Why AI Power Demands Could Spark an Energy Revolution

Joe Helfrich, managing partner at NosTerra Venture Capital, views the current energy crisis not as a roadblock, but as a unique opportunity to modernize the physical layer of the internet. As AI workloads shift from centralized training to real-time inference, the need for compute power near end-users is forcing a reckoning with existing generation, storage, and cooling systems. Helfrich suggests that data centers function essentially as factories, converting raw energy into compute, and the current limitations of the grid are simply highlighting where that conversion is least efficient.

To bridge the gap between capacity and demand, investors are increasingly looking for physical-layer technologies that move beyond incremental efficiency gains. One example involves Boston-based firms utilizing carbon-capture systems within modular edge data centers, a process that simultaneously improves cooling and produces marketable byproducts. By turning energy-intensive facilities into hubs for resilient and clean power, the industry hopes to solve the compute bottleneck while providing broader benefits to the national grid. For Helfrich, the urgency of the AI boom provides the necessary capital and customer base to accelerate energy technologies that have historically suffered from slow adoption rates.