Contextual benchmarks

What does Northeastern's
AI footprint actually mean?

Putting the numbers in context. The semester total of 51,763 kg CO₂ is hard to hold in your head — so here are five things it's equivalent to.

Semester total as of today: 51,763 kg CO₂(222.3 kg/day)

Five Ways to Feel the Number

All equivalencies use published emissions factors. Bars show the count as a fraction of a reference scale.

Transatlantic round trips

Equivalent to 0 Boston → London round trips

Scale indicator70% of reference scale

Reference scale: 50 transatlantic round trips

So what? Each transatlantic flight emits roughly 1,500 kg CO₂ per passenger. Northeastern's semester AI footprint equals that many flights.

Source: ICAO Carbon Calculator

Boston → New York City drives

Equivalent to 0 drives from Boston to New York

Scale indicator37% of reference scale

Reference scale: 2,000 boston → new york city drives

So what? A round-trip Boston–NYC drive in an average car emits ~70 kg CO₂. This many trips would stretch end-to-end across Massachusetts.

Source: EPA Fuel Economy Guide

Hours of Netflix streaming

Equivalent to 0.0M hours of Netflix streaming

Scale indicator72% of reference scale

Reference scale: 2,000,000 hours of netflix streaming

So what? SD video streaming uses ~0.036 kg CO₂ per hour. This semester's AI usage equals the streaming footprint of the entire NU community watching nonstop for weeks.

Source: Carbon Trust Streaming Report (2023)

Tree-years to offset

Would require 0 trees growing for one full year to absorb

Scale indicator49% of reference scale

Reference scale: 5,000 tree-years to offset

So what? A mature tree absorbs roughly 21 kg of CO₂ per year. Fully offsetting one semester of NU's AI footprint would require a forest of this many trees.

Source: US Forest Service Carbon Sequestration

Smartphones fully charged

Equivalent to charging 0.0M smartphones to 100%

Scale indicator100% of reference scale

Reference scale: 5,000,000 smartphones fully charged

So what? Charging a modern smartphone once emits ~0.0082 kg CO₂. Stacked end-to-end, this many phones would circle the Earth several times over.

Source: EPA eGRID + IEA device efficiency data

Cumulative CO₂ Across the Academic Year

Projected growth from September to May at current rates — and what a 30% efficiency improvement (lighter-weight models, fewer unnecessary queries) would look like.

Current trajectory
With 30% efficiency improvement
End-of-year baseline
60,698 kg
Sept → May at current rates
End-of-year optimized
42,488 kg
With 30% efficiency gain
Potential annual savings
18,209 kg
30% reduction in AI energy intensity
A note on uncertainty

These are estimates. The real numbers could be higher or lower — and that uncertainty is itself the research opportunity.

Northeastern has the data infrastructure, the research talent, and the institutional motivation to measure this precisely. This dashboard is a prototype of what that could look like — a live carbon accounting system for AI usage, validated against real API logs, updated in near-real-time, and integrated with NU's existing sustainability reporting. The methodology gaps are documented. The data partnerships are achievable. The first mover advantage is real.

See the Research AgendaRead the Methodology
NUCarbon

Built for Davis Bookhart, Senior Advisor for Sustainability, Northeastern University — May 6, 2026

Open Source

Prototype · Not official NU data