Estimated reading time: 4 minutes
Key takeaways:
- 63% of developers say AI contributes to climate change and 70% want employers to do more about it.
- They blame inefficient models and unwanted software, not their own habits.
- 80% want tools to write more energy-efficient code, starting with baselining AI energy use.
Developers care deeply about sustainability and the environmental impact of AI and software systems. They also think employers should be doing more to respond.
A clear majority of respondents (63%) believe AI systems contribute to climate change either a great deal or a moderate amount, according to a joint report from GitHub and the Yale Program on Climate Change Communication (YPCCC). As many as 70% say tech companies and employers should be doing more to address environmental impact.
“Developers want help putting their concern about climate change into practice, primarily through making software more efficient,” says Paull Young, GitHub’s head of sustainability.
The report surveyed 1,039 GitHub monthly active users in the US about climate change, AI’s environmental impact, and sustainable software engineering practices.
It arrives as software engineering teams increasingly rely on compute-intensive AI systems, while new data centers are being constructed to handle growing demand.
Your inbox, upgraded.
Receive weekly engineering insights to level up your leadership approach.
Developers care deeply about future sustainability
A hefty 86% of respondents say global warming is happening, with most linking it at least partly to human activities. Interestingly, that’s higher than among US adults in Yale’s national comparison survey.
More than three quarters (79%) are at least somewhat worried about global warming, while 82% believe it will harm future generations at least a moderate amount.
Data centers consumed about 1.5% of global electricity in 2024, according to IEA, which identifies AI as the most important driver of future growth.
MIT Sloan researchers estimate that the direct emissions associated with AI could contribute roughly 0.1°C of additional global warming by 2100. In one MIT scenario where AI eventually boosts gross world product by 25%, expected warming rises from 3.3°C to 3.6°C above pre-industrial levels.
Pascal Joly, founder of IT Climate Ed, a consultancy that analyzes the impacts of IT on climate change, adds that AI’s environmental footprint extends beyond data center emissions to air pollution from power generation and backup generators, noise from cooling systems, and a high level of water consumption.
“If you look at it from a global perspective, these operational emissions may look small but they are growing rapidly,” he says.
Against this backdrop, 71% of respondents to the survey expressed concern about the environmental impact of AI systems.
Developers: kill unnecessary resource drains
Three in four respondents say it is either “very” or “somewhat” important that their employer actively works to reduce its environmental impact. So, where do the biggest inefficiencies lie? The report’s anonymous qualitative responses point to several potential sources of waste.
One respondent highlights model inefficiency: “Give me more efficient AI models. I don’t care about frontier models. I care about local AI. I care about efficient AI.”
More like this
Inefficient processes and visibility gaps might extend well beyond AI, too. As another developer wrote: “There is a ton of wasted energy in the development and deployment of inefficient and unwanted software.”
The energy and environmental impacts are completely obscured from the developers as they are abstracted away, they add. “Highlighting the energy and environmental impacts to the development community needs more attention.”
Beyond the concerns raised in the report, inefficient AI-generated code could also contribute to unnecessary resource consumption. Other reports cite more duplicate code, rising technical debt, and declining code maintainability in the AI age.
Ways software teams can improve efficiency
Most respondents see their individual coding practices as having a relatively small environmental impact – 63% say its effect is “small.” Instead, they place more responsibility on governments and regulators, cloud providers, tech companies, and employers to act.
Nonetheless, 80% are interested in tools that help them write more energy-efficient code.
Selected qualitative responses suggest a number of ways GitHub could help developers reduce their software’s environmental impact:
- Use data centers powered by renewable energy.
- Provide tools to measure software’s environmental impact.
- Include energy-efficiency scores for large open-source projects.
- Add metrics for the computational toll of GitHub Actions.
- Delay unnecessary compute-heavy CI/CD steps.
- Promote more efficient, open-source local AI models.
- Help developers write more efficient code.
One respondent also emphasized training for the future: “Help train the current and future generations of developers in how to architect and design for sustainability.”
Young recommends engineering leadership start with something the team already wants to fix, whether that’s a slow build, an expensive workload, or code doing unnecessary work.
“Give developers time to tackle it, and bring your sustainability team into the conversation,” he says. “Cost, performance, and sustainability teams often have more shared interests than they realize.”
Joly says education is an important first step. “The first part is to get educated about the problem,” he says. “Software leaders need to clearly communicate to their team the magnitude of the problem, and why we should care at all.”
He recommends baselining your AI tools’ energy consumption and estimated emissions, tracking them over time, and sharing those metrics alongside token-cost dashboards.
Joly also advises considering smaller open-weight models, which can have a lower carbon footprint than frontier models. Matching model size to the task at hand can improve efficiency, too.
Berlin • November 9 & 10, 2026
Close the gap between what leadership expects and what’s actually possible at LeadDev Berlin.
Doing our bit for sustainable AI
The survey isn’t representative of programmers at large, but as the world’s largest Git hosting platform, used by over 200 million developers, GitHub provides a useful window into developer attitudes.
With most respondents saying it matters that their employer actively reduces its environmental impact, sustainability could also have implications for morale and talent retention. Another motivator is the rising cost of AI platforms.
While many of the recommendations above are aimed at GitHub, most could apply to software organizations more broadly as they look to build more efficient coding and AI practices.
