AI usage is booming, but activity isn't the same as value. Dr. Serena Huang talks with Dave Easton, Partner in the Growth Equity team at Generation Investment Management, about what separates real AI ROI from hype.
AI can transform any business, but only if there's enough power to run it. Dr. Serena Huang sits down with Dave Easton, Partner in the Growth Equity team at Generation Investment Management and WEKA board member, to explore what investors now look for before betting on an AI company. They dig into how to measure AI ROI that your customers actually feel, why culture and the right incentives matter more than any metric, and how to get the most intelligence out of your energy use.
Timestamps:
0:29 - Meet Dave Easton of Generation Investment
5:00 - What investors ask now vs. 18 months ago
7:30 - Spotting pilots vs. real deployments
9:38 - The simplest ROI wins
12:39 - Why AI activity is not the only efficiency metric
17:12 - Creating a winning culture
19:46 - How companies are approaching sustainability in AI
29:26 - The next 18 months of AI investing
33:24 - Dave’s two magic AI metrics
Links:
[00:00:00] Dave Easton: Am I doing everything I can to get the most intelligence out of the kilowatt hours that I'm drawing? How much power I'm pulling off the grid, am I getting the most out of that? Because we only have a limited number to go around. The demand for intelligence seems to be unlimited. The supply of kilowatts is very much limited.
So yeah, we've got to be doing everything we can at all layers.
[00:00:28] Serena Huang: Dave, welcome to Deep Geeks. It's so great to have you today.
[00:00:32] Dave Easton: Thank you for having me.
[00:00:34] Serena Huang: Let's talk about your work at Generation Investment. It's at such a unique intersection of growth and what you mentioned, sustainability. And as you have conversations with these AI companies, what patterns are emerging for you?
What are the strongest companies you see doing differently in the portfolio?
[00:00:55] Dave Easton: Yeah, that's a great question. So maybe just to quickly, um, back up. You know, Generation's an investment firm that was founded, you know, a couple of decades ago, really on the thesis that, um, if you understand, uh, long-term sustainability trends, and that could be, um, the energy transition and how do we decarbonize our grids, it could be access to affordable healthcare, it could be access to financial inclusion.
If you understand those trends really, really deeply, not only will you help to make the world a better place, but you'll find better ways to, um, invest and make more money, um, for your, for your clients. And so I think that's, that's, uh, sort of fundamentally our, our, our philosophy and approach. In terms of, you know, how that's working out in our portfolio, I think we see two sides of AI, really.
One is where AI can be used to fundamentally address real-world problems. Mm. Take healthcare, for instance. We have a number of healthcare companies, and you can now see a path where through AI triage, through AI note-taking- Right ... automated follow-up, um, of appointments using agents, you can get to the point where you could, you know, have zero clinician time spent on administration and all of it spent on clinical care.
Mm. That's, to me, like hugely transformative in a world where we have an aging population and healthcare becoming increasingly expensive. So, and we're seeing this right now, one of my portfolio companies, AlayaCare, is deploying agents right across the home care, home health agencies that they work with, that really is freeing up huge amounts of clinician time and improving, um, access to care.
So on the one side we see just an incredible set of opportunities around how AI can, as I say, transform healthcare, financial inclusion or energy. Um- On the other side though, we also see both opportunities and challenges in dealing with some of the risks of AI, right? So, so we are absolutely focused on, you know, how do you make sure that AI is energy efficient?
And that's maybe something we'll come back to, but one of the reasons we invested in WEKA really was around how do you make, um, energy, um, uh, you incre- improve the, uh, tokens per kilowatt hour ultimately. Uh, and there's a bunch of other risks around governance and privacy and compliance that we think also are a huge opportunity.
So we- we- we're bullish on how AI can, um, solve humanity's challenges, and we're also think there's a huge role for people like us in trying to make sure that that is done in a sustainable way.
[00:03:38] Serena Huang: Yeah. The AI note-taking has freed up clinicians' time, and I've seen that even in my recent visit with a doctor, and he asked me, he, he got my permission.
He said, "Is it okay if I use AI to take notes?" And I said, "Funny you should ask. I work in AI, so please, please do use it." Um, but something I do pay attention to is I'm watching whether or not AI makes them actually focus on the patient more, 'cause that's what it's supposed to do, right? As opposed to being even more distracted and multitask more and not focus on the patient.
