Rethinking Datacenter Refresh
We optimize when datacenter components are replaced to reduce cost and carbon emissions.
AI For Tomorrow is an open-source research lab focused on the long-term, scalable, and sustainable development of artificial intelligence.
Progress in AI is increasingly shaped by energy, hardware, and the cost of running experiments. We work on the algorithms, computing systems, and infrastructure that make AI more efficient and less resource-intensive at every layer of the stack.
See the latest challenges we're tackling here.
AI For Tomorrow was started in casual dorm-room conversations and extra-curricular hacking at the University of Pennsylvania in 2024 between close friends. Since then, we have formalized our organization, expanding our partnerships and team post-graduation.
Today, we are a 100% volunteer open-source lab of engineers and researchers working on efficient AI systems, sustainable computing, and the shared infrastructure that supports them.
Meet our team here.
We optimize when datacenter components are replaced to reduce cost and carbon emissions.
We benchmark neural-network training algorithms to identify faster, more efficient ways to train models.

We are improving energy accounting for language-model serving by connecting GPU measurements to individual requests.