Episode 76: Feeding the Matrix: JoAnn Garbin
The complete transcript for episode 76.

Molly Wood Voice-Over: Welcome to Everybody in the Pool, the podcast where we dive deep into the innovative solutions and the brilliant minds who are tackling the climate crisis head-on. I'm Molly Wood.
This week it’s the last OFFICIAL episode in our series Feeding the Matrix diving deep into AI and energy use with this special focus on data centers and how to make them more efficient less damaging to the environment and the places around them
Last week we talked about innovating on sources of energy and HOW data centers could be constructed alongside modular fusion energy reactors to make them clean-burning islands of computing power.
This week is the one I’ve been waiting for we’re going to talk about what would happen if you COMPLETELY redesigned a data center from the outside in with the goal of making it welcome in any community a positive contributor to the world around it.
That is an initiative that’s happening already and as it happens, it started at Microsoft.
JoAnn
I am JoAnn Garbin, hello, and I don't have a title and it's lovely. I have a couple roles, which is co-founder of Regenerous Labs, which is totally separate conversation, maybe. And then I'm also the co-author of the forthcoming book, The Insider's Guide to Innovation at Microsoft.
Molly Wood
Amazing. Tell us about, let's sort of get your backstory. You have this fascinating career that has combined sustainability and engineering and leadership. How did you get to where you are?
JoAnn
Zigging and zagging, but not really. Wiggling around a core path.
Molly Wood
So the best way.
JoAnn
That is something that I am grateful that I've done and also have done my best to be intentional about sticking to my values. So I am a very value driven, service oriented, do good, know, do well by doing good person. And I've been fortunate enough to
always find or make an opportunity to work in that space. And I've never had to compromise on that to pay the bills or, you know, just get through, since leaving college. So I've been, I've got 25 years under my belt of engineering and sustainability and entrepreneurism, largely because I never had the shit hit the fan.
and make me do something else.
Molly Wood
Honestly, we should all be so lucky. think that's pretty, it's a good path. All right, so you're here today. You are capping off this series on AI and data centers, AI and energy use, and this is the part of the conversation where we have turned to the ways that we can innovate so that we don't make mistakes that we don't need to make as this technology sort of takes over the globe. And so you are the perfect person to be this capstone because for two reasons, one, you were in charge of regenerative data center design at Microsoft for a little over four years, and you have a book out about innovation. We're gonna take those in the two parts of the show. Let's start with that role, regenerative data centers. That feels particularly relevant now. Tell us what that means.
JoAnn
Yeah, so the regenerative data center of the future. I was brought into Microsoft right before the pandemic actually. And our mandate was just figure out what the data center needs to be in the future. So we got what it is today. Obviously context change, environment changing, politics change, economies change. So we were trying to look out five, 10, 15 years and scenario plan, what might that future look like? And in that, those scenarios, what would the data center need to be? And now I was a little bit, you know, primed to find a regenerative sustainability path forward. came in having already explored biomimicry on my own and had 20 something years of sustainable advisory and innovation work. But I legit tried not to sway the crowd. I started with a whiteboard and we did the intentional discovery work that we talk about in the book of just stepping back from the problem and saying, what are all the variables? What are all the knowns?
What have you learned over the last 15 years? And honestly, when we looked at all of the business challenges, technical challenges, and people challenges, both employees, stakeholders, shareholders, but also people that have to live in the midst of data centers, it led us to the regenerative data center of the future. There was no, what we would call chip,
design of a data center that was going to solve all of those problems and capitalize on all the opportunities. So what we did was we flipped it and we designed from the outside in and it just opened up a world of possibilities and once we had decided that this was our goal and we created the you know the vision and the
JoAnn
the guiding principles, what does it mean to be regenerative? You know, it's that restore, renew, revitalize for all of the system interactions that you have. And that's with nature, with people, with economy. Once we flip that script, so many of the ways we would have solved problems before were no longer relevant because they just didn't live up to our own principles.
And so we would come up with the, we could do this technical thing. no, like that doesn't fix the other 12 things. It actually breaks those. And we weren't going to do a trade-off. We weren't going to try to turn Microsoft into a philanthropy, like, build a regenerative data centers because it's good for people and planet. It had to be good for Microsoft too.
But we also know that the way, and I say this as Joanne, not Microsoft, the way data centers are designed today from the chip out isn't good for people and planet. So we have a lot of opportunity.
