Bill Nye & Nicky Preston: Sustainable AI | Dreamforce 2026

Please welcome Sonja Norman, Bill Nye, and Nikki Preston.

We love that song.

Science does rule. I mean, maybe that should have been the title of this session. We missed the mark on that. And Nicky and I are a little sad that we didn't get our own chance, but we understand. We have a legend in the building.

It's okay.

My name is Sonja Norman. I lead Impact at Salesforce, which is all of our philanthropy, our corporate sustainability, and our tech for good. Such an honor to be with my 2 colleagues here today. No introduction really required, but Emmy Award-winning science educator, bestselling author, advocate for science literacy, Bill Nye.

Thank you, Nikki. Yes.

And then a very special customer of ours, Nikki Preston, who leads sustainability for One New Zealand, which is the largest telecommunications provider in New Zealand. Welcome to both of you.

Thank you.

Thank you. So those of you who were at the keynote this morning, you heard that one of Salesforce's longstanding core values is sustainability. And we've been working with our customers on a pilot for AI sustainability now that AI is the focus of the world and our field in technology. And so you've been in this customer pilot with us, Nikki. I'm curious, why did you say yes when we came knocking? What made you say, yes, we wanna explore AI sustainability with you, Salesforce?

It's a great question. And first of all, thanks. Thanks for having me. I'm super excited to be sharing the stage with Phil. So that's awesome. It's a good question. And at One New Zealand, we're really all in on AI, but also the positive benefits of AI. I think there's a lot of scaremongering out there, a lot of questions, and we really do believe that technology and AI can be for good. We've already seen some proof of that. We've been using machine learning in our nationwide network of cell towers, kind of putting them on microsleeps to save energy. We saved 14, 1-4% energy use last year by doing this. And so when Salesforce came to us and said, do you wanna be part of this program? We said, yeah. So, Super keen to get involved, really believe in the ideals that Salesforce has. And so great to partner.

Well, thank you for doing that with us. And Bill, I think artificial intelligence and kind of the magic of the AI revolution can feel really abstract and conceptual for a lot of us. You've spent your life making science tangible for folks. So how are you thinking about how to understand the physical impacts of AI, the energy use, the carbon, the water, How do you frame that in your own head? How would you explain it to all of us?

Well, for me, it starts with computer systems, which rely on transistors, which are made of silicon, like the valley, and you can imagine if you have enough transistors, you have enough switches to go on or off, and then you can imagine once you start using your computer, you can imagine a computer that could have been programmed or been used so many times that it has developed patterns. And the thing that's gone— the thing that's got everybody freaked out and running in circles screaming is now we're able to build computer systems so fast that they can have output that you can't anticipate. Like, when I was writing computer programs with punch cards— I was, you guys, back in the disco era. We had punch cards and things would go wrong, but you knew what was going to come out was a result of what you put in. But now these systems are so sophisticated, you don't know what's going to come out. But because it's so useful, because the systems are so useful, we're just throwing electricity at the problem.

Mm-hmm.

And then when you throw electricity, You gotta have cooling, and that's the second law of thermodynamics. There's no escaping it, people. And so this is why everybody is so— this is why everybody wants it and why everybody's concerned. Now, I'd like to remind everybody, I'm an engineer. I used to have a job. Okay? If you're ever on a 747, don't worry. I was very well supervised. Everything's cool. But I think about a commercial airliner. It is a very powerful thing, and yet built in day and night, all the time, everybody there, at least when I worked at Boeing, everybody there was worried about safety.

Mm-hmm.

And what I think has the public concerned right now about artificial intelligence is not everybody is sure the people building these things are concerned about safety. Am I wrong? So, Nikki, let me ask you, where does all the electricity come from for New Zealand One?

Great question. We actually have a 100% renewable energy target. We're at 80% of our energy use, and New Zealand's grid is about 86% renewable. So we're quite lucky, I guess, from design that our electricity is mainly—

So is it wind and solar? Is there—

Hydro. Hydro. Yeah. So massive water dams that were built kind of in the '60s. So yeah, that's where a lot of our energy comes from.

But that still leaves 20%.

100%. And actually you also think about the transport fleet. You know, we are on our journey to be putting more EVs in the country, but you know, a lot of, I have a petrol car because I live in an apartment and can't charge it. There's all of those infrastructure conversations that need to happen as we transition to more clean energy use. There's also, not all data centers are built equal. For example, there's closed water loop cooling systems that use, One lot of water through. So, you know, there's ways that we can think about good, clean infrastructure.

Plus, New Zealand's in temperate latitude, right?

100%.

So, you know, there are over 7 million people in the Bay Area.

That's more than New Zealand combined.

Yeah, New Zealand's just about 5 million.

So— It's a tiny country.

The longest journey starts with but a single step.

100%.

