Phaidra applies artificial intelligence to the controls that operate industrial facilities and, increasingly, data centers and AI factories. Its software aims to help operators choose efficient operating conditions and manage cooling and power use—not to sell a consumer energy device. The company was founded in 2019 by Jim Gao, Veda Panneershelvam, and Katie Hoffman, combining experience in Google infrastructure, DeepMind, and industrial controls.
What Phaidra’s AI controls do
Industrial plants and data centers run on networks of equipment and controls, from chillers and cooling loops to power systems. Phaidra’s original pitch was to use AI software to supervise those controls: analyze operating data, select efficient settings, and identify when equipment performance is deteriorating. Bloomberg News reported in 2022 that the company was pursuing power plants and other industrial facilities, including steel mills and vaccine manufacturers as target facility types.
The idea is to make use of operational data that facilities already collect. As Jim Gao put it in the 2022 Bloomberg report, “They’ve been collecting data for so long, but they haven’t been using it.” The software is aimed at complex, mission-critical operations; it is not a substitute for the plant’s physical equipment or a generic energy-monitoring gadget.
Why the founders’ backgrounds matter
Phaidra says Gao and Panneershelvam applied reinforcement learning to Google data-center cooling systems before founding the company with Hoffman in 2019. In its company history, Phaidra says that Google system achieved up to 40% energy reduction in an already highly optimized facility. That is the company’s retrospective account of a Google cooling application—not a general result for Phaidra customers.
#1 Best Overall
Hoffman brought industrial controls experience to the founding team. That experience is relevant because industrial systems have demanding operating constraints and can be difficult to automate safely. “These industrial systems are incredibly difficult to run on a good day,” she told Bloomberg in 2022.
From industrial plants to AI factories and data centers
The 2022 coverage emphasized power plants and industrial operators. Phaidra’s current public materials put more emphasis on data centers and AI factories, describing a coordinated approach to operating cooling and power systems. Its product lineup includes Prism for intelligence and observability, alongside specialized AI agents.
Rank #2
- Liquid cooling: an agent focused on managing liquid-cooling operations.
- Flexible power consumption: an agent intended to adjust power use in response to grid conditions.
- Chiller staging: an agent focused on coordinating when chillers run.
This is a visible change in the company’s public emphasis, not evidence that Phaidra has stopped pursuing industrial customers. The company describes deployments and partnerships, but public materials do not provide a complete named customer list or independently audited, site-by-site results.
How to interpret Phaidra’s energy and cooling figures
The reported percentages refer to different claims, settings, and sources. They should not be combined or treated as a guaranteed outcome for a new site.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →| Figure | What it refers to | Evidence and qualification |
|---|---|---|
| Up to 30% lower plant energy consumption | Phaidra’s industrial control system | A Phaidra claim reported by Bloomberg News in 2022. The article did not provide a universal result or a detailed measurement protocol. |
| Up to 40% energy reduction | Google data-center cooling | Phaidra’s undated company-history account, accessed in 2026, describing an already highly optimized facility. It is not a general Phaidra customer result. |
| Approximately 80% lower thermal overshoot; 30% lower cooling costs; 25% better energy efficiency | Phaidra’s current product and company messaging | Figures on Phaidra’s undated homepage, accessed in 2026. The reviewed page does not give detailed baselines or independent validation. |
For a facility evaluating control software, the relevant question is not whether a vendor advertises a large percentage. It is what equipment and operating conditions the result covers, how the baseline was set, and whether the measurement was independently verified. The reviewed public material does not establish universal savings, detailed site-by-site efficacy, current pricing, or full contract terms.
Why deployment is difficult
Controls in a plant or data center affect mission-critical operations. Bloomberg’s 2022 report noted that facilities have site-specific requirements, often rely on legacy controls, and may lack the engineering capacity to maintain advanced autonomous control. An AI system must work within the facility’s operational limits and integrate with its existing equipment and controls; an advertised efficiency gain alone does not show how those practical requirements are handled.
When assessing an AI controls platform, operators should ask how it integrates with current building or industrial controls, how safety limits and operator approval work, what happens if the software or connection fails, and how fallback behavior is tested. They should also request independently measured results against a clearly defined baseline, along with information about reliability, downtime impact, implementation effort, and ongoing maintenance.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What Phaidra’s funding and partnerships establish
Phaidra announced a Series B of more than $50 million in 2025 through a release distributed by PR Newswire. The release lists NVIDIA among investors, and Phaidra identifies NVIDIA as a partner. These details indicate funding and an ecosystem relationship; they do not validate the company’s energy-performance claims or establish a particular customer outcome.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
