Our own Emilio Vicens shared some thoughts with Pipeline Magazine about the AI roadmap and why onsite power is the future of AI innovation. Below is a brief excerpt from the article.

The AI boom is moving at lightning speed. The U.S. power grid is not. Data center developers are running into multi-year interconnection queues that threaten to stall AI capacity expansion just as demand is inflecting. On-site firm power offers a practical path forward. Purpose-built generation can bypass the interconnection queue, deliver megawatts in months rather than years, and transition into long-term resiliency infrastructure once grid service arrives, all without adding upward pressure on consumer electricity rates.

The AI-driven power gap

U.S. data center power demand is expected to nearly triple by 2030, rising from 62 GW to more than 134 GW (S&P Global), against a backdrop of essentially flat total U.S. electricity consumption over the past two decades. Utilities were never sized to absorb this kind of concentrated, regional load growth, and it has resulted in multi-year interconnection queues, extended substation upgrade timelines and delayed project starts — even on fully entitled sites.

The 2025 Uptime Institute Global Data Center Survey found that 63 percent of operators now cite power availability as a top concern, ranking it alongside cost and capacity forecasting as the industry’s most pressing challenge. In the speed-to-compute market, certainty of power is the new competitive advantage.

We are excited to share our thoughts on this critical topic so that we can help educate our colleagues and the broader industry. Read the full article here: Pipeline Magazine