If you can solve the problem in New Zealand, then, you know, why not the world?

And this is actually—

I was at a conference last week, the Climate Change and Business Conference. which happens every year. And a big thing that comes up in New Zealand is everyone goes, well, we're such a small country, does it matter? Why should we be pushing harder on this? But one of the conversations they were having is if every person globally thinks about becoming a little bit better, becoming a little bit more efficient, using more clean energy, the impact can be really great. So think about that compounding impact terms as opposed to just thinking about your individual actions.

Well, the old saying is we all share the air.

100%, yeah.

You're not going to meet anybody, you're not going to have much of a conversation with anybody who's not breathing.

No, and I want to keep breathing as well.

So everything every one of us does affects everybody in the world. It's like, whoa, dude, dude.

The interconnectivity, isn't it? Even, I mean, I flew over 12 hours from New Zealand here in a place with other companies have similar but different challenges. We're all fighting these same concerns. How do we become greener? How do we become more efficient? How do we become more sustainable, more ethical? Um, so it's really cool having these conversations and trading ideas around how we can do those things.

So Nikki and I, you and I are both representing businesses, and I think it's important to name that, of course, we feel deep responsibility to take action on these issues, but there's real business case in making sure that we're more sustainable, you know, thriving planet, thriving business. At the end of the day, all of our businesses rely on natural resources and human capital. We have human workers, which really, that's what a company is.

So hold on, you still have human workers?

I hope so.

Tell us about that. How does that work?

Absolutely. I'm one of them in the flesh. But it's so important for business continuity to make sure that our customers are successful. And one of the controversial takes that my team has around AI sustainability is AI is a tool that actually isn't for every single use case. And I know, Bill, you've been advocating this too, the idea of smart use.

Oh, doing more with less.

Yeah. Yeah. Can you talk a little bit more about what you're seeing?

So 100 years ago, okay, maybe, maybe it was 20 years ago, I met Rick Smalley, who got a Nobel Prize for discovering buckyballs, Buckminsterfullerene Molecules. And his big thing was not to do less. I remember, I remember, I do, the first Earth Day. I grew up, some of you may know, I grew up in the city of Washington, D.C. There was a few of us. I rode my bicycle to the Washington Monument. I locked it to one of the flagpoles. If you were to try that today, it'd be a different outcome. But anyway, The deal in those days was do less. You just eat less, drive less, don't wash your clothes, just— but that's not what people want. People don't want to do less. So Rick Smalley's big thing was, let's do more with less. And he just— a detail, he felt that electrical transmission lines could be made out of—

Mm-hmm.

Barkminster flooring. But it's really an important idea that The goal is not to do less, not to criticize people for using energy to do anything, but you want to do it efficiently. Of course, it makes perfect sense. If you can do things efficiently, then you'll have more out for less in, and everybody wins, including the shareholders. No, really, it's a real thing. It's a way of looking at any problem, and you work in cell phones, mobile phones, as they say in New Zealand, and you guys, you drive on the other— the whole thing.

I almost got in the wrong side of the car yesterday because I forgot that it's the other side.

You get used to it pretty quick. I was driving a stick shift on the wrong side, but all that aside, everything that I think— almost everything that we're talking about when we're talking about sustainability is how we use electricity. And go ahead.

Well, I was just going to talk about trade-offs as well, because I think one of the key things with sustainability is making better decisions. Knowing more, having more data can help us to make the decision that may be a trade-off. So working with the Salesforce Impact Team, shout out to Stefan and Alex, who have done an incredible job. They did modeling and showed that us moving our agent force workloads from the US to Sweden saved an estimated 83% in terms of carbon emissions. So great reduction in carbon emissions, awesome. But the trade-off was that we used significantly more water. So Sweden—

Why?

I think the data centers, they were using—

Cooling.

For cooling, yeah, yeah. But Sweden's a low water, or in a low water stress region, so that trade-off felt like an okay one, you know, less carbon emission, fewer carbon emissions, more water, but water's not an issue there. So the lesson there is not move everything to Sweden. I mean, you know, there's other countries like New Zealand that has a renewable grid and it's, we use, we've got an abundance of water. But the point is, if you know what your footprint is, you know your impact, you can then make decisions that hopefully create better outcomes. But there is a trade-off and I think it being clear about that and being transparent about that Is, is okay. And we try to do that within our use of AI with our customers. We found customers and stakeholders just want a bit of honesty, um, and try and outline those things and try and do better.

Do you guys make a deal with customers who use less?

In terms of mobile phone use less?

Yeah, like I'm trying to think, use less electricity. You were just comparing tweeting to texting or something.

Well, an interesting one actually is, um, each year we see data use go up. by about 20— well, 26% last year. Customers use their phone's data more, 26% more.

Fucking amazing.

Exactly.