Molly Wood
So yeah, what were the problems? Were the problems like the government of the Netherlands might kick us out and say you can't build data centers here anymore because there's too much water and we don't like them and they aren't real jobs. you know, what were the
JoAnn
Yeah, it wasn't even a might. It was a reality. mean, there were Singapore at that point in time had put a moratorium on new data centers. The Netherlands, which is a key hub in the world, just started blocking permits and making it run from six months to get a permit to two years to get a permit. Facebook and Google, I think both pulled out of the region and it was like, that was not...
anymore a one-off that was becoming the norm, what we call NIMBY, right? So in the start of the data center world, was, hey, this is amazing economic development, come build your data center. That ship sailed. It was like, oh wait, when you build a data center, you clog our grid, you take our water, you put this blight on our landscape. Like these are actual words coming out of newspaper articles.
Molly Wood
Right.
JoAnn
Like when we did social listening about the sentiments being expressed, communities were enraged. They're like, you create a couple dozen jobs, but you take up, you cut down our forest and you put this ugly thing up with a big metal fence around it and you go behind that fence and we never see you again. And so over the years, more efforts were made for community development.
There's some really great programming that's been developed for as far as skilling and connectivity and things like that. But nothing at that point, the infrastructure itself had never been looked at as how does that provide benefit. So again, we didn't start with that problem statement, but as we explored permitting delays, land scarcity, community backlash, competition. Just pure like, hey, Facebook, Google, and Microsoft show up in the same town at the same time because the demand signals there. And now they're saying, which one are we going to build? And now it, you know, I think a lot of,
Molly Wood
Yep.
JoAnn
The industry grew a long way serving the engineers that care about the wiring and the nuts and the bolts and, you know, all the little nitty-gritty things.
People outside the walls of the data center in general don't care. They want performance and then they want all the other stuff they can see, right? Like performance, price and experience. And so we really, that's, we go into this in the book, Xbox has a great framework for it called BXT, business experience technology. For years, the T has been leading the data center world.
Molly Wood
Right.
JoAnn
All we did in our analysis was say, what happens when we let the experience lead with a good business model behind it and technology making it happen? So you put technology in the place that belongs, which is at the support role in the problem solving. That led us to a sustainable regenerative solution. wasn't, let's figure out how to make it regenerative. And that's because people
Molly Wood
Yeah.
JoAnn
People value it, right? Like, I don't want to see you cut down my forest. I want the economic development of a data center, but I want my forest too. How to make that possible, that's the innovation.
Molly Wood
Right, totally.
Molly Wood
And then it seems like this would be a good place to do a little bit of a definition about regenerative too. So something that doesn't, not only does not take excess resource, but potentially gives back.
JoAnn
It, I always caveat it by saying I am not hippie dippy and woo woo. I'm not going to start my, my speeches with all let's, you know, breathe deeply for five minutes. I love those people. They're my friends. I am a business person and an engineer. And I do get a little like with that stuff in the workplace.
Molly Wood
you
JoAnn
But it is the definition of regenerative is switching your mind where sustainable or efficient, even renewable is do less harm, hold the line, net zero. Regenerative is net positive. It's aiming for abundance. It's giving back to the systems you rely on. And the simplest framework I've found
to describe it is feedback loops. Just create reinforcing feedback loops, right? If you're in a place and you need water to do what you do.
Understand the water system, the water cycle, and make sure you're contributing to the health and revitalization of the water cycle, which leads you down really interesting paths. For instance, everybody would, I would bet that most people walking around would say the best data centers use no water. That's not necessarily true. If you're in a place whose
micro water cycle, the small water cycle has been disrupted by land use, know, poor land use and forests being cut down and everything. And evaporative cooling data center that puts water into the local atmosphere can revitalize that small water cycle. And some much more savvy water people than us showed us how
If you're disrupting at the coastline, you could cause a drought 20 miles, 50 miles inland. Or you can say, okay, we're going to be a part of this and we're not going to do no water here. We're going to do evaporative here because evaporative will help restore the ecosystems around us. It's counterintuitive.
Molly Wood
Right.
But it also sounds like it's very local. Back to your point about community, as you're designing, sounds like it's a... I could imagine some corporations being allergic to the idea that there isn't just a template for a data center. There is in fact an individual, a personalized design process, if you will, that has to take place here.
JoAnn
Well, we had to design for that too. And so one of the principles that we had four guiding principles and one of them was to be globally applicable, locally desirable because we can't afford to build snowflake data centers. It cannot be designed from scratch every time. So we just pushed the envelope of what can be held constant.
and what can change, and how do we create that kit of parts. And the data center engineers doing the current designs use that phrase, kit of parts. But their kit of parts is very small. And all we did was say, let's make that bigger and incorporate more aspects of the interactions so that you have a kit of parts for.
the ecology and biodiversity of an area. have a kit of parts for the economy of the area. You have a kit of parts for, you know, the specific place in the world, what we call the nature of place, which is everything that exists there. And then yes, you do have to expand your design process to pull in more Lego bricks, but it's still Lego bricks. And then
we designed trade-offs in other ways, can stuff be manufactured and drop shipped instead of stick built? Like it's not, there wasn't novelty in those discoveries. It's, know, the pieces and parts themselves aren't novel. It's how do you put the whole end to end process together to expedite and scale your operation of
developing data centers while being locally desirable and beneficial.
