Next in Tech | Ep. 269: Smart buildings and AI

View Full Transcript

Eric Hanselman   00:00:00 

Welcome to Next in Tech, an S&P Global Podcast with the world of Emerging Tech lives. I'm your host, Eric Hanselman, Chief Analyst for Industry Research at S&P Global. And today, we're talking about a quiet revolution in technology in smart buildings. With me to discuss it is the analyst leading the research, Zöe Roth. Zoe, welcome back to the podcast. 
 

Zöe Roth    00:00:23 

Thank you. It's always great to be back. 
 

Eric Hanselman   00:00:25 

It's great to have you here. We've just come off a discussion of industrial control systems and everything that's happening around the AI aspects of this. But you've been covering smart buildings and a lot of the technology that's been going into that for ages. And there's some research that you've been working on that has shed some interesting insights about where this is and the state of what's happening in smart buildings today. Can you tell us a little bit about it? 
 

Zöe Roth   00:00:50 

Yes, of course. So I'm actually just coming off of a series where I got to talk to 4 CEOs from building controls vendors. So picking their brains about what's on their road map, how they're targeting different verticals, what their drivers are looking into 2026. So I'm excited to talk about some of those trends and also some of the evolution that is going to happen from an energy technology perspective, but also on the interesting topic of autonomous buildings. 
 

Eric Hanselman   00:01:16 

Well, so you talked about -- you mentioned the energy aspect of this. This has been one of the core pieces of building automation for ages, which is helping to go manage energy consumption. And it is something that we are entering a day and an age in which those energy focus points are even more critical. How are the ways in which buildings are being managed changing? And what are the different aspects that are coming into play as we start to do more sophisticated control and better capabilities around that? 
 

Zöe Roth   00:01:47 

I think the biggest area of change is the investment that these vendors are making into their digital controls platforms. So you can think of Honeywell Forge, Siemens Building X, Schneider EcoStruxure Buildings and Johnson Controls OpenBlue. So in the past, where they might have sold a building management system or an HVAC system once every 20 or 30 years, now there's these applications that you can use for building management, but also energy efficiency, occupancy, lighting. So it's really kind of the software-defined building. 
 

Eric Hanselman   00:02:18 

Well, a lot that's been going on in terms of the capabilities that are there. But you mentioned autonomy. What are we doing to get towards greater levels of autonomy? And I guess, how far are some of these systems looking to go? And what are the implications? 
 

Zöe Roth   00:02:31 

Right. It's an interesting topic. And when we think about the average building, it is not quite ready for autonomy yet. It's an interesting -- the possibilities for what an autonomous building could do are quite interesting from an energy efficiency perspective, but also from an occupant experience or a physical safety and security perspective. I think the most important thing is there is this framework you have to start. The first step is getting visibility, connecting to your assets, censoring your data, censoring these legacy hardware systems. From there, you can pull in the data and it's such a point solution game now in buildings. Even within the vendors that I talked about, those are just a few of the dozens, if not hundreds of vendors that have some sort of hardware in a smart building. So the goal of these autonomous buildings is eventually that they would be able to connect to not only their own proprietary data, but all of the data in the building in order to detect anomalies in order to potentially bring in utility rate data and say, maybe we can come off the grid right now and we should draw into battery backups if that's a possibility so that we can save money on our utility bill. A lot of it is around energy efficiencies and energy savings, especially in these mission-critical verticals like data centers, hospitals, health care environments. 
 

Eric Hanselman   00:03:43 

It seems like more sophisticated building environments like those kinds of environments have got much stronger motivations. You're typically talking about big building campuses where the potential savings are significant. And especially as we start thinking about interaction with the grid, it sounds like we're starting to get towards the same kind of situations that we've been in with data centers, which is that they go from just being plugging into the grid to draw power to now actually being a more active participant. Are you seeing that kind of transition in buildings as well? 
 

Zöe Roth   00:04:18 

Yes. I think we are moving in that direction. And to your point, it is driven by these mission-critical facilities that don't really have a choice to go down. I mean no building wants to go down, but when you have SLAs or you're serving patients in a health care perspective, you definitely want to have that resilience baked in, and that now looks like potentially having battery energy storage systems, participating in demand response. So you have your legacy infrastructure in the building, your building management system, your energy management system, but now you're thinking of how am I going to connect that to potentially a battery energy storage system, integrating distributed energy resources. And it's a very regulatorily complex in each state, even at the state level, the country level regulation, and it's a huge investment, honestly. So it's not like you're just going to do it for -- you need to have -- be able to make that justification for the ROI, and then that's where the software comes in to be able to bridge the gap in all of these systems, some that you might have and some that you might need to be making investments in order to be able to participate in these virtual power plants or in these demand response programs. 
 