Yes.

But last year—

Boom.

Yes.

Not going to comment on that, but yes. Yeah. But with all of our energy efficiency initiatives, we've been doing a bunch of work across our network. We've got thousands of cell sites across the country. We actually used about the same amount of electricity So doing things like the microsleeps and things. So even though 20% more use, 26% more use, but same amount of electricity. So if we can do that more, think about how we can be more efficient, it's good from an environment, also good from a shareholder perspective, saves you money. So finance team loves that. So thinking about how do we create some win-wins in this space.

Yeah, we're seeing some win-wins in domain-specific models.

So, oh, domain-specific model. Yeah. Yeah. Oh, you don't need to know. Everybody's hip to domain-specific model.

We just made an announcement of partnership with NVIDIA. And essentially the takeaway there is that for enterprise use cases, you might want a domain-specific model, a model trained on enterprise data. So if you're a healthcare company, healthcare-specific data, so you need to train on the entire internet. for a healthcare use case. And so we're really excited because that actually provides better outcomes for the customer and the end user. It's more accurate, it's more grounded in, um, the context that's valuable, but also will have lower compute and, um, and energy efficiency as well. So there's a lot of creative thinking and innovation happening, I think, in pockets. I know, Bill, you're thinking about global scale and policy as well. When you're talking to decision makers who are policymakers or in a position of power, what are you advocating for?

What am I— with pos—

what?

So, excuse me. Yes, I, I'm often consulted by people in—

You should be.

This administration. So before we go on, isn't that hilarious? Before we go on, you know, I'm still associated with The Planetary Society, and we use Salesforce products. Woo! And I guess we are domain-specific. I guess we choose which products, you know, we have 42,000 members. Who's not a member of The Planetary Society?

Anyone?

Is there anybody who is? Come on. Yeah, I love you, man.

Hello!

All right. So I am not consulted, and I encourage everybody to vote, This year, you know, this is really important.

At the conference?

Well, just as you know, the US has now an administration, a third of the government doesn't accept the fact of climate change, and they're not— don't seem to be hardly concerned about efficiency or especially sustainability. And the word energy now has been equated with oil use, which didn't— that didn't used to be the way. We'll turn things around as long as everybody votes. Vote, vote, voting, vote, people. With that said, when I was consulted, what we want to do is get especially young people excited about understanding energy, understanding where electricity comes from and how we use it and the remarkable the magical properties of electricity, and want everybody to have a rudimentary understanding of computers, or transistor logic that makes computer decision-making possible. It's easy to say, but if everybody just appreciates it, then I think we will have better— we'll have a better society writ large. It's easy to say, but it has to do with education, and you want education at every level, and you want to educate your customers, right? And provide these opportunities. But opportunities for people to work together, for— to use, to do more with less. And then everybody wants what I— what can I do? What can I do if I recycle this box? Well, you laugh, but that is empowering for consumers. It's a relevant thing. So What we want, I say all the time, 3 things for everybody on Earth. First of all, the big idea is to raise the standard of living of women and girls. When you raise the standard of living—

Woo!

Thank you, thank you. I'm the first guy ever to come up with that. When you raise the standard of living of women and girls, everybody is better off, everybody is happier. Woo, yeah. Happy women and girls, happy life, something like that. And then for that, To me, there's 3 things: clean water, access to the internet, the electric computer machines that the kids use, and then renewable electricity that's reliable, reliable renewable electricity. If we had those 3 things, we could do everything, and that to me is where artificial intelligence is going to change the world because it enables these efficiencies. You can move electricity around. Automatic, perhaps, automatically. You can influence consumers to charge their electric vehicles off hours when the grid is not as strained. And then you could have policies made that enable all 3 of these things. If I were king of the forest. And so did I mention voting? Yeah, so—

You did.

I got that. Let's vote. And, you know, it's never been this weird. And I, you know, I remember when Nixon resigned, and that was a big deal.

Well, I have—

It was a huge deal.

I have heard that voting and where you work are the 2 biggest impacts you can have on your environmental footprint. So I 100% agree with you.

There you go.

I love that. And I love what you called out about this being magical. Like, we take so much for granted. The fact that you can fly on the plane that you helped to create at Boeing, the fact that you can pick up a device in the morning and get access to news from around the world, you know, communicate with another human in another country.

Everything in this room owes its existence to electricity. You wouldn't have any metal without mines, you wouldn't have these lights, you wouldn't have the upholstery, you wouldn't have this, right? You wouldn't have any clothing, nothing without electricity. And so we want, we want it available to everybody in the world, and artificial intelligence systems, I think, will make that easier and easier if we don't blow it.

Nicky, we've talked a little bit about the business case for sustainability. For us, we find that it helps to deepen trust with our customers, our partners, folks like One New Zealand. What are you seeing from your stakeholders?