Molly Wood
Right. I'm going to try not to geek out too much, but I just can't help it. What are the parts of data center design that need solving? Like the resource use is obvious, and it sounds like it's water and power. There's the construction. There's the aesthetics. Walk us through the different chunks of this that have to be solved when you imagine doing this totally differently.
JoAnn
That's cool, I'm a geek.
JoAnn
Yeah, it is from the outside in. It's what systems is it plugging into?
Are you creating bi-directional flow in those systems or one direction, right? So let's take green energy. Amazing work creating more green energy supply in the grid, right? And there aren't many island mode data centers right now. And then at least what I knew of Microsoft before I left.
Molly Wood
Mm-hmm.
JoAnn
That was not the game plan. The grid is still the most reliable way to power anything.
Molly Wood
Island mode being that there's some sort of power generation situated next to the data center. The dream.
JoAnn
Yeah, and you're off, you're not connected into the grid and you're just doing your own thing. Which is an age old way to do things. A lot of critical operations have the ability to switch to island mode and then switch back to grid. If you don't want to lose time and money. But for data centers, and they do that in a sense with their backup generators, right? But it's not meant to be long term.
Molly Wood
Mm-hmm.
JoAnn
So one of the things if you look through the social listening, so this is all public information, you'll hear a common complaint is you bought the land, you said this is going to be green energy powered, great. You partnered with a wind farm, awesome. They built a wind farm in our town. You built a data center in our town. And now all of the energy from the wind farm goes to your data center. And we're sitting here with no more green energy.
but we're looking at a wind farm and a data center. So the aesthetics are in there, but really the problem in that respect is you're taking the energy and it dies with you. You use it and then you exhaust it as heat. And only in a few places in the world is it being turned into high quality heat. That's hot. The conditions are appropriate to reuse the heat.
It's almost always in the Nordics where they have district heating and it's cold out and there's degree of difference.
What happens when you hold that I'm going to build a wind farm, I'm going to build a data center, but that green energy needs to come in to my data center, be used to compute, and then passed on into the community to be reused again, no matter where I am in the world. All of a sudden, the very affordable air cooling that creates low quality heat that can only be used in the Nordics.
That's not an option anymore. It pushes you to use liquid cooling because liquid cooling with the, the, just the heat transfer dynamics and everything, you can get a higher quality heat. can reuse it more readily. And now you can go into Atlanta and you can co-locate next to an airport or a hospital or a condominium association.
JoAnn
probably all of those things, because the scale of the data center is so large, that you have enough heat for laundry services and heating and hot water and sterilization and beer making and you name it. But you have to be intentional to think about the throughput instead of consumption, its use and reuse. And you get as much.
Molly Wood
Mm-hmm.
JoAnn
You know, it's all starting with the sun, not to, again, get hippy-dippy about it. But every time that sunlight comes down and we collect it and we lose some of it because of efficiency losses and then we transmit it and we lose. And if we can use and reuse that sunlight as much as possible before it's just, you know, peters out, that is the most sustainable answer.
Molly Wood
Right, totally.
JoAnn
And now you're not having that conversation with the town where they're already got their backs up saying, eh, you know, we get these two big ugly infrastructure things and we get nothing for it. A couple dozen jobs. It gives a more visible, tangible, integrated experience and business case to the places that you put these things.
Molly Wood Voice-Over: Time for a quick break. When we come back, more on how communities themselves are actually in the driver’s seat when it comes to demanding better data center design and how the principles of innovation may be the key to solving the climate crisis.
Molly Wood Voice-Over: Welcome back to Everybody in the Pool. We’re talking with Joann Garbin, formerly of Microsoft, now the author of a new book about innovation at Microsoft and let’s get back to how the NEED for better built data centers was obvious in that company for a while.
Molly Wood
Mm-hmm.
Molly Wood
Totally. And I want to put a fine point on the fact that Microsoft, at least, had started this conversation with some urgency before the pandemic and before this, you know, the advent of generative AI and the kind of boom in data center construction that we suspect we're going to see, or at least resource consumption that we're going to see as a result of this. And I would imagine...
it points out that this was already a thing. And I would imagine it's an even bigger thing now, by which I sort of mean to ask you, like, this is probably still a business imperative, right? I think that's all of our fear is that everybody's gonna sort of throw out their sustainability goals because the economic drive is there and the returns seem to be there in generative AI. And so it sounds like maybe you're still hopeful that companies will stick to these principles if only because they will have to. They will not be allowed to build new data centers willy-nilly.