Um, so I- I've had a few friends who are very skeptical and anxious about the privacy aspect, and I tell them, you know, of course you can always opt out and, um, but pay attention to whether or not they are focusing more on you as a patient 'cause they should be able to as a result of AI. Th-
[00:04:30] Dave Easton: that's right. I mean, I think in healthcare in particular, I think the, th- the promise of AI is to have technology that allows the clinicians to be more human so that the more administrative or kind of data input, data extract can be done by the machines so that the humans can do what humans can do, which is care for other people.
And I think that's, like, a hugely important design principle when I think about AI, how AI will be applied to healthcare.
[00:05:02] Serena Huang: I remember personally sitting through s- a few startup pitches two years ago, and when people said, "We're building AI," that was enough for the whole deck almost. A lot has changed since then.
Where, where do you see investors go these days? What are they asking now that they weren't 18 months or even 24 months ago?
[00:05:25] Dave Easton: Yeah. I think, look, nearly all of our portfolio companies, um, have always been deploying some form of AI or machine learning and, and most of them have built, you know, basic versions into their products.
I think what we've seen over the last 18 to 24 months is really bu- people building the kind of harnesses and scaffolding for, like, true agentic solutions. So things that can not just auto summarize or not just do a simple optimization algorithm, but can really do sustained work. And so I think now, like, you know, these agents are now live in the field working with customers, and so the question I think moves from, you know, can you go and build an agent-like platform that works in demos, to can you now deploy it in customers a- and manage the change management of that, right?
Because AI as a bolt-on to the way people do things currently is not realizing the full potential of AI, you know, in the same way that, you know, this is probably an overused analogy, but, um, just putting electricity into a factory that was run on steam power before didn't change the factory. You needed to change how it worked.
Similarly, if you're gonna deploy agents into your customers, they're gonna have to change some of the ways that they work, and that I think requires change management. So I think now we're in the age of not just can you build them, but can you deploy them, and can you help your customers get the value out of them that they're looking for?
That's really, to me, the question is not can I build and sell the agents. It's like what's the real world impact hav- happening, happening on the back end? And I think that's what, uh, investors really are looking for now.
[00:07:10] Serena Huang: There's such a gap, right? I, I think a few years ago we see a lot of, we hear a lot of what's more pilots, and I remember in our prep call you mentioned companies moving from, um, experimenting to deploying.
Um, as you look at your portfolio companies now, what tells you that they've actually moved from pilot to real deployment?
[00:07:37] Dave Easton: And ultimately this is about what the customers say, right? What I say is relatively irrelevant. It's what those customers say. It's are you making those customers' lives better? You know, you take an example, you know, again, from a, you know, from say the, the energy sector now.
So are you, you know, one of my companies, um, Volue's deploying, um, uh, AI to improve the, uh, function of control rooms in power sec- in, in, in the power sector, so making them more efficient, um, and doing optimizations. Uh, the question really becomes like does it work? Like does... Do your customers see the value?
Do your customers think, "Wow, this is something that I wanna get more of"? And so, so, so, so to that extent, it's really about ultimate customer satisfaction and their own ability to measure hard ROI. I, I think if you're selling into innovation budgets, you know, still, I think that's gonna be, you know, not where it's at.
It needs to be selling into real business leaders who, uh, really see this either reduces their risk or reduces their cost, or ideally, uh, grows their revenue by optimizing the business even further. So I, I, I, and I should say as a sustainability focused investor as well, you know, we're always asking the business ROI question because ultimately that's how our customers sell, that's how our companies sell to their customers and make money.
But we're also asking the impact question. So is it, you know, you know, is it improving the lives of clinicians? Is it im- improving medical outcomes? Is it reducing waste and loss on the grid? So these kinds of impact questions are hand in hand to us with customer ROI questions.
[00:09:18] Serena Huang: Yeah. I like that a lot. Um, and everyone is talking about how to get ROI on AI these days, and we, we sometimes think about in the AI world, ROI is not just return on investment, but return on intelligence.
I'm curious when you hear a company explain the value they are creating with AI, what convinces you that they are creating real business value, um, and not just AI activity? If you can give me a concrete example.
[00:09:51] Dave Easton: The simpler the ROI to explain, the realer it is. What I mean by that is I think if the, if they, if there's a complicated logic chain required to explain it to customers, it's probably not gonna be what lands ultimately, right?
