The hardest problems in modern biology are now compute problems. Protein structure prediction, computer vision applied to biomedical imaging, deep learning on light microscopy data: each one consumes serious processing power, and the energy footprint of that work is no longer an afterthought. As the science scales, so does the power bill, and so does the carbon.
That tension is exactly why we are proud to sponsor the UCSF Quantitative Biosciences Institute (QBI) Hackathon. 639Cloud is providing cloud compute credits to the teams building at this year's event.
The QBI Hackathon is a two-day sprint that brings the Bay Area developer community together with scientists from UCSF, UC Berkeley, and UC Santa Cruz to take on cutting-edge biomedical problems. The same algorithms that learned to recognize faces and drive cars are now reshaping how researchers predict protein structures and interpret biomedical imaging. The hackathon is a bet on how far that science can move when developers and researchers build side by side.
That kind of work needs real compute. It does not need to come at the cost of the climate.
The energy demand of AI is not a distant concern. Global data center electricity use reached roughly 415 TWh in 2024, about 1.5% of the world's total, and it has been climbing about 12% a year, more than four times faster than overall electricity demand. The compute behind AI is the engine of that growth: GPU-accelerated servers, the kind that train the models a hackathon team builds, are growing around 30% a year. The International Energy Agency projects data center consumption will roughly double to about 945 TWh by 2030.
Biomedical AI sits right at the center of this curve. Training and inference on imaging and structure-prediction models are among the most power-hungry workloads in computing. As this research moves from the lab to everyday clinical and scientific practice, the question is not whether the compute will scale. It is what will power it.
That is the choice 639Cloud exists to remove. Researchers should not have to weigh scientific ambition against environmental cost. With renewable-powered infrastructure, they don't have to.
639Cloud runs on 100% renewable energy. Every model a hackathon team trains, every dataset it processes, runs on infrastructure powered entirely by renewable energy.
For biomedical AI specifically, this matters more than it does for most workloads. Pairing compute-intensive research with renewable-powered data centers means the science can scale without the energy footprint scaling alongside it. Sustainable cloud computing is not a marketing line for us. It is the architecture.
Supporting the QBI Hackathon puts that principle to work where it counts: in the hands of the people pushing biomedical science forward.
The next Hackathon is set for June 27-28, 2026 at UCSF Mission Bay. At the main event, participants will come together for two days to collaborate and develop their results. At the end of the Hackathon, teams will present their work in front of a panel of judges and the best teams will receive prizes.
The teams at the QBI Hackathon are working on problems that could shape the next generation of biomedical research, from faster protein analysis to better tools for reading medical images. We are glad to give them the compute to do it, on infrastructure that does not ask them to choose between ambition and sustainability. The breakthroughs this work points toward deserve to be built on power that doesn't cost the planet.
