Larry Fink Highlights AI's Energy Challenge; Google Aims for Efficiency
BlackRock CEO Larry Fink has emphasized that electricity is becoming a critical constraint in advancing artificial intelligence, suggesting countries like China are preparing by expanding nuclear and solar power. In contrast, Alphabet Inc.'s subsidiary Google is reportedly developing an advanced AI chip called Frozen v2 to improve energy efficiency. Frozen v2 aims to embed parts of the Gemini AI model directly into hardware rather than relying solely on software execution. This approach could reduce computing resources needed for AI tasks without compromising performance. If successful, it may offer a solution to growing power demands in data centers by increasing computational efficiency per watt used. As Alphabet prepares for its upcoming earnings report, attention is shifting toward how the company balances investments in AI scale with efforts to enhance energy efficiency.
Topics: AI capital expenditures electricity demand semiconductors