The AI Boom Needs Clean Energy
Why we need new infrastructure to power the AI revolution.
September 18, 2025
Salesforce Ventures has long focused on backing companies that enable the future of enterprise. Today, that increasingly means addressing the foundational infrastructure powering AI. Every new breakthrough in generative models, autonomous systems, or intelligent automation is powered by an increasingly intertwined set of physical infrastructure: energy generation, power transmission, energy storage, complex grid infrastructure, and, ultimately, data centers.
With our focus on investing in enterprise innovation, we’re keenly interested in how AI is shaping the future of energy. The insatiable need for compute to power AI models is fueling a boom in data center capacity and, consequently, a rise in energy consumption.
BCG projects U.S. data center electricity demand is set to increase to up to 130 GW by 2030 due to AI. For context, this would represent close to 12% of total U.S. annual demand today, or the equivalent of the electricity used by 100 million U.S. households. Said another way, the U.S. economy is set to consume more electricity for processing data by the end of the decade than for all manufacturing of energy-intensive goods, including aluminum, steel, and cement.
The challenge with AI’s demand for energy is that the current grid is simply not built to support this level of energy demand. Even the best-run hyperscalers are facing multi-year interconnection delays and rising operational costs. Analysts warn that 40% of AI data center projects could hit power bottlenecks by 2027, and three quarters of CIOs may face energy constraints by 2030.
To sustainably meet this surging, short-term energy demand, diversification of the energy mix is necessary. Today, a majority of data center energy is supplied by traditional sources, and those sources are projected to meet a portion of new demand. This highlights the urgent need for accelerated innovation in higher-capacity energy solutions to ensure the long-term viability and growth of AI infrastructure.
The Salesforce Ventures team actively invests across the AI/energy value chain. We look at technologies that optimize data center development and operations, next-generation energy supply companies, and grid solutions that enable AI infrastructure at scale.
Ahead of Climate Week, we wanted to share a few opportunities in the energy space we’re leaning into and are excited to learn more about.
Energy Supply
Energy supply companies offer baseload power and distributed renewables designed for AI-scale loads. Within this category, we’re excited by power source businesses like Base Power, which offers affordable backup battery power, and WeaveGrid, which is working to optimize EV charging to unlock energy on the grid. There are also energy demand response companies like Emerald AI (AI-powered data center energy optimization), Piclo Energy (demand response marketplace), Leap (demand response aggregation platform), and Texture (AI platform for coordinating energy operations) that are working to help power grids better manage energy supply and demand. Meanwhile, companies like KoBold Materials and Vulcan Elements are working to unearth the rare minerals needed for alternative energy development (e.g., cobalt, lithium, and nickel needed for batteries). New companies have also spun up, bringing AI into the fold for optimizing power generation, such as Juna AI (AI agents for manufacturing optimization), Applied Computing (AI optimization platform for the oil, gas, and petrochemical industries), and Equilibrium Energy (software for managing power grid volatility).
Grid and Power Operations
Grid operations companies are working to modernize the aging U.S. power grid by introducing advanced grid intelligence, alleviating congestion, and ensuring uptime. Within this space are transmission companies like Meter, which provides high-performance internet infrastructure for enterprises. There are also grid management companies like Gridware (AI-powered electrical fault detection), Rhizome (AI platform for modeling and mitigating climate-related risks to electrical utilities), and our portfolio companies Urbint (AI platform that identifies threats to workers and infrastructure) and Arcadia (AI-powered utility data).
Data Center Development & Operations
Data center development and operations companies are working on data center site selection, build profile, permitting, and interconnection tooling. Examples include Phaidra (an AI system for industrial optimization), CivilGrid (site selection software), Trunk Tools (an AI-powered construction project management), Ocient (data analysis for compute-intensive workloads), Claros (data center power delivery optimization), Fluix AI (AI copilot for data centers), and Waterplan (AI-powered water security and compliance).
AI Infrastructure
AI infrastructure companies are those building, supplying, or operating the physical and digital foundations required to train, deploy, and scale AI. The Salesforce Ventures team is focused on companies in this category, prioritizing sustainable solutions. One example includes Crusoe, which is building vertically integrated AI infrastructure.
_
It’s no longer enough to optimize AI workloads — we need to optimize the infrastructure that powers them. We believe founders building the energy stack for the AI age will play a critical role in enabling the next wave of enterprise innovation.
We’ll be in New York for the Climate Capital Summit during Climate Week and look forward to meeting fellow investors and founders who work at the intersection of AI and energy.