Eric Hanselman   00:05:21 

As you start thinking about more localized grid dynamics and a lot of the power side equations, these start to become a much bigger part of that conversation. I guess we're not yet at a point where we're consuming power at data center scale, there aren't yet any gigawatt hospital campuses, but maybe we're getting there closer to it. 
 

Zöe Roth   00:05:41 

Yes, exactly. And I think the good thing about whatever happens in data centers, eventually, whatever what we can solve there from an energy technology perspective will eventually cascade into mission-critical facilities, but also a commercial real estate building. what we're able to solve from -- it's mostly a convergence thing from a mechanical, electrical system perspective. If you're able to bring those together mostly with software, then eventually all buildings will hopefully be able to benefit from the same innovation. 
 

Eric Hanselman   00:06:07 

Well, this is -- I was thinking about some of the data center side management capabilities, the data center information management, the DSIM market. It seems though that's a lot simpler typically, granted you're operating at very high scale, but the kind of systems you're managing are probably something where the controls got built in or the various odds and ends and monitoring systems get built in when the data center got built. With your average building, you're coming into an environment, as you're pointing out, that there are a lot of disparate systems doing a lot of different things. And it's not just making sure the lights and the cage come on when the tech turns up. these are much more complex systems. And the kind of problems that need to be tackled around that, it seems are much more complex. 
 

Zöe Roth   00:06:50 

Yes. That fragmentation is a huge issue. And it's nice in data center that the data center industry has something like DSIM that provides this single point of visibility into the IT or to the white space, whereas other industries don't necessarily have some equivalent of DSIM. So DSIM becomes that natural IT/OT convergence point for the building management system and for the EPMS, -- but in other industries, that hasn't been parsed out as clearly as it is for DSIM. 
 

Eric Hanselman   00:07:18 

So what do you see as being the catalyst for a lot of this happening now? It seems like there -- this has been bubbling along in the background for a while, but there are some significant changes that are taking place. So what do you see as driving that? 
 

Zöe Roth   00:07:31 

Yes. I think from the conversations I have had with some of these vendors and also the research I've done is honestly data center. What's happening in data center is this rising tide that lifts all boats. So as these controls vendors look to build their capabilities and they've been selling building management systems and energy management systems into the data center vertical for dozens of years. But now they're thinking about how can we create more value, how can we potentially look at some things like liquid cooling. In the past, you would sell an HVAC, air cooled data center. Now they're making these investments like Johnson Controls is making investments into liquid cooling. Some of the other vendors, Schneider acquired Motivair for liquid cooling. So they're thinking of how they're going to bring their businesses into this next generation and data centers is just such a hot market for them from a thermal management perspective historically. So now they're thinking how can we get deeper, what else can we provide in this market? 
 

Eric Hanselman   00:08:22 

So this really is an adjacency in terms of what they've been looking at and some of the capabilities they have now translating into this more complex space, although none of the vendors that you're identifying have been strangers to the building management world either. 
 

Zöe Roth   00:08:36 

Yes. Yes, that's right. And I think while a lot of the conversation has been around data center, I think there's a lot of opportunity around some of these other verticals like hospitality. Honeywell has made a big push around hospitality. Siemens is big in health care and campuses. Johnson Controls is looking at some of these industrial buildings and Schneider is so native to data centers. But beyond that, looking at these other mission-critical, these campus-like environments that are just consuming so much energy that it makes sense to rethink how they're going to be designed, especially now that they have all these greenfield opportunities, like if you wanted to build a data center from scratch with all this new technology, you can do that. 
 

Eric Hanselman   00:09:12 

And you've got the technology at hand and you can put it to work in these applications as well. So talking about autonomy, one of the things that, of course, is the question around autonomy is workforce integration. And it seems like there are 2 sides to that coin. I'm sure building management is also one of those areas that is facing the same kinds of workforce transition issues that we've seen. And yet on the other side of it, full autonomy as we were talking about in the last step can be a little challenging to achieve. I take it one is the driver? And what do you see is the other side of that equation in terms of how far autonomy can go to manage that? 
 

