Sridhar: AI computing loads rapidly swing between 10% and 100% capacity
K.R. Sridhar · Frontier Forum: The new power map for AI infrastructure · Oct 21, 2025 · at 16:42
Bloom Energy CEO K.R. Sridhar discusses the unique power delivery challenges of artificial intelligence computing clusters compared to traditional CPU data centers.
0:00 / 0:32exact quote · 32.4s
720p mp4 · rendered on demand · StarZero watermark
“The AI load, unlike the CPU load, is very volatile. It has milliseconds, seconds, and minutes variation that go from 10% of a load to a hundred percent of the load. So you're talking about, if it's a hundred megawatt data center as an example, you're talking about the load swinging from 10 megawatts to hundred megawatts up and down And some of those variations happen in millisecond frequency, some of them happen in seconds frequency, some in minutes frequency.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from K.R. Sridhar
Sridhar: AI revolution cannot be powered without natural gas
“And without natural gas, we are not going to be able to power the AI revolution. That's a fact.”
Sridhar: Politics will block taxpayer-funded grid expansions for hyperscalers
“Now, if the grid were to build excess capacity purely to meet this huge demand and expects the taxpayer to pay for it so the hyperscaler can benefit, That politics is never going to fly.”
Sridhar: Grid upgrades cannot match AI power demand within 5–7 years
“Realistically, over a five, six, seven year period, it cannot catch up with the demand base that exists out there.”
Sridhar: High-power AI data centers must adopt captive on-site power
“We are finally building high electricity intensity AI factories. These are, these data centers are AI factories. They have been the exception and not the rule. They are now going to have to follow the rule.”
Sridhar: Inference will account for 90% of AI power load within four years
“Three to four years from now, I predict that 90% of the load will be inference and 10% will be training, even though the 10% in training then will be significantly larger than the 90% training today.”
Sridhar: Data centers will shift to 800-volt DC architecture
“The second pain point that they are working on as we speak is shifting the mindset to an 800 volt DC bus going into Iraq. Let me explain this to you. When you are a large power plant that produces 50 megawatts, 500 megawatts worth of power, because so much pow…”