CarbonRunner

Product Introduction

1. CarbonRunner is a cloud compute optimization platform that automatically shifts workloads like AI training and CI/CD pipelines to data centers in regions with the lowest carbon intensity across AWS, Azure, and Google Cloud. It uses real-time grid emissions data to dynamically route compute tasks while maintaining cost efficiency. The platform integrates directly with GitHub Actions through a one-line code modification, enabling developers to reduce infrastructure-related emissions without operational overhead.

2. The core value of CarbonRunner lies in its ability to reduce compute-related carbon emissions by up to 90% while cutting costs by 25% compared to traditional cloud compute strategies. By prioritizing regions with cleaner energy grids and optimizing multi-cloud resource allocation, it enables organizations to meet sustainability targets without compromising performance or requiring infrastructure overhauls.

Main Features

1. CarbonRunner automatically routes workloads to cloud regions with the lowest real-time grid carbon intensity, using live energy mix data from 47 global regions to achieve an average operational intensity of 25gCO2eq/kWh. This system continuously monitors emissions factors and excludes regions exceeding 100gCO2eq/kWh thresholds, ensuring compliance with environmental targets.

2. The platform supports multi-cloud workload distribution across AWS, Azure, and Google Cloud, providing automatic failover and resilience while optimizing for both emissions and cost. This cross-cloud architecture enables dynamic resource scaling based on environmental and economic factors, with no vendor lock-in for users.

3. GitHub Actions integration requires only a one-line modification to existing workflows, replacing standard runners with CarbonRunner's optimized instances (e.g., changing runs-on: ubuntu-latest to runs-on: carbonrunner-4vcpu-ubuntu-latest). This seamless implementation maintains existing pipeline configurations while enabling immediate emissions reductions.

Problems Solved

1. CarbonRunner addresses the environmental impact of cloud computing, where traditional GitHub Actions runners operate at an average of 332gCO2eq/kWh compared to the global grid average of 437gCO2eq/kWh. This results in significant unnecessary emissions for organizations running frequent CI/CD pipelines or compute-intensive workloads.

2. The platform specifically targets development teams and DevOps engineers managing cloud-native applications, AI/ML workloads, and continuous integration pipelines who need to meet corporate ESG goals. It also serves sustainability officers in tech organizations requiring verifiable emissions reductions across IT operations.

3. Typical use cases include optimizing AI model training jobs, video rendering pipelines, gaming infrastructure, and automated testing workflows. One enterprise customer reduced emissions by 89% across 12,000+ monthly CI jobs while maintaining execution SLAs and reducing cloud spend by 18%.

Unique Advantages

1. Unlike static carbon offset programs or single-cloud optimization tools, CarbonRunner combines real-time grid data with multi-cloud orchestration to actively prevent emissions rather than compensate for them. This approach achieves 13x lower emissions intensity than standard GitHub runners compared to typical 2-3x improvements from basic region selection tools.

2. The platform's proprietary algorithm balances three variables simultaneously: current grid carbon intensity, cloud provider pricing fluctuations, and workload performance requirements. This multi-objective optimization considers 15+ data points per region, including renewable energy availability and fossil fuel dependency.

3. CarbonRunner holds a technical advantage through its granular emissions tracking system, which provides per-job carbon accounting validated against the Greenhouse Gas Protocol. Competitors typically offer only aggregate estimates, while CarbonRunner's methodology enables precise reporting for Scope 2 emissions compliance.

Frequently Asked Questions (FAQ)

1. How does CarbonRunner achieve 90% emissions reduction compared to standard GitHub Actions? CarbonRunner routes jobs to regions where the local energy grid averages 25gCO2eq/kWh versus GitHub's 332gCO2eq/kWh average, prioritizing areas with high renewable penetration and avoiding fossil-heavy grids. Actual savings vary between 70-95% depending on workload timing and geographic constraints.

2. What modifications are required to existing GitHub Actions workflows? Users only need to replace the runs-on specification in their YAML files with CarbonRunner's machine types (e.g., carbonrunner-4vcpu-ubuntu-latest). No changes to build steps, environment variables, or artifact handling are required.

3. How are carbon savings calculated and verified? The platform uses real-time regional grid intensity data from authoritative sources like electricityMap.org, combined with cloud provider energy disclosure reports. All calculations follow ISO 14064-2 standards and are auditable through the dashboard.

4. Does multi-cloud routing impact job execution times? CarbonRunner maintains performance parity by excluding high-latency regions from consideration and prioritizing compute instances with equivalent specifications. Most users report identical or faster build times due to optimized resource allocation.

5. What happens during grid outages or regional service disruptions? The system automatically fails over to the next-lowest emission available region while maintaining SLA compliance. All major cloud providers maintain redundant capacity across geographies to ensure service continuity.