AI: AI need to 'play nice' with other Power uses. RTZ #503

...the calculus of future AI power demand curves

AI is the new hungry kid at the Power smorgasbord.

The immense Power requirements of AI infrastructure ramps to scale AI in this AI Tech Wave, is something I’ve discussed regularly in these pages. From the ‘AI Table Stakes’ that include power as a key and supply constrained input for AI data centers, to scaling the AI data centers from tens of thousands of AI GPU super clusters to hundreds of thousands to soon millions of GPUs.

All require so much power, that Microsoft, Amazon and other AI ‘hyperscalers’ are signing up recommissioned Nuclear power plants in the US.

There are some recently updated projections on this AI power ramp, and how this demand will have to contend with other growing sources of energy, from fossil, clean, sustainable, renewable, and beyond. As Axios outlines in “Why AI isn’t the whole data center story”:

“AI's huge power appetite is the talk of energy circles, but what's less certain is precisely how much it explains the wider data center growth underway.”

“Why it matters: It's a key question as policymakers and other stakeholders weigh AI's benefits against its carbon footprint.”

“The big picture: Data centers are one reason U.S. electricity use is rising after about 15 static years.”

• “The International Energy Agency projects data centers will be 6% of U.S. power demand by 2026, up from 4% in 2022.”

• “Further out, Barclays researchers see data centers accounting for 9% of demand in 2030 and far higher in their "upside" case, up from 3.5% today. McKinsey analysts estimate it even higher at 11%-12% in 2030.”

But it’s important to keep in mind that AI is not the only element in data centers, a half trillion dollar industry, poised to a multi-trillion dollar industry in a few years.

“State of play: Right now, constellation of other needs accounts for much more than AI in data center power use.”

• “Think streaming services, storage and databases, payment processing and various business management systems, to name a few.”

• “Rhodium Group director Jeffery Jones estimates AI is around 5%-10% of U.S. data center power use today.”

“What's next: Energy use for training and using large language models like ChatGPT are growing fast from a small base — and spawning huge new data centers.”

• "We're two steps into an ultramarathon, and I mean a really long ultramarathon," Jones tells Axios, noting that generative AI only arrived in earnest about two years ago.”

“Zoom in: Jones estimates that in 2025, "legacy" uses will account for 80% of U.S. power demand from data centers, while AI is 20%. By 2035, he projects, it's 50%-50%.”

• “Absent generative AI, he said there would still be growth in traditional data centers, but it would be fairly steady at 1.5% to 2% annually.”

“The intrigue: For now, there's a range of views in wonk-world.”

“Electricity analyst Rob Gramlich estimates half the growth in data center energy suck in the next 3-5 years will be from AI.”

• “Goldman Sachs research shows that soaring growth from a small base could still leave non-AI uses playing the largest role in the medium term.”

• “It sees non-AI data center energy use rising from 142 terawatt hours in 2023 to 304 in 2030. AI-related uses soar from 4 TWh last year to 93 TWh in 2030.”

The AI Gold Rush thus includes not just the AI chips infrastructure from the likes of Nvidia, but power too from a whole host of industry and critical AI data center component providers like cooling, to increase power efficiencies.

As Axios continues:

“Reality check: It's hard to predict how computing efficiency gains will stack up against rising AI use.”

• “And there's no bright line between what's AI vs. other advanced forms of machine learning. Complicating things further, traditional data centers handle some generative AI use.”

• “This is also a local story. Clusters of huge data centers for AI can bring big localized challenges for utilities and grid regulators.”

“The bottom line: AI is turbo-charging demand, but it's not the only factor.”

The calculus of these power projections will shift of course on a lot of variables. And it’s global.

These including major types of AI use I’ve discussed like AI training, inference, synthetic data, synthetic content. As well as the transition of AI along the AGI roadmap from chatbots, to reasoning to agents and beyond. And of course the expected shift of some AI compute from cloud to local sources.

And the tech industry will have to navigate the types of energy sources from fossil fuels across the spectrum, to renewable sources like Solar, and of course the current focus on Nuclear power. Microsoft’s headline-catching deal with Three Mile Island facility is just a beginning case in point.

As I said before, AI is the new big kid at the power smorgasbord. And will have to play nice with all the other kids also wanting their fair share.

Overall, the industry will navigate building and funding this demand ramp with adequate supply to Scale AI. Not always in a straight line at as fast a pace as desired, but the road ahead is becoming clearer. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)