Phaidra Secures $12M for AI-Driven Energy Efficiency in Data Centers

  • Home
  • Industry News
  • Phaidra Secures $12M for AI-Driven Energy Efficiency in Data Centers
DateJul 4, 2024

Phaidra, an innovative AI-based control system specializing in enhancing the energy efficiency of facilities like data centers, has secured $12 million in new funding led by Index Ventures. This investment brings the companys total capital raised to $60.5 million. The additional funds would enable Phaidra to further develop and expand its autonomous control systems, which leverage reinforcement learning to optimize power usage in data centers while ensuring reliability.

This funding announcement would come at a crucial time for the data industry, where rapid advancements in AI are heavily dependent on reliable and scalable compute infrastructure.

Data centers, the critical backbone of AI, are experiencing exponential growth. In the United States, which represents 40% of the global data center market, demand is projected to soar from 17GW in 2022 to 35GW by 2030. Data centers are notorious for their high energy consumption, using 10-50 times more energy per square foot than a typical commercial office building.

With the proliferation of AI models that depend on resource-intensive GPUs, existing markets would find it increasingly difficult to meet demand. For instance, in Northern Virginia, the world’s largest data center market, only 0.2% of grid capacity remains available. In Ireland, data centers are expected to account for 32% of the national electricity consumption by 2026. This dire situation would underscore the urgent need for significant improvements in data center infrastructure and efficiency to sustain AI’s growth

Martin Mignot, Partner at Index Ventures, emphasized the core issue: “For all the talk about AI and expanding compute capacity, the fundamental problem is power. There is no AI without sufficient energy. Through our research, we realized the critical importance of investing in solutions that address the downstream effects of computational growth. Phaidra exemplifies this with its innovative team and product, perfectly timed to address this challenge.”

Translating into Millions of Dollars in Cost Savings

Phaidra’s AI control platform would transform industrial facilities into intelligent, self-learning systems that blend the precision of manual methods with the scalability and efficiency of AI. The platform would integrate seamlessly with existing building management systems to capture real-time data, which is then processed by its cloud-based reinforcement learning agent. This agent autonomously adjusts the settings on individual components to achieve optimal performance for the entire system. Unlike traditional control systems, which degrade over time and require regular manual updates, Phaidra’s system would continuously learn and improve. Early adopters of Phaidra’s technology have reported significant efficiency gains, translating into millions of dollars in cost savings and increased power availability for revenue-generating services.

Phaidra’s founding team brings together a unique combination of engineering expertise and cutting-edge AI research. Co-founder and CEO Jim Gao was inspired by the AlphaGo documentary to explore the application of machine-learning technology to improve efficiency in Google’s data centers, where he worked as an operator. His enthusiasm attracted Vedavyas Panneershelvam, now Phaidra’s CTO, who was a primary research engineer on AlphaGo. They collaborated at DeepMind, developing AI-powered solutions that reduced the energy needed to cool Google’s data centers by 30%. After publishing their findings, they were joined by Katie Hoffman, an industrial control systems engineer from Trane Technologies, who became Phaidra’s third co-founder and COO.

“The data center industry is rapidly evolving and reaching a pivotal point due to the accelerated demand for AI,” said Jim Gao. “Data center operators recognize the need for greater efficiency to meet this rising demand. At Phaidra, we aim to revolutionize data center management with an AI-powered control system that adapts in real-time and continuously improves. This leads to substantial cost savings and more revenue to fuel further growth.”

Top Talent from Google, DeepMind, Meta, Amazon

Phaidra’s team of around 100 includes top talent from Google, DeepMind, Meta, and Amazon, as well as specialists from leading engineering firms like Trane and Johnson Controls. The new funding will support ongoing investment in research and development, implementation, and customer success, alongside expanded go-to-market efforts.

Martin Mignot of Index Ventures expressed confidence in Phaidra’s potential: “Meeting the team and hearing from their customers makes it clear that Phaidra is uniquely positioned to address this problem head-on. No other startup combines domain expertise and AI research prowess like Phaidra, evidenced by their impressive performance and results. We are excited to support Phaidra as they advance solutions crucial for the growth and sustainability of data centers, AI development, the technology industry, and the world.”

This investment marks a significant milestone for Phaidra as they continue to push the boundaries of what AI can achieve in optimizing energy usage in data centers, ultimately supporting the broader tech ecosystem’s growth and sustainability.

Leave a Reply