17 - Voice of Industry Experts - Why the Future of Code Must Be Green?

Why the Future of Code Must Be Green 

By Kedar Deo – Green IT Practice Head, Tech Mahindra

Introduction: From Code to Carbon

When I wrote my first lines of code almost 30 years ago, my biggest concern was whether my application would run without crashing. Performance tuning meant saving a few milliseconds off execution time. “Sustainability” wasn’t in the vocabulary of most developers.

Fast forward to today, and the stakes are far higher. The code we write, the platforms we use, and the architectures we design have a direct impact on the planet.

💡 Fact: By 2040, the Information and Communications Technology (ICT) sector could account for 14% of the world’s carbon emissions — nearly 10x higher than in 2007.

The good news? We can change that. The better news? It doesn’t mean sacrificing innovation — in fact, it’s an opportunity to innovate smarter.

🌍 Why Sustainability in IT is Not Optional

The United Nations defines sustainability as:

Meeting the needs of the present without compromising the ability of future generations to meet their own needs.

In technology, this translates to three imperatives:

1. Economic viability – Build profitable, innovative solutions that create jobs and deliver value.
2. Environmental responsibility – Minimize resource usage and carbon emissions at every stage.
3. Social equity – Ensure technology is accessible, fair, and beneficial to all.

Real-world warning signs

• AI Training Footprint: Training a large language model can emit as much CO₂ as five cars over their entire lifetimes.
• Crypto Energy Use: In 2019, Bitcoin’s network consumed more electricity than Switzerland.
• Small Gains, Big Costs: A neural network’s accuracy increased by 1.74% (from 96.17% to 97.91%), but its energy use tripled.

These examples make one thing clear: efficiency is not just about speed — it’s about sustainability.

💻 What is Green Software Development?

At its core, Green Software Development is about avoiding waste — in CPU cycles, storage, and energy — while still delivering the required functionality and user experience.

Think of it like building a car:

• The driver cares about speed, comfort, and fuel economy.
• The engineer should ensure the vehicle is efficient under the hood.

Key Practices

• Optimize performance – Avoid heavy memory usage or CPU overconsumption.
• Right-size infrastructure – Don’t keep servers running at 15% utilization all year.
• Choose the right storage class – Over-provisioning storage tiers wastes both money and energy.
• Minimize dependencies – Every extra library has a carbon cost.
  

🏛 The Three Pillars of Green Coding

1. What is generated?
   Is your code efficient in terms of the energy it consumes to produce the required output?
   Example: A cleaner SQL query that retrieves only the necessary fields instead of SELECT * can drastically reduce computation time.
2. How is it generated?
   Is your development lifecycle efficient? Can you avoid unnecessary builds, redundant tests, or unused features?
   Example: Using a CI/CD pipeline with automated test triggers instead of running all tests after every minor change.
3. Where does it run?
   Does your application run on energy-optimized platforms and configurations?
   Example: Deploying on a cloud region powered largely by renewable energy rather than fossil fuels.

 

🤖 Enter Agentic AI — The Sustainability Multiplier

The future is shifting from traditional AI to Agentic AI — AI that is autonomous, context-aware, and capable of taking independent action.

Benefits

• Efficiency: Automate tasks with minimal resource usage.
• Scalability: Handle fluctuating workloads without over-provisioning.
• Personalization: Deliver exactly what the user needs, avoiding wasteful over-processing.

Real-world example

A logistics company using Agentic AI for route optimization reduced delivery miles by 20%, saving both fuel costs and carbon emissions.
Similarly, smart data center cooling powered by AI can cut electricity usage by 30–40%.

🗑 The Hidden Wastes of IT — And How to Fix Them

In Lean manufacturing, waste is any step that doesn’t add value to the customer. In IT, it’s no different.

Common types of waste:

• Overproduction – Building features nobody uses.
• Waiting – Idle servers or teams waiting on approvals.
• Over-processing – Running expensive computations for negligible gains.
• Under-utilization – Highly skilled staff doing repetitive tasks.
• Inventory – Backlogs of untested or unused code.

Pro tip: The “Lean AI Waste Matrix” approach maps waste types to AI optimization strategies, helping teams target the biggest efficiency wins first.

🔍 Measuring the Carbon Footprint of Your Code

If you can’t measure it, you can’t improve it.

Example: Using CodeCarbon in Python

from codecarbon import EmissionsTracker

tracker = EmissionsTracker()

tracker.start()

# Your code here

tracker.stop()

The tool calculates CO₂ emissions for your workload, so you can track progress over time.

Other tools:

• Website Carbon Calculator – For web applications
• EPA Household Carbon Footprint Calculator – For personal use

🏗 Designing the Modern, Sustainable Application

A sustainable architecture doesn’t happen by accident — it’s intentional.
The 7-layer modern application architecture ensures that from user experience to infrastructure, every layer is optimized for performance and sustainability.

Example:

• Experience Layer – Use lightweight front-end frameworks.
• Business Process Layer – Automate workflows to reduce manual processing.
• Data Layer – Store only what’s needed, archive intelligently.
• Infrastructure Layer – Use cloud providers with strong renewable energy commitments.

🔮 The Road Ahead: Green + AI + Governance

To build a truly sustainable future, we need:

1. Holistic AI strategy – Combining efficiency, ethics, and innovation.
2. Clear governance – Policies that address environmental impact alongside compliance.
3. Cultural change – Developers must see green coding as a core skill, not an optional extra.
4. Early adoption – The sooner you start, the easier the transition.

Conclusion: Plant the Seeds Now

There’s an old saying: “Don’t change the wheels when the cart is moving.” But when it comes to sustainability, we can’t afford to wait until the system fails. Like any harvest, we need to plant the seeds — of green coding, Agentic AI, and waste-free architecture — today.

The future belongs to those who can innovate and protect the planet.

 

About the Author

Kedar Deo is a Green IT Practice Head, Adobe Experience Manager Certified Architect, and industrial academic advisor to multiple engineering institutes. He has designed Green Software Development curricula, mentors sustainability-focused projects, and champions Agentic AI for environmental impact reduction.