Sustainable AI at UC San Diego

Driving Innovation Responsibly

TritonGPT is hosted at the San Diego Supercomputer Center, a facility designed for energy efficiency and lower emissions. By running AI on SDSC infrastructure rather than commercial cloud services, UC San Diego delivers powerful AI tools with a smaller environmental footprint.

Reduced Energy & Emissions

SDSC reduces energy use and emissions through efficient building design, innovative cooling systems, green battery storage, and access to UC San Diego’s low-emission cogeneration plant. This allows AI queries to run on cleaner, more sustainable infrastructure.

• Optimized building orientation, sun shading, and low emissivity glass reduce heat gain and cooling demands.
• CoolFlex rack systems keep hot and cold air separate, improving cooling efficiency and reducing fan energy.
• A solar reflecting roof and high efficiency air handling units further decrease energy use. Sensors throughout the building track energy performance and guide future improvements.

SDSC draws most of its electricity from UC San Diego’s cogeneration plant, which produces far fewer emissions than conventional sources. Traditional lead acid backup batteries have been replaced with safer, environmentally friendly alkaline technology that provides more than double the backup capacity.

SDSC also powers research into cleaner energy technologies, including wave energy converters and advanced battery development, reinforcing UC San Diego’s broader sustainability goals.

What This Means for TritonGPT Users

Hosting AI at SDSC means every TritonGPT query benefits from:

• Lower energy consumption
• Reduced emissions
• Safer, greener backup systems
• Campus controlled infrastructure aligned with UC San Diego sustainability goals
• TritonGPT users get campus tailored AI tools supported by infrastructure that prioritizes environmental responsibility