Axios Pro Exclusive Content

Dan Goldman's electricity reality check

Photo illustration of Daniel Goldman of Clean Energy Ventures with a photo of electric transmission towers and blocks of color

Photo illustration: Sarah Grillo/Axios. Photo courtesy of Clean Energy Ventures

The surge in electricity demand unleashed by AI and data centers has troubling implications for certain sectors of the energy transition.

State of play: The future of green steel and green hydrogen rests on access to cheap electricity, and these nascent, thin-margin startups have to compete with ultra-profitable tech giants.

Context: Veteran climate investor Daniel Goldman talked with Axios at the CERAWeek energy conference in Houston this spring, and again yesterday.

This interview has been edited for length and clarity.

What have you been paying attention to?

  • "There is a fundamental risk-return imbalance that is leading infrastructure development to be developed much slower than it otherwise would be.
  • There are structural problems: interconnection, an outdated, balkanized grid, higher interest rates, supply chain, continuing inflation.
  • Then there's the challenge of load growth. Even if we had 0% load growth, we would have a huge decarbonization challenge. Now we've got doubling in the next decade or so.
  • The need to decarbonize is an insurmountable problem. I wake up at night thinking, 'That's an impossibility.'"

You think it is insurmountable?

  • "As viewed from today it's insurmountable. Electrons are going to float to their highest value. That's probably data centers. So if you need electrons to make hydrogen, if you need green electrons to decarbonize steel or cement or manufacturing, it's going to be really challenging.
  • One question we've been asking is, what does it take for data centers to go off the grid? There's also opportunity in efficiency."

That term "insurmountable" — does that mean that the climate crisis is insurmountable? The energy transition is insurmountable?

  • "If we believe that green hydrogen is necessary to decarbonize industry, it's going to be really tough to develop the scale of green hydrogen, at least in North America, to address the climate crisis.
  • There are other parts of the world that might do it effectively. China's building some 25 nuclear plants — we are not building more than one or two nuclear plants, because we can't figure out how to do it in a structured, very low-cost modular approach.
  • That would be a huge help: to bring large, firm, zero-carbon power onto the grid in a shorter period of time. But it takes 10 years for us to build a nuclear plant. If every nuclear is going to take 10 years, we're not going to be able to depend on that power source.
  • Geothermal could provide alternatives for firming low-cost power. That's a promising area, but also a long timeframe for development."

What do you see filling the gap now?

  • "Energy efficiency. There's a massive amount of low-hanging fruit across residential, commercial and industrial."
  • In yesterday's conversation, Goldman added, "The intelligence we're getting out of AI could have tangible benefits to the grid. There's this other side, where we may be able to do things faster, cheaper, in a more cost-effective and easily financed way with AI aids."

Are there industrial sectors that have been trying to electrify, which before you were open to, but because of scarcer electrons you're moving away from?

  • "Not yet. We tend to think globally about electrification, and some of this is going to move offshore.
  • If you think about the desire to on-shore mining and the whole supply chain, that becomes challenging with more expensive or fewer electrons.
  • Electrons are going to be available. They're just going to have some price associated with them. We're not going to run out of power, we're just going to have higher prices."
Go deeper