AI's Hidden Crisis: Heat is the True Bottleneck

What we talked about.

In this follow-up conversation, Sunny Ray talks with sustainable finance veteran Jonathan about the overlooked physical constraint now limiting AI growth: heat. Jonathan explains why cooling, not chips or power generation alone, has become the binding bottleneck for hyperscale data centers, where thermal management can consume 40 to 60 percent of electricity spend. He argues air cooling has hit its physical limits, pushing the industry toward liquid and immersion solutions. The discussion widens into energy history and finance, covering why renewables failed to insulate China and the EU from Strait of Hormuz oil shocks, why green hydrogen was one of climate finance's costliest missteps, and why efficiency investments like Boeing's winglets matter more than chasing new energy sources. Jonathan also introduces his new venture, the Sustana Cool Climate Fund, which invests across buildings, food cold chain, grid and electronics cooling in North America, Asia, Europe, Latin America and Africa. He makes the case that emerging market cooling, from Indian small business AC subscriptions to solar powered rural refrigeration, is a venture opportunity, not charity.

A sustainable finance veteran explains why heat, not chips, is the real bottleneck choking AI's growth.

The questions, and the answers.

You said renewables didn't protect China or the EU when Hormuz got threatened. What did people who built the renewable narratives miss?

Most oil isn't used for electricity, it's used for heat and materials like the parts and lubricants in solar and wind equipment. Electricity is roughly a third of oil use, so people mistook that third for the whole. There's no facility making polysilicon with solar power. Renewables can be a component, not a substitute, for high temperature chemical processing, so China and Asian economies stay far more exposed than the US or Brazil.

China has more solar and EVs than anyone. So what's the failure mode you're pointing at?

China has more coal than the rest of the world combined, plus cheap hydro, so it makes sense to save oil for manufacturing since electricity can't substitute for making high performance materials. They use electricity for citizens' EVs while husbanding oil for the third of use that can't be replaced, keeping people employed in manufacturing. A prolonged blockade would still hurt many Chinese businesses badly.

You've been in sustainable finance since 1999. What's the most expensive narrative the climate finance industry got wrong, and who paid for it?

Green hydrogen is probably the most egregious one, you burn more natural gas to make it than if you just used the gas directly, and countries like the US and Norway spent billions on it. I'd also say stymieing US natural gas after it let America decarbonize faster than Europe was a strategic mistake, since it cleaned the air near power plants and hyperscalers are now rediscovering gas as their fuel of choice.

You keep showing up early to the right energy bet. What are you actually looking at when you make those calls?

I put together pieces that already exist and check what's consistent and inconsistent, recognizing that energy has to come from somewhere, you can't claim a perpetual motion machine. A lot of clean tech financing is political geometry, getting language into a bill so your technology gets covered, rather than independent analysis. Lithium batteries got treated as a renewable resource despite losing energy every cycle and carrying fire risk, because government got it wrong.

You held Exxon and Boeing. Why was that the right call, and is it still?

The real environmental gains come from energy efficiency, not a new form of energy, and bringing capital to efficiency accelerates improvement. Boeing's winglets, a three foot nub on a wing, save a huge amount of fuel multiplied across the whole fleet. The engineer who invented them left Boeing after management wouldn't back it, built the company independently, and Boeing eventually bought it. That's a tangible efficiency investment.

Why cooling, why now, why a dedicated fund instead of a generalist climate fund?

Cooling is the climate adaptation most critical to humans surviving on Earth, it's why we can live in hot places, why food is safe, and why data centers can manage thousand watt chips. Old air based cooling isn't adequate anymore, so there's an explosion of innovation across aerospace, buildings and other industries. We want to crystallize the benefits of that future energy efficiency by investing directly in the technologies enabling it.

You've got over 100 companies in your pipeline. Can you give me the shape of it, categories, stages, geographies?

Geographically it's mostly North America, Asia and Europe, with some from Mexico, Brazil and Africa. By category roughly 30 percent focus on buildings, 20 to 25 percent on food, another 20 percent on grid and electronics, and the rest span multiple verticals. Compared to solar's ten year payback, energy efficient cooling often shows four year simple paybacks, so there's real opportunity to bring modern finance into the space.

AI data center cooling gets all the media attention. How much of your fund is that versus cold chain, real estate, food? Where's the actual alpha?

Food actually uses more energy than data centers and is growing just as fast, even though data centers get all the media oxygen. We'll likely see faster exits in the AI adjacent space, but many of these companies didn't start there, one we're bringing to the Department of Energy began as a real estate hydronic cooling company later adapted for data centers. There's no silver bullet, it's gold buckshot across many industries.