The card discloses that Llama 4 model training consumed 7.38 million GPU hours and generated an estimated 1,999 tons of CO2 equivalent on a location-based basis, with Meta asserting a market-based emissions figure of 0 tons based on renewable energy matching and open-source release.
This analysis describes what Meta's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
This provision constitutes a voluntary ESG disclosure quantifying the energy and emissions footprint of Llama 4 training. The distinction between location-based and market-based emissions figures reflects recognized greenhouse gas accounting methodologies and may be relevant to institutional investors and ESG compliance teams assessing Meta's climate disclosures.
Interpretive note: The market-based emissions figure of 0 tons depends on Meta's renewable energy matching claims, which are asserted but not independently verified in this document; the methodology for verifying renewable energy certificates is not described.
The document discloses that training energy use and greenhouse gas emissions are borne by Meta under its open-source release model and will not be incurred by organizations that deploy the released weights. The market-based emissions claim of 0 tons depends on Meta's renewable energy matching assertions, which are stated but not independently verified in this document.
Cross-platform context
See how other platforms handle Greenhouse Gas Emissions and Energy Disclosure and similar clauses.
Compare across platforms →Monitoring
Meta has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"Estimated total location-based greenhouse gas emissions were 1,999 tons CO2eq for training. Since 2020, Meta has maintained net zero greenhouse gas emissions in its global operations and matched 100% of its electricity use with clean and renewable energy; therefore, the total market-based greenhouse gas emissions for training were 0 tons CO2eq. Since Meta is openly releasing these models, the training energy use and greenhouse gas emissions will not be incurred by others.— Excerpt from Meta's Llama 4 Model Card
(1) REGULATORY LANDSCAPE: Voluntary greenhouse gas disclosures of this type engage the SEC's climate disclosure rulemaking, which has been subject to legal challenge and implementation uncertainty, as well as the EU Corporate Sustainability Reporting Directive for EU-operating entities. The methodology referenced is the arxiv paper at arxiv.org/pdf/2204.05149. (2) GOVERNANCE EXPOSURE: Low. The disclosure is voluntary and quantitative, following recognized GHG accounting methodology. The primary governance consideration is whether the renewable energy matching claims are supported by verifiable instruments such as renewable energy certificates, which the document does not address. (3) JURISDICTION FLAGS: EU entities subject to the Corporate Sustainability Reporting Directive should assess whether this disclosure format and methodology aligns with applicable reporting standards. California's climate disclosure legislation may impose additional requirements on Meta as a covered entity. (4) CONTRACT AND VENDOR IMPLICATIONS: Organizations with scope 3 emissions reporting obligations may wish to assess whether use of openly released model weights, rather than API-based inference, affects their own emissions accounting. The model card's assertion that training emissions are not incurred by deployers does not address inference-time energy consumption. (5) COMPLIANCE CONSIDERATIONS: ESG and sustainability teams should note that inference-time energy consumption is not addressed in this disclosure, and that organizations deploying Llama 4 at scale should assess the emissions footprint of their own inference infrastructure separately.
Full compliance analysis
Regulatory citations, enforcement risk, and due diligence action items.
Free: track 3 platforms + weekly digest. Monitor: 25 platforms + same-day alerts. No credit card required.
Compliance Governance Intelligence
Need to monitor specific governance provisions?
Compliance includes provision-level monitoring, governance timelines, regulatory mapping, and audit-ready analysis.
Built from archived source documents, structured governance mappings, and historical version tracking.
This provision constitutes a voluntary ESG disclosure quantifying the energy and emissions footprint of Llama 4 training. The distinction between location-based and market-based emissions figures reflects recognized greenhouse gas accounting methodologies and may be relevant to institutional investors and ESG compliance teams assessing Meta's climate disclosures.
The document discloses that training energy use and greenhouse gas emissions are borne by Meta under its open-source release model and will not be incurred by organizations that deploy the released weights. The market-based emissions claim of 0 tons depends on Meta's renewable energy matching assertions, which are stated but not independently verified in this document.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Meta.