System Problem

Universities increasingly run data analysis, machine learning, rendering, and other computing tasks that can consume substantial electricity. Some flexible workloads do not need to run at a specific hour.

Technical Proposal

University computing centers should test carbon-aware scheduling for non-urgent workloads, shifting selected tasks toward periods when the electricity grid has lower emissions intensity or campus renewable generation is highest.

Eligible Workloads

• Apply scheduling only to flexible jobs that users approve in advance.
• Track energy use, completion delay, and grid emissions estimates.
• Allow researchers to opt out when timing affects scientific or operational needs.

Measurement

The program could reduce emissions from computing without requiring less research activity.

Limitations

Carbon-intensity data can be uncertain, so claimed savings should be reported as estimates rather than exact reductions.

Researchers should be shown estimates of both emissions and scheduling delay. That allows them to decide whether the environmental benefit is meaningful enough to justify waiting for a lower-carbon period.

A water budget should not reward departments for allowing landscapes to deteriorate. The purpose is to identify avoidable use and encourage appropriate plant selection, not simply to minimize consumption at any cost.

The comparison should include convenience. A theoretically efficient sharing system will fail if users must travel too far, wait too long, or complete complicated registration every time they borrow a tool.

Students could publish a short annual report comparing water, energy, and crop output. That would help future classes understand that controlled-environment agriculture involves tradeoffs rather than simple technological superiority.

Emergency contracts signed in advance can reduce confusion after a disaster, when municipalities are under pressure to remove debris quickly and may have little time to negotiate responsible handling.

Graduation traditions should be preserved where they carry meaning. The strongest changes will focus on invisible waste rather than removing symbolic elements simply to make the event appear more sustainable.

Certification results should be understandable to travelers. A short public explanation of what each level requires would be more credible than a logo whose standards are invisible.

Seedling production should be coordinated with real restoration schedules. Growing thousands of trees without confirmed planting sites, water, or long-term care would only move the sustainability problem from one stage to another.

The project should be expanded only if the first phase shows that its environmental benefit can be maintained without creating larger costs or operational problems elsewhere.

---

I. Chen