AI’s power race is shifting leverage from chipmakers like NVIDIA to the grid - CryptoSlate

The AI computing power race is fundamentally reshaping the semiconductor industry's power dynamics, shifting leverage from chipmakers like NVIDIA to grid infrastructure. As AI applications intensify e... Read original →

The AI compute race is pivoting from raw performance to power efficiency—a structural shift redistributing leverage across the tech stack. Technically, NVIDIA must now co-design with liquid cooling, smart grid, and power management ecosystems; even EDA flows will prioritize watts over flops. Regulatory risks are mounting: proposed PUE caps in the EU and U.S. could inflate operational costs and trigger supply chain realignments, especially for fabs in Taiwan, China and Arizona. Competitively, AMD and Intel will weaponize energy-per-inference metrics, while niche players like Groq may gain traction by optimizing 'performance per watt.' Over the next 12–24 months, access to clean, affordable power—not just transistor density—will become a core valuation driver, rewarding firms with PPAs or microgrid capabilities with market premium.