The discussion, guided by Precourt Institute for Energy director William Chueh, was part of Big Ideas in Sustainability, a Stanford Doerr School of Sustainability series that invites global thought leaders for conversations that engage the Stanford community.

Stanford President Jonathan Levin introduced the talk as an opportunity to learn how sustainability can serve as “an enabler as opposed to a constraint and, in fact, a driver of competitive advantage,” he said.

Sustainability lessons from Delta

According to Hai, Delta Electronics reduced its energy use and carbon dioxide emissions while increasing revenue and profitability over the past 25 years.

Hai urged companies and countries to decouple greenhouse gas emissions from revenue and GDP growth to achieve net-zero emissions. Even if emissions grow more slowly than revenues, he said, they’re still growing and will never hit zero.

It’s important to put energy saving into all managers’ key performance indicators, Hai said. “It’s going to be related to your performance review, then your year-end bonus. Also, the good performance will get rewarded.”

Improved efficiency is essential

When Lovins started publishing articles and books about energy strategy in the 1970s, the world was focused on producing more energy. He started urging people to consider how they wanted to use energy, and the quality, source, and scale that “would do the job the cheapest way.”

The U.S. is using 65 percent less total energy per dollar of real GDP (adjusted for inflation) in 2024 than in 1975, Lovins said. “That just scratches the surface, though, in how much efficiency is available and worth buying.”

Efficiency is not just low-hanging fruit, but fruit that has “fallen down off the tree,” Lovins said.

He proposed that the world could double or triple its efficiency – while saving $5 trillion a year – by shifting from fossil fuels to renewable primary electricity, mainly from solar and wind. Even greater efficiency gains could become profitable – and more appealing to business leaders – if buildings, vehicles, and factories are designed as “whole systems for multiple benefits,” he said.

According to Lovins, the pace of improving efficiency needs to double.

“When you’re going to be making big changes anyway for some other reason, that’s when you do your big retrofits,” Lovins said. “That set of changes can often make your diminishing returns expectation turn into this kind of a curve, where the cost goes down again and big savings become a lot cheaper than small savings.”

Carbon pricing and AI data centers

The conversation also touched on two timely energy issues: carbon pricing and AI data centers.

Hai said the price of carbon has to be high to make more people care. Delta’s internal carbon pricing is $300 per ton, and in 2024, the company collected about $150 million by charging its business groups and factories for carbon emissions – which the company reinvests in projects to reduce its carbon footprint.

“We have seen some dramatic reduction in certain business groups,” Hai said. One group, for example, “used to be the highest emitter, and now reduced [emissions] by almost half,” Hai said.

Chueh asked Hai and Lovins to share their perspectives – from inside and outside of industry – on AI’s energy needs.

Lovins said the current revenue model for AI data centers doesn’t add up. He questioned whether AI computing demand will need all the proposed data centers and whether the data centers will thrive long enough to justify construction of new power plants.

“If you’re counting on this data center to pay its electric bill for several decades to pay off the power plant, it’s going to have to sell about four times more AI service every year to keep that up,” he said. “Is that plausible?” He encouraged companies to run data centers differently, for example, so they draw power only at times when the electric grid has energy to spare.

Hai challenged Stanford students to design a 20-watt integrated circuit chip to allow for more efficient servers.

Or, Lovins suggested, solar microgrids could power data centers and electronics on-site and potentially send excess power back to the grid. “A negative power supply may be a better solution,” he added.

Hai agreed, noting that it’s better to save energy than to build new power generation facilities.

Inspiration for students

Chueh closed the conversation by asking Hai and Lovins to share a challenge they’ve overcome that could inspire students.

Hai said the most critical moment in his career was switching from banking to electronics and joining Delta. When founder Bruce Cheng asked him to lead the company, Hai was worried he didn’t know enough, but Cheng encouraged him.

“You really need to have a very good team working with you to do things, because you alone, even if you are a superstar, you cannot accomplish so much,” he said.

Lovins said it took him a long time to learn to focus on outcomes, not motives, when talking to people with diverse views.

“Whatever change we are each trying to make in the world, I would listen a lot to what other people are interested in and talk about that, and see what agreement you can discover for their reasons, which are just as good as your reasons,” he said. “Don’t argue about the reasons.”

Stanford Doerr School of Sustainability Dean Arun Majumdar closed the event. “What we learned today is a combination of not only engineering, systems-view economics, but it’s also, as you pointed out, the culture, the mindset of how you think about an issue,” Majumdar said.