So if you take, um, one of our portfolio companies, Octopus Energy, now, you know, the leading electricity retailer here in the UK, they've been deploying AI through their software platform, Kraken, um, for two, three years to improve the way in which customer service messages are handled. And they're the mess- it's, there really are two north star metrics, which have always been Octopus' north star metrics, which is number one- Does it improve the efficiency of their team?
And number two, does it improve the happiness of their customers? And if it does those two things, which they have proven that it does, then that's ROI. I think to me, that's really where the, where the kind of the best form of... It's where you can really boil it down to what does this business exist to do, and does this improve those metrics that matter most?
I think where it becomes harder to measure, obviously, is when you're talking about things like internal productivity around questions to do with speeding up R&D, right? Um, you know, clearly we've moved from lines of code as a metric to commits as a metric, and now we're moving towards, okay, like, but like features and business value features.
And so I, I do think there it's harder, but when you've got a, um, yeah, I say a company with a very clear mission, which is how do we like make our customers happier whilst being more efficient, if you hit both of those, then I think you know you're winning.
[00:11:44] Serena Huang: I love that. And I don't know if, um, if every company has truly moved away from, uh, um, from lines of code to commit.
I think we're still getting there, um, because I, I've talked to a lot of leaders recently, and they're all getting... trying to get help on how do we measure AI ROI, and I see most metrics are around AI activity, I'll call them, right? Whether it's lines of code or prompts or, uh, gosh, even tokens, right? And, and it's, um, yes, if you don't have any activity then, then you're not going to get ROI.
Fair. But, um, but the connection, as you said, to things like productivity and efficiency, innovation, um, that's, that's sometimes harder to, to measure. Um, I am curious if you, um, if you see this shift really happening or, um, is this just a lack I'm seeing in the enterprise that we're still thinking too much about activities and, and not able to connect to business outcomes quite yet?
[00:12:51] Dave Easton: The simpler the activity, the easier it is to measure, right? So when you're talking about relatively easy to measure activities where you can measure the quality of that activity, so say, I don't know, scheduling an appointment for a, for somebody and, like, whether that appointment was the right appointment, where you can get relatively fast feedback loops, I think that's clearer.
It's clear- clearly as you move towards more complicated multi-stage tasks where you can't measure a single outcome- That's obviously harder. I think there's gonna be an adoption curve, right, that says, look, at the start, until you let everyone try stuff, then you can't have any view of anything. So I think everyone's now at the point where, you know, they're letting their teams try and experiment and these kind of things.
I think then it's gonna be about a bunch of best practice, um, sharing. And then really I think honing in on what processes do we need to change, and ultimately, like, have we incentivized our people in the right way- Yeah ... to go and achieve the right outcome? I kind of think, and this is true, been true for engineering leadership forever, right?
It goes back to the book, "The Mythical Man-Month." If you try and sort of force, uh, overly bureaucratic measurement of, uh, you know, productivity within engineering organizations, you end up incentivizing the wrong things, because ultimately what you want to incentivize is business value. And so really to me, this is about are you giving people access to the right tools?
Are you helping them share best practices? Are you getting out of their way so that they can, um, use those tools efficiently as possible? And do you have the right people? 'Cause the right people always want to, you know, achieve the right outcomes. If you try and create too many false forms of measurement, I think you'll end up having people, um, grading to the test.
Whereas if you have incredibly empowered individuals who believe in the mission and just want to succeed, then I think that's gonna be a far better outcome than any kind of, um, measurement chart is gonna produce.
[00:14:58] Serena Huang: Yeah. I, I like what you said a lot. Um, there's a few things that stood out, so I'll just summarize for our, our listeners.
One is the focus on simplicity. I think the KISS principle is what I'm hearing, right? Um, when it comes to metrics, whether it's, um, for AI adoption or even the business outcomes and ROI, it has to be simple to explain, easy to explain, and easy to understand. Uh, um, and the other thing that I like a lot is your focus on customers.
Are customers happier, better off as a result of what you're doing? 'Cause if not, frankly nothing matters. So you could be doing a lot more with AI, but if the customer impact isn't there, it, it doesn't really matter. Um-
[00:15:48] Dave Easton: A- and I do think this is where if you have really clearly mission-driven, um-
[00:15:53] Serena Huang: Yes ...