JoAnn
I think two things, at least, on this front.
Molly Wood
That's okay. There were like 11 questions in there. you know.
JoAnn
Yeah. So one, think it's up to us, the population of the places. Put your foot down. The business imperative isn't to build a data center. The business imperative is to bring cloud services and AI to certain places in the world. If you are a place they want to be, put your foot down and get what you want.
That's, and trust me, the people on the other side are just looking for that demand signal. Like these are not bad evil people like, I want to do, no, but if they don't have a clear cut demand signal to take the risk of changing a design they know works, very hard path for anyone involved to change direction.
But if they're like the moratorium in Singapore, that was a clear signal and Singapore is a super valuable market. And they just snapped a line and said, no, you want to come here, here's what you got to do.
Molly Wood
Yeah.
JoAnn
We don't need national rules for that. Every municipality has the power to do that. That's the interesting thing. In the north part of the Netherlands where there's this concentration, this hub, the farmers put up a fight. They're like, this is fertile ground we're giving up for industrial infrastructure. We're one of the bread baskets of the world. What are we doing? And they just...
snapped a line and they made it really difficult. But it may change. So that's our part in it. It's hold them to what you want.
The other part, if I can remember it, is, yes, this is my main point.
JoAnn
I would say to all of the executive leadership of these companies, if you are treating sustainability like a line item cost and after the fact compliance or PR correction, yeah, there's no return on investment on that. That is an endless spend.
If you shift it left and make it one of the many lenses by which you're exploring your problems and looking for solutions and then you design with it for innovation, it has a direct return on investment. And there's just the philosophical simple notion that what we're doing today is not sustainable.
So by definition, sustainability is novel. And you're going to have a competitive advantage.
Molly Wood
It's... That's sort of the perfect segue then to, mean, what we do on the show is try to reframe the climate conversation to solutions by way of innovation and ingenuity. And the thing that humans actually, we are awesome at destruction. We are also awesome at innovation and ingenuity. And you have this book out, The Insider's Guide to Innovation at Microsoft.
Molly Wood
Yeah, yes. So give us a high level what is in the book and then how does it relate to this? Like, it's thinking, you know, this change in thinking that that keeps you ahead of the game as a business.
JoAnn
Yeah. So the book is a collaboration of many people. We interviewed over 40 different people throughout the history of Microsoft. So current and past. And we wanted to hear the stories behind some of the biggest breakthroughs the company has had. And yes, Dean and I picked Microsoft because we work there.
Molly Wood
Dean, your coworker, give us this, or co-author rather.
JoAnn
Yes, Dean Carrigan is my co-author and he is in the responsible AI part of Microsoft. He works for Eric Horowitz, who wrote our forward. He's the chief scientific officer of the company. Dean has good friends. His network, he's been in the company 20 something years. His network was absolutely invaluable to be able to write a book like this.
Because what we wanted to do since we were Microsoft and they're Microsoft, we want to write a book about Microsoft in the sense that with Microsoft turning 50 this year and having had only three CEOs, it gave us this really deep and broad canvas to explore for innovation. We could find different kinds of businesses and different methodologies and
A lot of innovation books kind of say, well, we learned this from this company and this from that company and this, and then you kind of have to try to cross-pollinate. We wanted to remove that necessity. We're like, this is a prime opportunity to just see if there's some real fundamental truths of innovation that aren't so attached to who is the company and what do they do. Right? Cause we were looking to create.
Molly Wood
Yeah.
JoAnn
really cheat sheet. Get good at innovation foundational principles and practices. because I'll be candid. I've had a couple of experiences over the last decade where I spent more time trying to get people to collaborate effectively.
And you're spending all that time and energy on good collaboration and creativity practices. And then you're too tired to actually do the work. So we're like, you have to, especially for big complicated things, right? Nobody can know everything.
Molly Wood
That's why I don't like to work in teams. No, I'm just kidding. That's not, that's the wrong attitude.
