Rapid CO2 hydrate formation was investigated with the objective of producing a negatively buoyant CO2-seawater mixture under high-pressure and low-temperature conditions, simulating direct CO2 injection at intermediate ocean depths of 1.0-1.3 km. A coflow reactor was developed to maximize CO2 hydrate production by injecting water droplets (e.g., approximately 267 microm average diameter) from a capillary tube into liquid CO2. The droplets were injected in the mixing zone of the reactor where CO2 hydrate formed at the surface of the water droplets. The water-encased hydrate particles aggregated in the liquid CO2, producing a paste-like composite containing CO2 hydrate, liquid CO2, and water phases. This composite was extruded into ambient water from the coflow reactor as a coherent cylindrical mass, approximately 6 mm in diameter, which broke into pieces 5-10 cm long. Both modeling and experiments demonstrated that conversion from liquid CO2 to CO2 hydrate increased with water flow rate, ambient pressure, and residence time and decreased with CO2 flow rate. Increased mixing intensity, as expressed by the Reynolds number, enhanced the mass transfer and increased the conversion of liquid CO2 into CO2 hydrate. Using a plume model, we show that hydrate composite particles (for a CO2 loading of 1000 kg/s and 0.25 hydrate conversion) will dissolve and sink through a total depth of 350 m. This suggests significantly better CO2 dispersal and potentially reduced environmental impacts than would be possible by simply discharging positively buoyant liquid CO2 droplets. Further studies are needed to address hydrate conversion efficiency, scale-up criteria, sequestration longevity, and impact on the ocean biota before in-situ production of sinking CO2 hydrate composite can be applied to oceanic CO2 storage and sequestration.