[00:15:54] Dave Easton: uh, a really clearly mission-driven company that knows what change it wants to see in the world, whether people join for that mission and are incredibly motivated by that mission, that alignment I think means that you'll get way more out of these tools than if you are trying to force adoption.
Like to, I mean, to, like, we think winning cultures are an incredibly important part of it, and winning cultures aligned with strong missions, um, are always better. I think in a world of, uh, of AI enablement where your people can do more than they ever could before, that combination of mission and culture, we think is, is gonna be, you know, even more powerful than it was before.
[00:16:34] Serena Huang: Yeah. De- definitely. Um, and, and I think for, for engineering leaders who are listening, what you said really stood out as well around are we incentivizing the right behaviors and do we have the right people? Are we building the right culture? Um, um, I, I love the focus on humans because we, we talk so much about AI, but we know that humans need, you know, not only need to be in the loop, but we also sometimes worry what happens to us when AI gets more capable.
Um, and as far as building winning teams, I'm curious if you have some real examples or what kind of leadership trades or what practices have you seen that create the right incentives and right environment for that winning culture?
[00:17:23] Dave Easton: That's a great question. Um, and again, it's core to, um, it's core to how we, um, we think and assess companies.
Sometimes, uh, um, very often when we're doing diligence on, on companies, we'll have to ask to speak to chief people officers. Uh, and it's, it's amazing how rarely investors do that. It's always surprised me that very often people say, "You're the first person that's wanted to speak to our chief people officer."
Uh, you know, now chief people officers are incredibly important. Obviously, you also need a CEO that can lead from the front on culture, so I don't think the people function as a replacement for that. But to me, like it's just, it's an unbelievably important part of, um, due diligence. Is this sort of, is do you have companies that simultaneously have a winning culture, so people are turning up because they wanna win and because they believe that they can win, and have very, very high standards for themselves, whilst at the same time a sufficiently kind and mission-oriented culture that they are doing things for more than themselves.
And I think, like, that's what we look for in companies. You know, we-- I'm very, very blessed to be on the board of, uh, a number of companies where, um, we have that together. But that's the, that's the, that's the challenge, right? Sometimes there's a view that says, "Oh, because a company is trying to have an impact in the world or mission-driven, then, you know, maybe it needs to have a culture that is less sort of focused on winning, less, um, sort of ha- has a less high bar."
But from our perspective, no. Like, if you wanna change the world, if you wanna transform the lives of millions of people, if you wanna save gigawatts of power for the grid, if you wanna do all of these things, you need to have a higher bar for yourselves than, you know, if you were doing something less important.
[00:19:19] Serena Huang: Wow. Well, I hope our chief people officer listeners are, are lighting up as much as I am, uh, 'cause I'm going to send this episode to a few of them that I know would really appreciate this. Yes, to get to know your culture, talk to the chief people officer. Um, there's no one better who can tell you the truth of what's happening inside the organization.
Now, you've mentioned energy earlier and, um, sustainability, and that's what potentially hold companies back from making progress with AI and that sort of become a constraint. Um, from your history of investing in energy technology and software, where energy we're now seeing becoming the constraint that determines success, are you seeing companies actually planning for this?
Um, how are, how are companies actually approaching this?
[00:20:14] Dave Easton: Yeah. It's, it's a great question, and it's really core to how... I mean, it's, it's funny, I started as an energy technology investor. That's where I came to AI through the energy problem. And so this is probably six or seven years ago, talking to people who were thinking about the sizes of cluster that they might end up building.
Back then, this was for autonomous vehicles or computer vision or whatever, and seeing how much the power consumption was gonna increase Uh, and I think, uh, uh, if I'm a slightly naive view perhaps that said, well, wind and solar are now increasingly cheap, so the power problem is kind of solved. Which is wind and solar are incredibly cheap.
The, the reduction in the cost of renewables is one of the wonders of the world, but you do have to connect them all. And we don't have a grid yet that can move at the speed of renewables or at the speed of AI. And so we became, I think, increasingly concerned about the grid build-out as a constraint on AI, as I say, like six, seven years ago.
And I think when we were speaking to utilities back then, we would say things like, "Hey, how are you getting ready for this AI power wave that's coming?" Uh, and they would say, "We don't see anything yet." And so I think that's really what got us super interested, um, uh, back, you know, I'd say like 2020, 2021 around the energy problem, uh, of AI.