Very similar. We run an annual AI trust report. Well, we've done it twice, so now I think it's annual. And we ask consumers, so everyday Kiwis, New Zealanders, how they think about businesses using AI. The most recent survey, we asked whether people care about how businesses use AI and the environmental impacts of it, 45% said yes, they do care, and 63% of people aged 18 to 24-year-olds, so probably more conscious consumers, cared. And the things that they were asking for was renewable energy, efficient use of electricity, considerations around water use, e-waste, or the physical infrastructure behind it. And so, We know consumers care. And so what we are doing about that is we've got responsible AI principles of which sustainability, environmental sustainability is one of the 6. We run annual training. So all of our staff, our thousands of employees all are empowered to A, understand, and B, think about how they can be more efficient in their use. 'Cause I can do so much from my role, but if we've got all of our engineers and all of our customer service people and everyone thinking about being a little bit more efficient or using a little bit less electricity or a more efficient model, then that impact could be much greater.

So, Nikki, when you say stakeholders, you're not just talking about cell phone users. Who are you talking about?

For us, it's, I mean, the government is one, regulators, but also our partners. So we partner with a lot of government agencies. So sort of that broader piece around stakeholders.

And the consumer.

And yeah, consumer. So business and we've got millions of customers, business and consumer customers. So yeah, it's that broader—

God, that sounds like a full-time job.

It really is. I mean, it's tiring.

Nikki, when you think about AI, what's the use case or example from One New Zealand or just more broadly that you're really inspired by? What's working in applying AI for sustainability?

That one I gave at the start around the microsleeps is kind of the most obvious one. But then you also think about sort of ethical behavior, how we think about how we're— we've got this thing called a trust trail. So every time before we put any application, AI application in use, we think about how do you take people through or take our people through a trust trail to make sure that we are doing things in the most sort of responsible and ethical way. So there's so many great examples, but I think what it comes down to is really thinking about at the core, what is it that, what is most material to your business? So for us it's electricity, it's thinking about how do we deliver our services to a way, to consumers in a way that is—

Plus you have to import phones, right?

We do. And so this is a big part of our footprint and actually working with partners is a key part because, you know, for us our Scope 3 emissions, which is our supply chain, is the vast majority of our emissions. So yeah, you do have to work with partners on that and the people who make phones. Thankfully, Apple and Samsung are doing some really great things to make their phones a lower carbon footprint, make them more efficient. So that really helps.

Bill, what are you inspired by? You mentioned that AI could be used for energy load management. What else do you think is potentially a step change for us all in addressing climate?

As an engineer, we use a lot of things, cars, transportation systems, stuff, and there's a tremendous amount of energy and real estate used for spare parts. I can imagine artificial intelligence systems just creating all the spare parts you want each time you need one.

Wow.

And, well, I mean, and I think the 3D printing people out there, of course, where have you been?

Yeah.

But it would— that technology's going to get better and better. more and more accessible to more and more people. And so I think that'll be a way to do more with less, especially in the transportation sector. You know, the US is 50% of our energy in transport. What is it in New Zealand? It's probably—

Transport system.

If you count coming to and fro.

Oh, I think we're at like 60% or 70%. It's quite high in terms of permits.

So you all, every time you want to go anywhere, oh yeah, it's 12 hours on a plane.

It's a very long way to everywhere.

You mean 12 minutes? No, 12 hours. And so by making the need to store old stuff less important, I think we'll be able to do a lot. And then the main thing with artificial intelligence for me is the efficient use of the electricity we do make. That seems like a huge opportunity, moving it around continents, where you need it. The sun comes up in the east. What? Yes. And moves across, then you move the electricity back and forth using very sophisticated sensing systems. But boy, the fear is that once somebody hacks that, everything goes to heck very fast. Man, oh man.

Yeah, that's why trusted AI is so important. You'll hear all week long about, about that topic. Well, the time has sped by. I would love for you to close out. What should this room do to support a more sustainable future and ensure sustainable AI?

Nikki? Well, other than the vote and be conscious of where you work, um, I would say just that concept of 1% better every day or encouraging your people to think about how do you just do everything a little bit better because that compounds over time.

Uh, make the world more efficient.

And safe.

There. Thank you. No, but I really think about it as I was in Waymos today, big fun. Woo, Waymo, yeah. But one of the interesting things when you're in the backseat of these self-driving cars is that's different from, not you, of course, not you, but people I've met, the cars obey all the rules.

Wow.

And so if you guys could make systems that obeyed all the rules inherently, you would be well received and you would change the world for the better. Get her done.

I love that. I've never heard that as a call to action. You're incredible. Thank you, Nicky. Thank you, Bill.

Thanks, you guys. Thanks for having us. Thank you so much. Good fun. Carry on!

Woo!