Zöe Roth   00:09:50 

Right. It's a very nuanced conversation. And I know when Ian Hughes is here, we talk about this concept of the silver tsunami, where people in manufacturing who had been working in the floor of the plant for 20 years are retiring and they have all this knowledge in their brain and now they leave and you're like, wait, that's the only guy who knew how that elevator worked or how that thing works. So there is that bringing it together, the human plus AI is obviously the best possible case scenario, and there are real workforce and labor shortages. And when you're thinking of these new systems that nobody has used before, obviously, training is a huge part of that. A lot of the controls vendors have made big investments into upskilling technicians to manage and train people on how you're going to use this new cloud-based building management system or something that you've not used before. But on the other hand, with all of these baked-in agent-like capabilities and you're having a conversation with your building now, that obviously opens up people that previously weren't trained on some of these skills like HVAC skills to go in and have a conversation what's going on with this chiller, what's going on with this valve, predictive maintenance, remote access and management, so you can access your system from home and you don't have to roll a truck to go see what's going on in your building. So it is a double-sided coin like you're saying, but I do think there are a lot of helpful applications that these vendors are thinking about, and they're not this full autonomous building, it's like set point changes, schedule-based automations, things that look to augment this workforce challenges that buildings are having. 
 

Eric Hanselman   00:11:15 

So how do you see that balance in terms of really working out what are the capabilities that are able to run autonomously. And I guess it's like with any autonomous system, the real challenge is how do you build trust in the building managers and operators and administrators to be able to get to that point where they actually are comfortable handing off some of the automation of this capability and moving that forward? 
 

Zöe Roth   00:11:39 

Yes, that's the million-dollar question. That's what everybody wants to find out because on one side, you have the controls vendor saying we can develop, we can deploy an autonomous building and the owners and operators are not so sure. So I think that when we think of what some of these vendors have done, like Siemens, for example, looking at this human-centered autonomous building, that's how they're going to market with this, and they're really putting a human in the loop to look over and be involved in these conversations. I think as far as some of the use cases for what people are willing to give up or give the provision access or the ability for autonomy, it's things like set point changes. So if you have a room that's getting hot or too cold, automatically identifying that with a benchmark, things that the impact of them going wrong, like a room being too hot or being too cold is not the same as the impact of a data center going down because the chiller, there's the levels to it. And I think that a lot of them are these very -- it's like boring, like these boring autonomous use cases, that's where the value is, honestly. 
 

Eric Hanselman   00:12:41 

No, it makes sense. And I guess when you start thinking about some of that transition, I guess the real question is how much information can you get? And how much can you understand -- and we've talked in the smart building world forever about integrating with calendars and knowing that the simple things like Saturday and Sunday, you're keeping the building, you're not going as far in terms of heating or cooling because your expectation is there are fewer people there, which works great until there's some Saturday kickoff or something that's going on for project meetings or something. I guess the big question, a lot of that is how much information can you bring together? How do you integrate it? And the other question, of course, on the AI side is where does that information live? And now we start getting into those sort of tricky questions about data, the telemetry that's generating all the data and where does that go and how does it get managed? 
 

Zöe Roth   00:13:36 

Yes. I mean with this question, it truly depends on the vertical. So when you're asking me, I'm thinking of a commercial real estate, a commercial office. So these digital buildings platforms, they're aiming to be the single pane of glass that can aggregate the data from all of these different hardware systems and point solutions. But there is -- most of these buildings are mixed legacy hardware estate. So you're not just going to have a system from 1, you're going to have 2 or 3. So then it creates an opportunity for these next-generation cloud-based platforms like Code Labs is an example in this space, where they can act as this intelligence layer, but also act as this cloud BMS without any hardware. So there's -- it's created an opportunity, all this fragmentation and all these point solutions because nobody wants to look at 12 dashboards and people say they can integrate and they can be open, but it does get to a point where you actually need that to happen from an architectural perspective, and it's created opportunity not just for the controls vendors and their digital platforms, but for this next generation of start-ups that has identified this is a wonky process right now, and there's a market to solve it. So... 
 

Eric Hanselman   00:14:36 

Are these systems that generally are -- have a footprint on premises for some of these capabilities? -- cloud-based, the challenge, of course, you got to make sure that you ensure connectivity. And you can do that with backup connectivity. Presumably, these are high data demand kinds of applications. I'm curious about some of the approaches and some of the design ideas about how much lives on-prem, how much is cloud side? 
 

Zöe Roth   00:15:00 

So we've definitely seen a rising adoption of cloud-based building management systems. I would say something like a commercial real estate office or a university is more likely to have a cloud-based BMS, whereas in a data center, that's not likely to have either the digital platform or the cloud BMS just because there are concerns. One way to get around that is some of these edge analytics. So Honeywell does this, for example, in data centers, where they have an edge algorithm that can detect smoke from lithium-ion batteries. Like it's -- these individual applications that are very domain specific, they're very small and they can run on the edge, and there's not that concern about where is all this data going, -- all of a sudden, I've given access to my entire building operation, it's online, it's on the cloud. That becomes a concern, especially in data centers and with the legacy vulnerabilities that we've seen in some building management and building automation systems. So there are workarounds, but the architecture totally depends on what vertical it's being deployed in. 
 