And I think there's really, um... The way we think about it is at two levels. One is how do we rapidly, but in a low carbon way, get more capacity on our grids? And we can talk about that. And secondly, how do we optimize the tokens per kilowatt hour? Um, ultimately, we've got a constrained number of kilowatt hours.
We've got a constrained number, amount of power we can put into this. So how do we get more intelligence out? And that was really, back then it was a thesis we called green data. Uh, and if you saw when we announced our, uh, investment in WEKA, it was all about green data, and it was all about this question of, hey, like- If you've got GPUs that are burning power but are not doing useful calculations because the data isn't fed in fast enough, then that's wasteful.
If you can feed more data in so that you massively improve the efficiency of your cluster so that, you know, in the kind of, in the experience of one WEKA customer we spoke to during diligence, you can take a training run down from two weeks to four hours, then you're optimizing your intelligence per kilowatt hour.
And similarly, you know, now obviously as we're moving into the world of inference and WEKA's augmented memory grid, seeing the same thing happen. You know, so for us, we just think this is an incredibly important North Star for anyone building large scale, um, uh, machine learning or AI applications. Am I doing everything I can to get the most intelligence out of the kilowatt hours that I'm drawing?
[00:23:31] Serena Huang: Wow. That's a very high bar, Dave. So I think the, the metric that I hear you say is intelligence per kilowatt. Is that right?
[00:23:41] Dave Easton: Intelligence per kilowatt hour, yes. So basically, for how much power I'm pulling off the grid, am I getting the most out of that? Because we only have a limited number to go around, and there's, the demand for intelligence is, seems to be unlimited.
The supply of kilowatts is very much limited. So, um, yeah, we've got to be doing everything we can at all layers, as I say, you know, what WEKA's doing in the augmented memory grid, what WEKA's doing in neural mesh, uh, and other solutions as well, in order to really, you know, get the maximum intelligence we can out of the resources that we're drawing.
[00:24:18] Serena Huang: Hmm. Such a good reminder, and, and I think a lot of companies are probably wondering, wow, where, where is that metric on our scorecard? It's probably not shown to the board on a regular basis, I bet. Um, but investors are going to start asking, 'cause we, we know there could be constraint on how much we can grow.
Um, how- what are your practical recommendations for companies that do want to optimize this metric if they haven't really thought about it and just kind of getting started, can you help them out?
[00:24:50] Dave Easton: Y- yeah. I mean, I think this... And this-- Look, this is, this is the, this is the flip side of tokenomics, right?
Which I know that obviously that many people have talked about. Um, I know Val's talked about extensively. You know, because in some, in many ways, like, you're maximizing, uh, resource efficiency is the same thing as maximizing return on, on cost. So it's many of the same things that Val, for instance, would've talked about, you know, um, in a bunch of places.
But ultimately it's about in economics there's the concept of the binding constraint, right? It's about looking through your stack and saying, "Okay,
[00:25:25] Serena Huang: you know,
[00:25:25] Dave Easton: I, I've-- How-- Do I have the best power efficiency in the, in the physics of my data center?" Well, increasingly people do, but then am I running storage that is optimally, um, uh, is optimally feeding data in?
Am I running networking that's efficient? How am I moving towards, um, photonics, which obviously will reduce, um, energy intensity quite significantly of networking? Um, how am I thinking about, you know, the use of, like, KV caching and other things to, like, one of our portfolio companies, Redis, has done a bunch of work to try and make sure that, you know, you only hit the model when you need to.
How do you do smart model routing to make sure that you're using the right model for the right task? It's really all of these elements, and I think as power becomes more and more and more the bottleneck, I think we are gonna be, have to be asking ourselves this, this question at all levels of the stack.
[00:26:21] Serena Huang: Yes. Um, thank you. That was very helpful and very practical as if you are just getting started, it's, um, those are great questions to think about. And, um, I- when I run into people who are very concerned about energy and they ask what can I personally do, um, it's that on the personal level, are we using the smallest model possible for, for the task?
And, uh, it's a great, uh, I think parallel to what you just said, uh, on the company level, but we can also individually try to minimize, um, energy consumption to, to some extent, too.
[00:26:58] Dave Easton: Yeah, that's right. But I do think the onus really is on companies deploying AI in order to be the people to, to do it, because ultimately...
