C
CARTIEAI
PricingTry free
Back to home
Proprietary engine

Scope-3 Carbon

Watershed / Persefoni report scope-3 at the ENTERPRISE level. We attribute it per workload.

Hardware embodied carbon (instance + GPU + network + storage) + SaaS-vendor service-category emissions (LLM, monitoring, vector DB, feature flags) → per-workload Scope-3 footprint.

Scope-2 (grid electricity)
2,240 kg
covered by Sub-system P
Scope-3 (value chain)
128.7 kg
= 5.4% of total
Total workload
2,368.7 kg
ai-recommendation-engine
Scope-3 breakdown
hardware-embodied-instance
NVIDIA A100 GPU embodied = 580kg CO2eq, amortised over 5yr × usage 730.0hr/mo
9.7 kg
hardware-embodied-network
Networking gear embodied amortisation @ 18000.0GB egress
0.2 kg
saas-vendor:llm-inference
$4200.0/mo × 0.024 kg-CO2/$ (CEDA 2024 service-category factor)
100.8 kg
saas-vendor:vector-db
$680.0/mo × 0.017 kg-CO2/$ (CEDA 2024 service-category factor)
11.6 kg
saas-vendor:monitoring
$320.0/mo × 0.018 kg-CO2/$ (CEDA 2024 service-category factor)
5.8 kg
saas-vendor:feature-flag
$140.0/mo × 0.005 kg-CO2/$ (CEDA 2024 service-category factor)
0.7 kg

We value your privacy. Cookies help us improve your experience. Learn more

Install CARTIE AI

Add to your home screen for quick access and offline support