According to @paoloardoino, 95% of enterprise generative AI pilots are failing, pointing to weak ROI despite heavy energy usage. Source: @paoloardoino on X, Sep 6, 2025. He adds that VC-funded startups are attempting to brute-force AI development and that the sector remains early-stage. Source: @paoloardoino on X, Sep 6, 2025. MIT-affiliated research has similarly reported that only about 10% of organizations achieve significant financial benefits from AI initiatives, underscoring widespread challenges in scaling pilots. Source: MIT Sloan Management Review and Boston Consulting Group, 2023 State of AI report.

According to @DeepLearningAI, Mistral published an 18-month life-cycle analysis of Mistral Large 2 that measures greenhouse-gas emissions, energy use, and consumption of water and other materials across data-center construction, hardware manufacturing, training, and inference (source: @DeepLearningAI post on Sep 1, 2025). According to @DeepLearningAI, the analysis covers total training impacts and includes inference-stage measurement, but the post does not provide numeric results, requiring traders to review the full report for metrics (source: @DeepLearningAI). According to @DeepLearningAI, the scope explicitly extends beyond model operations to include upstream infrastructure like data-center buildout and hardware production, clarifying where environmental costs accrue in the AI stack (source: @DeepLearningAI). According to @DeepLearningAI, the post does not mention cryptocurrencies or related tokens, and no crypto market impacts are specified (source: @DeepLearningAI).

According to @GoogleDeepMind, the team has published a comprehensive methodology to measure the energy use, emissions, and water impact of Gemini prompts to clarify AI’s environmental footprint for users and stakeholders; source: @GoogleDeepMind on X, Aug 22, 2025. For traders, standardized AI workload metrics arrive as electricity demand from data centres, AI and crypto is projected to roughly double by 2026, heightening sensitivity to power and water costs across AI and blockchain infrastructure; sources: @GoogleDeepMind on X, Aug 22, 2025; International Energy Agency, Electricity 2024 report (Jan 2024).