As again, like, you know- I, I, I think that we are very bullish on the transformative effect that AI can have. So we are definitely not people who say, "Ooh, try and reduce your usage," or whatever. Like, no, there's like a, i- i- intelligence can help solve the world's problems, so we should have more of it. But what we gotta do is we gotta do that within the physical constraints, and that is just about not being wasteful.
It's about looking through, as in, you know, looking through everything that we're doing when we're architecting our, our, our clusters to make sure that we're doing the most. And this is, this is, this is not gonna be an option, because if you simply look at the The demand for data centers coming on the grid in places like PJM in, in the US, a-a-and the amount of power we can build and interconnect, like it's, it's very clear we're gonna be at a power constraint.
We already are at a power constraint. You know, if you look at the capacity auctions run by PJM, they're, you know, they've massively increased the p- the price of capacity, power capacity in PJM. We're getting to like levels that are sort of, that clearly show that, um, demand outstrips supply. So I don't think companies are gonna have a choice.
Like, I think comp- consumers are gonna demand intelligence in a variety of ways. The grids are not gonna be able to build it as fast as they can, and so everyone's gonna need to come back to saying, "Okay, how do I get the absolute maximum efficiency out of the systems that I'm building?" Um, because otherwise they're gonna lose to their competitors who are adopting more efficient technologies and being able to get more out of the, um, the, the kilowatts that they have.
[00:28:47] Serena Huang: Yeah. Excellent reminder. Um, yes, and I like the focus as well on the, the, the responsibility is on the companies to make the right decisions and, um, and make sure we can get the most intelligence 'cause that, that demand will continue to go, go up. Um, all right. Well, let's, um, let's talk about the future for a m- minute here.
Um, as you think about the next 18 months, what do you think the AI investment landscape will look like?
[00:29:20] Dave Easton: I mean, the really hard thing about being as an in- an investor is that you have to try and predict the future. Uh, and it turns out predicting the future is very, very hard. Our companies are generally some of the most advanced companies in terms of adopting AI, because those are the kind of founders that we back.
Those are the kinds of, um, companies that we... So I think probably most of our companies are in the top 1% for like how much they're using AI to improve their internal processes. That will, I think, cascade across companies, big, small, large enterprises. I think we will see more and more companies in the next 18 to 24 months deploying AI in R&D, in customer support, in the ways that our companies already are.
So I think you're gonna see a mass growth in the use of AI inside the enterprise for a whole series of things, both for internal efficiency and for building better products. I think you, you may see from some of our companies, um, their spending on AI for internal use, you start to plateau as they start to think about, okay, well, which are the right models to use?
When do I use open source? When do I not? Like, how do I... Like, learning the optimization. So it wouldn't surprise me if spend continued to go up from here, but it wouldn't in my companies, but it wouldn't surprise me if it was flat. Um, I don't think it'll be growing at the 20% month on month that it was growing a few months ago.
I, I, I don't think that's sustainable, but I think that more and more companies in main street, in mainstream corporate world will start to use AI in this way that our companies already are. So I think that's one thing, uh, um, I would say. And then the second thing is I, I think the power issue is gonna become front and center.
I think Uh, power shortages will be real. You can just see this in any form of the interconnection queue, and I think this is gonna for- force the big data center builders to get really, um, creative about how do they do things like fund demand response like Google just did with, um, uh, a company called Voltus, which is trying to help do, uh, demand response.
Um, or how do they help, uh, actually speed up the, um, interconnection process for improving the way in which we can bring new renewables and new batteries onto the grid. So I think that is gonna become, uh, a front and center issue because to be honest, if we don't solve the power problem, two things are gonna happen.
One is AI is gonna be constrained, and secondly, we risk losing our social license to operate, where local communities will say, "Hey, you can't just take the power and force the power price up. If, if you do that, we're just gonna resist." And you're seeing this obviously across the US right now with backlash against the new build of data centers.
So I, I, I think data center developers are gonna have to become grid developers and to help fund the upgrade of the grid that we need, or else there's gonna be, I think, a real social license to operate problem.
[00:32:33] Serena Huang: Yeah, absolutely. Um, find, find ways to give back to the community as well. We, um, we, we heard earlier in, uh, in our series that, um, data centers in other countries in Europe actually, um, sometimes they're able to, for example, convert the heat from data centers and share it back with the community so there's free heat for, um, everyone in, in the city.
And, and I think in, um, different parts of the world, because regulations are different, because incentives from the government are different, um, it's not always top of mind and the approaches are different. But I like the reminder a lot to, um, to think about how do we get ahead of, of that power constraint.
