Will Supercapacitors Come to AI’s Rescue?
May 6th, 2025Musk had to use expensive Tesla Megapacks to handle the massive power fluctuations at xAI:
AI datacenters, especially during training, bring extreme power fluctuations due to the nature of neural networks (gradient descent).
xAI has the largest AI data centers. xAI is using Tesla Megapacks to handle millisecond power fluctuations.
Supermicro provided the racks and systems for the xAI 100,000 GPU data center. They installed 64 H100 GPUs in each rack.
Elon Musk has said that they are adding 50,000 H100s and 50,000 Nvidia H200 GPUs to double the compute power of the data center.
In an interview with Lex Friedman, at 14:53, Musk said, “If you suddenly see giant shifts, 10, 20 megawatts, several times a second, this is not what electrical systems are expecting to see.”
Via: IEEE Spectrum:
Because training is orchestrated simultaneously among many thousands of GPUs in massive data centers, and with each new generation of GPU consuming an ever-increasing amount of power, each step of the computation corresponds to a massive energy spike. Now, at least three companies are coming out with a solution to smooth out the load seen by the grid—add banks of huge capacitors, known as supercapacitors, to those data centers.
“When you have all of those GPU clusters, and they’re all linked together in the same workload, they’ll turn on and turn off at the same time. That’s a fundamental shift,” says Joshua Buzzell, vice president and data center chief architect at power equipment supplier Eaton.
These coordinated spikes can strain the power grid, and the issue is promising to get worse rather than better in the near future. “The problem that we’re trying to solve for are the language models that are probably 10 to 20x maybe 100x larger” than the ones that exist today, Buzzell says.
Related: Inside the small town where Elon Musk’s supercomputers have left residents struggling to breathe
