SBTi ICT Sector Compliance¶

Guide for setting science-based emission reduction targets under SBTi ICT guidance.

What is SBTi?¶

The Science Based Targets initiative helps companies set emission reduction targets consistent with climate science. Targets are validated against pathways needed to limit warming to 1.5°C or 2°C.

ICT Sector Guidance¶

SBTi provides sector-specific guidance for ICT (Information and Communication Technology), including:

• Software and services companies
• Cloud computing providers
• Telecommunications

AI inference emissions fall under software/services using cloud infrastructure.

Target Types¶

Near-Term Targets (5-10 years)¶

Required for SBTi validation:

openclaw green targets:add \
  --name "2030 Near-Term" \
  --base-year 2025 \
  --target-year 2030 \
  --reduction 42 \
  --pathway 1.5C

For 1.5°C alignment: ≥42% reduction by 2030 (from 2020 or later base year)

Long-Term Targets (by 2050)¶

Net-zero commitment:

openclaw green targets:add \
  --name "Net Zero 2050" \
  --base-year 2025 \
  --target-year 2050 \
  --reduction 90 \
  --pathway 1.5C

For net-zero: ≥90% reduction by 2050

Pathways¶

SBTi defines temperature-aligned pathways:

Pathway	Near-Term Reduction	Ambition
1.5°C	≥42% by 2030	Highest
Well-below 2°C	≥25% by 2030	High
2°C	≥15% by 2030	Minimum

Recommendation: Use 1.5°C pathway for leadership position.

Setting Targets¶

1. Establish Base Year¶

Choose a recent year with reliable data:

# Check current emissions
openclaw green status

# Emissions become base year reference

2. Create Target¶

openclaw green targets:add \
  --name "SBTi 1.5C Aligned" \
  --base-year 2025 \
  --target-year 2030 \
  --reduction 42 \
  --pathway 1.5C

3. Track Progress¶

openclaw green targets

Output:

Emission Reduction Targets

Target                      Reduction  Progress  Status
SBTi 1.5C Aligned (2025→2030)    42%      15.2%  ✓ On track

4. Report Annually¶

Export target progress:

openclaw green export --format tcfd --period 2026 --baseline 2025

On-Track Calculation¶

Progress is calculated linearly:

Expected reduction at year Y:
  = (Y - base_year) / (target_year - base_year) × target_reduction

On track if:
  actual_reduction ≥ expected_reduction

Example: - Base year: 2025 (100 kg) - Target: 42% reduction by 2030 - Year 2027 (2 years in): - Expected: 2/5 × 42% = 16.8% reduction - Actual: 20% reduction → On track

Reduction Strategies¶

1. Model Selection¶

Choose efficient models:

# Compare model efficiency
openclaw green status
# Look at "Top Models" - lower avg/request = more efficient

Smaller models (Haiku, GPT-4o-mini) are 5-10x more efficient than large models.

2. Caching¶

Cache reads are ~10% of input carbon:

• Enable prompt caching
• Reuse system prompts
• Cache common responses

3. Batching¶

Reduce per-request overhead:

• Batch similar requests
• Use longer contexts vs. multiple calls
• Optimize prompt length

4. Provider Selection¶

Compare provider efficiency:

openclaw green status
# Check provider breakdown

Some providers use more renewable energy.

5. Timing (Future)¶

When real-time grid data available:

• Schedule non-urgent requests for low-carbon times
• Use carbon-aware scheduling

SBTi Submission¶

To submit targets for validation:

1. Commitment Letter¶

Sign the SBTi commitment letter.

2. Target Documentation¶

Provide: - Base year emissions inventory - Target boundary (Scope 3 Cat 1) - Target reduction percentage - Target year - Pathway alignment

3. Supporting Data¶

From OpenClaw:

# Base year inventory
openclaw green export --format ghg-protocol --period 2025

# Progress data
openclaw green targets

4. Validation¶

SBTi reviews and validates targets (~$9,500 fee for SMEs).

Example Target Statement¶

## Science-Based Target

[Organization] commits to reduce Scope 3 Category 1 emissions
from AI inference services 42% by 2030 from a 2025 base year,
aligned with the 1.5°C pathway.

### Base Year (2025)
- Emissions: 12.45 kg CO₂eq
- Scope: AI inference API calls
- Boundary: All operations

### Target (2030)
- Emissions: 7.22 kg CO₂eq (max)
- Reduction: 42%
- Pathway: 1.5°C

### Progress (2026)
- Current: 10.50 kg CO₂eq
- Reduction achieved: 15.7%
- Expected by pathway: 8.4%
- Status: On track

### Reduction Strategies
1. Prioritize efficient models (Haiku, GPT-4o-mini)
2. Implement response caching
3. Optimize prompt engineering
4. Evaluate provider renewable energy use

Monitoring¶

Dashboard¶

View target progress in Gateway UI: - Navigate to Green tab - Scroll to "Target Progress" section - See progress bars and status

CLI¶

# Quick status
openclaw green targets

# Detailed export
openclaw green export --format tcfd

Alerts (Future)¶

Configure alerts when off-track:

{
  "green": {
    "targetAlerts": true
  }
}

Related Standards¶

• GHG Protocol — Inventory methodology
• CDP — Recognizes SBTi targets
• TCFD — Target disclosure framework
• ISO 14064 — Verification standard