Well, if you could wave one magic wand, um, and, and there's one metric that you would like to see from AI companies, uh, to create long-term value, not just hype, what would you want to see? What is that one metric?
[00:33:40] Dave Easton: Can I have two?
[00:33:42] Serena Huang: Sure. You have to pick the primary, though.
[00:33:45] Dave Easton: Well, I think there's two things, right?
'Cause I think there's what you build and there's how you build it.
[00:33:49] Serena Huang: Ah,
[00:33:49] Dave Easton: okay. So I think in what you build, it's really about customer satisfaction, however you measure that. It's about are you improving the lives of the customers and having the, like... And achieving your mission by improving the lives. And, and that's gonna be a different metric if you're a healthcare company or if you're a energy optimization company.
But fundamentally, it's about are you helping your customers in a way that makes the world a better place and achieves your mission, and are they happy with that? That's the first one. And then in the what you build, it's intelligence per kilowatt hour. It's really that. It's really as simple, am I, with the resources that I have, am I doing the maximum I possibly can to get the most intelligence out of the fewest resources?
Hmm. I
[00:34:39] Serena Huang: like that. Okay. So customer satisfaction, and then intelligence efficiency? Can we call it that?
[00:34:48] Dave Easton: Yeah, sure. Sure.
[00:34:50] Serena Huang: Yeah. Okay. Um, thank you. Thank you. That's, that's very helpful. All right. Everyone who's listening, those are the two magic metrics you want, uh, go, going to your, your next meeting. And, and Dave, I, I think let's, um, let's end with some advice for tech founders who might be listening.
Whether they're building an app to coach employees or an app that help clinicians take notes, if they are listening to us today, what's one thing you think they should get right before going to a meeting with an investor?
[00:35:27] Dave Easton: I think ultimately it's about, um, w- why does your company exist, and like what's its higher purpose and higher mission?
Like, you cannot- From my perspective, there's no point building companies that don't impact the world in a positive way. Now, that doesn't mean they shouldn't make money. They have to make money. Like, you know, but you have to... You're not gonna inspire your team, you're not gonna inspire your customers, you're not gonna inspire yourself if there's not a reason that you're doing what you're doing, and that, that, that, that reason is both to, um, make money and create value for your shareholders and to, um, uh, have a huge impact on the world.
So for me, it's about being really clear about what's that driving force that, that, that is gonna drive you and your team to wanna work incredibly hard. And to me, that, that, that always comes best from mission-driven companies.
[00:36:23] Serena Huang: I, I couldn't agree more. I, I think the startup founders that I've seen who are really successful do have that really strong mission, and they're able to bring people along.
And when things get hard, they can get themselves out of bed of why am I still doing this? Why the grind?
[00:36:42] Dave Easton: Yeah, 'cause it's definitely gonna be hard. I mean, that's the only thing I think you can promise, uh, any entrepreneur is that it's gonna, you know- Yep ... there's gonna be moments of deep suffering. And so, like, in those moments of suffering, like, why are you doing what you're doing, and does it motivate you more than the pain?
And if it doesn't, there's no point.
[00:37:02] Serena Huang: Yes. I think that's it. Does it motivate you more than the pain? That's, that's the bar. All right. Well, thank you so much for all the insights you have shared today with Da- with us, Dave. Is there any last word that you want to, uh, leave with us?
[00:37:21] Dave Easton: No, I just wanted to say thank you.
Like, thank you to the whole WEKA team. It's been a, um, heck of a journey so far working with WEKA, and I think, like, look, there's a huge, huge, huge, um, wave of opportunity coming for everyone building, um, the right companies in the right way i- in the AI wave and, you know, um, being on the right side of history is hugely valuable.
And so yeah, I'm just, I'm excited to be on the journey.
[00:37:44] Serena Huang: Thank you. And I really appreciate the behind the scene view a little bit as a investor as well. We, we so rarely get this kind of opportunity, so really thank you so much for your time today, Dave.
[00:37:55] Dave Easton: Thank you.
[00:37:57] Serena Huang: Thanks for listening to "Deep Geeks." A huge thank you to my guest today, Dave Easton.
If today's episode make you think differently about how AI gets built or powered, share it with someone who needs to hear it. Find "Deep Geeks" on Spotify, YouTube, or wherever you get your podcasts. Until next time.