When Fidra CEO Gym Its Seattle Organization describes AI agents built to improve the efficiency of the Data Center, who can imagine an army of digital mines in the back of the screen to work away.
Their job is to concentrate on cooling electronics-Fine-Tune Data Center operations — which usually accounts as 30% of the use of any advantage, earns second with electricity required for data processing operations. They are tracking temperature, voltage, spinning of pumps and other infrastructure to understand how well they are working the center.
The agent’s AI agents work autonstrated and developed through learning reinforcement – observes the results, adapt and improve. Startup reduces the use of energy from being cool by 25% through its technology. Savings subject.
Gao said, “We live in a limited world of energy,” mention that data centers are being built quickly to meet AI’s claim, but cannot hold fuel supplies. “The ability to earn income for these big AI companies is literally limited [electricity providing] Electron available. “
Fedra announced the help of anecdotal funds, helena, index Ventures, Nvidia, Sony Innovation Funds and other investors last week for new funds, its total capital has increased to $ 120 million.
Since Big Tech and Data Center Operators are planning to spend a few billion dollars to create more energy-intensive facilities-or “AI factories” as their calls-the startup will help to further its technology to cut even bigger energy.
AI is coordinating AI agents to optimize an emerging region of focus.
The benefits are also trying to surpass the cooling infrastructure to help manage the data workflies that come through the centers, which can create spikes on energy needs. It requires a system for more power than most of the time, so reduce these peaks can create important savings.
The GAO said, periodically, the benefits should be taken slow time for certain tasks, Gao said.
“This does not happen today because power, cooling and workload management systems all work independently from each other, without adjustment, without orchestration,” he added. “But this is the future that we see – significantly more skilled AI factories.”
Related: Fidra enhances AI Data centers by enhancing energy efficiency ‘not smart, not just tough’ to help $ 50 million to help