OpenAI will devour as much power as NYC and San Diego combined
13 hours ago
- #Environmental Impact
- #AI Energy Demand
- #Nuclear Power
- Sam Altman and partners plan AI data centers consuming up to 10 gigawatts, with additional projects totaling 17 gigawatts.
- AI's power demand could reach 10-12% of global electricity by 2030, comparable to powering entire cities or countries like Switzerland and Portugal.
- Nuclear energy is favored by Altman for AI's steady power needs, but experts doubt its scalability by 2030 due to slow deployment.
- Short-term solutions for AI's energy demands include renewables, natural gas, and retrofitting older plants, as nuclear won't arrive fast enough.
- Environmental concerns include increased carbon emissions, water usage for cooling, and electronic waste from rapidly outdated hardware.
- Experts warn that AI's growth may make it impossible for companies to meet previous clean energy commitments.
- Investment in AI infrastructure is massive, with OpenAI sites valued at $50 billion each and Nvidia pledging up to $100 billion.