Eric Hanselman   00:15:55 

You make an important point there, which is things like urgent control, life safety, those sorts of things, you probably want an on-prem capability, if nothing else, just because of the speed of response and those kinds of concerns and characteristics. But it seems like that we're heading towards something that sort of is balancing what happens on-prem, what happens beyond. And certainly being able to have the analytics of a cloud-based system, if you had multiple buildings, if you're managing a campus, those sorts of things, when you get to scale, that sounds like that might be something that there would be benefits on the cloud side. 
 

Zöe Roth   00:16:32 

Yes, absolutely. And this even -- it's bring up a conversation that I think is not talked about much in the industry of small and medium buildings. In a lot of cases, it doesn't make sense for them to have an entire building management system because it's too expensive. They're not going to have the technician. If you think about QSR or multisite retail, multisite food chains, it doesn't make sense for them to have their own building management system. But the corporate headquarters is going to want to know like portfolio level analytics about what's going on. So the cloud-based BMS makes a ton of sense in that case because it's this small packaged on-site and all the vendors have a capability in this, and they're starting to target this market as well as some of the start-ups in the space because it's been a lot of focus on the enterprise huge buildings, but there's -- not every building is going to be this massive energy-generating site. In most cases, a lot of these buildings are smaller and medium and there is untapped potential to get insight into what's going on in those buildings as well. 
 

Eric Hanselman   00:17:29 

Well, I could imagine a building fleet, if you're small retail, those kinds of environments where you actually want to -- you probably want to track building capabilities and performance relative to retail sales. If we keep the building cooler, does it generate more sales? If you keep it warmer, does it generate more sales? -- those sorts of things. Those seem to be a big part of it. But it also sounds like there are opportunities for -- in those sort of small and medium environments for there to be service capabilities that start to integrate this for building managers and building managed services that start to wrap around it. 
 

Zöe Roth   00:18:03 

Yes, yes, absolutely. And Schneider and Honeywell have both made investments into this small, medium slice of the market. So it just goes to show it's quite an interesting space they're in because they're targeting all vertical industries and you can develop something for data center. And at some point, it cascades to mission-critical, but you're also targeting small and medium. There's just a world of innovation and really prioritizing where you want to place your chips as far as what are we going to do next as these huge controls vendors who have the world to solve pretty much. 
 

Eric Hanselman   00:18:33 

Well, I mean that's always that challenge about what you're going to target. Do you go for the broad market, very high volume, but in small increments? Or do you try and pick the larger lower volume or higher size, lower volume kinds of approaches in terms of where it sits -- in building environments, you got a little bit of everything. What are the things about autonomy, of course, and all of the autonomous control pieces are the issues about security and some of what are the, I think, the friction points that we've seen historically and some of the transition between approaches that are more IT focused, again, from the AI side and others that are more operational technology or OT focused, which are that you put in place, it doesn't change forever and ever and ever, and some of the friction that you get in differences in pace for those environments. Of course, you've got to be ready. You were talking about being able to get to remote access and things like that, a lot of those convenience management pieces, those start to open pathways for risk on the cybersecurity side. It sounds like these are things that are going to have to balance carefully in these applications as well. 
 

Zöe Roth   00:19:38 

Yes, yes, absolutely. And I will say cybersecurity is a big focus for most of these digital platforms as well, whether the vendors have made acquisitions or it's a big part of their messaging. But as you're saying, oftentimes, your OT security posture is not necessarily on par with your IT security posture. And we've seen building systems be the victim of cyber attacks because they are such a high-value target. If you can take down a building management system, you can stop operations. There's been ransoms associated with that. So it is -- I think the industry is moving from this post-breach to pre-breach. Let's not let this happen. Let's maybe be a little bit more proactive. And that's why there's these concerns about the cloud-based systems to begin with because they're afraid of what this could open up from a security threat perspective. 
 

Eric Hanselman   00:20:22 

Well, it's that question that we deal with so often, which is where are the better places to secure your environment? Because a cloud environment, if you're amassing large amounts of data can be easier to secure than what you've got on-prem. -- if you do it right. And we've seen this wave of OT-related cyber compromise over the last couple of years. Last year seemed to be the year of water systems being compromised. Attackers are going to go places where they see high value, especially now that we live in the age of ransomware, the ability to go ransom a particular operational system starts to become pretty high value. But a lot of this got driven originally from remote access. It seems like for some of these systems, if you can build in secured remote access as part of the platform, -- now you start to eliminate some of the big security risks, which were somebody hooked up a VPN connection or a worse yet dial-in connection somewhere the buildings to be able to go get to them, at least a formalized approach so that you've got something that's going to manage access such you got role-based access control. All these other good things can help you track who's in the building and when those can be advantages, again, as long as they get done right. 
 

