivity is clearly lower for depleted reservoirs, giving them a potential net economic advantage, whereas uncertainty on well containment favors saline formations, which are intersected by fewer wells. Injectivity in depleted reservoirs may be much more difficult to ensure than for saline formations or O&G reservoirs where pressure has been maintained. Saline formations have a lower, mostly unproven, safety margin between injection and fracturing pressure, resulting in a potential advantage for depleted reservoirs where repressurization will lead to a final pressure lower than or equal to the original value. Each reservoir type has a different risk profile, different advantages, and a rightful place in a portfolio of injection sites.