JoAnn
And so that was the, you know, all of the motivations for doing the book was Dean and I being practitioners of innovation with this access, with this great case, you know, library of case studies. And so the first half of the book is case studies, but told as narratives where like not everybody wants to read a white paper. So they're told like stories. We, we did our best Walter Isaacson. We did not come close, but we tried and Then the second half, lot, again, we looked at, we're both book nerds. We looked at a lot of innovation books and we're like, a lot of books tell you the stories and kind of just give you the conclusion to synthesize it or leave you to find the patterns. And we're like, no, we're going to do the work. Our whole point is to make this easy for people. So the second half of the book is the analysis to find those patterns and threads. And so we broke it down into how do you innovate? Every day, over the years, with everyone, and beyond technology. If you stack those up, you've got the foundations of everything you need to be innovative as an individual and an organization. You're going to learn and grow and add to it, but it's going to get you started in a solid direction so that you don't spend the next five years figuring out how to innovate.
And then in the middle of the book, because Dean's responsible AI and on sustainability, we wrote a chapter on responsible innovation and it bridges between the two halves of the book. And, you we, we cover privacy, accessibility, security, responsible AI and sustainability, and we focus on holistic sustainability. So not just environmental sustainability, although I lean heavy in that that's where most of my experiences, but
You can't disconnect people and other things from ecology. So that's what the book is. by the end of it, we believe if you're reading it and underlining it and trying things out with your teams and experimenting with it, you can get good at, what do I do when I have an idea? Most people are like, fail fast. No.
JoAnn
Learn. Failing is one way to learn. But there are a whole bunch of other ways to learn that are much cheaper and quicker. Step back from your problem. What did we do with the data center of the future program? We didn't say, how do we make the data center regenerative? We were like, what does the data center of the future need to be? Who does it serve? How many of them will there be? Where will they be? You know, what will the universe look like at that point in time? And you get a whole bunch of scenarios. So
Molly Wood
Right.
JoAnn
There's just a bunch of work that happens before you land on an idea that's worth taking forward. And then there's all these things you gotta do once you have the idea.
Molly Wood
All right. Do you believe that this is a framework that can be successfully applied to solving the climate crisis? Yeah.
JoAnn
Yeah. I think that I'm very confident about that because it's what I've been doing for 25 years.
Molly Wood
You've been out of Microsoft for a while. What, if anything, can you tell us about the status of the regenerative data center project?
JoAnn
Yeah, I left in October of 24, but I've been offline for a little while longer. It is an aspirational goal still. I've heard them, numerous people say that. They use it as the big picture. And I know that they've invested in biomimicry. So...
One of my teammates was, we found her in another part of the company and she's one of the first people to get a biomimicry certification, like professional cert. And we were like, come. And she was one of the first directors of biomimicry in any company. I think she was like the second and the only one in big tech. Well, she now has a couple more biomimics with her and they're over in our land group.
Molly Wood
Fascinating.
JoAnn
So they're positioned well to have impact. I would say from my perspective, it's, you know, I'm an entrepreneur. That's the core of who I am. Nothing moves fast enough in corporate. It's just a big machine, right? It's a very big boat to turn. But there are people working to turn the boat.
Molly Wood
Yep. Yeah, it'll take all those signals it sounds like.
JoAnn
And like I said, if the communities hold that line, you're just gonna make their job easier inside to get the work done. So we have a role to play.
Molly Wood
Exactly.Yes. And now you know, like some version of these plans exist communities. although I will, I'll say it so JoAnn doesn't. exactly.
JoAnn
Yeah, they're out there. Like the pictures, I can't share anything that they haven't put in the public domain. So the pictures that are on my LinkedIn profile, those are out there. Use your imagination.
Molly Wood
I cannot wait to hear more about that as it evolves. Joanne Garbin, thank you so much. The book is an insider's guide to innovation at Microsoft, but also everybody just follow this lady. She's amazing. Yes.
JoAnn
And just final sentiment, all proceeds from the book are being donated to support STEAM education charities. So creating that feedback loop, trying to help whoever reads it and help the next generation of change makers.
Molly Wood
Go get it. Go get it everyone. JoAnn, thank you so much. What a treat.
JoAnn
Thanks, Molly. Thanks for having me.
Molly Wood Voice-over: Full disclosure that I’d like to do a whole OTHER episode in the future about what Joann is doing with her new venture, Regenerous Labs, but I’ll save it for now.
That's it for this episode of Everybody in the Pool. Thank you so much for listening.
And that’s it for Feeding the Matrix although as you might imagine this topic isn’t going anywhere in fact next week, I’m talking with the former CTO of Facebook who founded a climate tech investment firm and you’d better believe I got his thoughts on AI and the energy race, as well
Please let me know what you thought of this topic and these interviews and anything else that’s on your mind as we all try to navigate innovation solutions emotions activism find me at in at everybody in the pool
Subscribe to the newsletter and find all the latest episodes at everybody in the pool dot com
And remember together, we can get this done.
See you next week.