Zöe Roth   00:21:37 

And you've identified that's a whole separate market now because, again, these native hardware vendors have the capability to do that, and they are doing that for their own systems. But when you think of the building itself as a whole with 12 different point solutions, all with different vendors and all with different software, there's vendors now that are coming to address that problem of you want remote access for all of your applications. You don't just want it for this one BMS. So there's vendors like Neve that have emerged to solve this challenge of you have the data aggregation challenge anyway to run the building and optimize the building. But from a security perspective, you need that security baked in with each of the vendors on top of their own native security that they've assured you that they have, you want to take it to the next level and potentially work with somebody else to get secure remote access and some of these other use cases that might not be baked into each of the point solutions that you're using. 
 

Eric Hanselman   00:22:28 

And being able to have that kind of overlay capability so that you not only can provide things like secure access, but you can also provide a layer of protection around systems that may not have as much as you might otherwise desire, especially that's the problem we see in OT environments in that we've got a system that's in place. There are risk concerns about upgrading it or updating it and being able to wrap it in a more secure wrapper has a lot of attractiveness. So what should building operators and I guess a lot of the folks that are focused on this space really be thinking about in terms of directions and the kinds of things they can leverage? 
 

Zöe Roth   00:23:06 

I think there's 2 main areas that I'm excited about from a research perspective. One is what we talked about earlier with the energy technology advancements. So when we're thinking about some of these new demand response programs, participation in microgrids, virtual power plants, I think that will allow buildings to move from being just active or energy consumers to these prosumers where they're producing, they're consuming. Of course, that brings in the whole utility grid. It becomes a larger issue than just the building. But I think thinking about this energy ecosystem as not just the building, but it's also your connection with the utility, your connection with the data center. This is all getting tied together now more than ever with all the grid constraints. So the more you can participate and produce your own energy, the more you're able to have resilience in this time where there's grid outages and shortages, you can just be -- you have that assurance in yourself based on the investments you've made in these technologies. And I think the other area is autonomy. With the caveat, I think it's very interesting and cool what could be possible mostly from an energy, again, perspective, if you're able to optimize using utility data and drawing from your own backups and doing real occupancy integration. That's been like the dream is that you can say, normally on Wednesdays, we have a lot of people on this floor, but today, we don't. So there's that schedule-based automation, but it's actually still learning. There are these things like a building assistant that's actually able to reduce the energy usage by 20% in one building if every building has something like that would actually have a huge impact. 
 

Eric Hanselman   00:24:35 

I'm pulling in other data sources. You know it's going to be hot tomorrow, so it's holding down overnight in the inverse. 
 

Zöe Roth   00:24:43 

Exactly. 
 

Eric Hanselman   00:24:44 

Well, this has been great. So there are so many different aspects, and it's really interesting to see some of the parallels that we see across all of this. And of course, you said the one word that is important in a lot of the operational pieces, resilience in all this, which more and more of the discussions we're having around capabilities are all getting back to that point of ensuring that you build in resilience in the fundamental technology approaches that you're taking at this. 
 

Zöe Roth   00:25:09 

Yes, absolutely. I'm curious to see how it all shakes out. I'm glad to see that data centers are shaking it up and people are doing a lot of new interesting things in data centers. But again, that's just one end market for buildings. There's tons of buildings that aren't data centers that can benefit from these evolutions and from these innovations. So I'm excited to see how it scales beyond, of course, in data centers, but beyond that as well. 
 

Eric Hanselman   00:25:29 

The benefits of a lot of this, the cutting -- what's cutting edge actually gets out into the broader market. 
 

Zöe Roth   00:25:34 

Yes, totally. 
 

Eric Hanselman   00:25:35 

Well, this has been great. But we are at time for this episode. Thanks for being here, and thanks for all of your insights. 
 

Zöe Roth   00:25:41 

Yes. Thank you for having me. 
 

Eric Hanselman   00:25:43 

And that is it for this episode of Next in Tech. Thanks to our audience for staying with us, and thanks to our production team, including Sophie Carr, Feranmi Adeoshun, and Kyra Smith with Marketing and Events teams. If you liked this episode, please subscribe and like this.  

I hope you'll join us for our next episode because there is always something next in tech.