When Raj Singh, managing partner in JLL Spark Global Venturesis, of course, a technology enthusiast, he also has a cool analytical detachment experienced proptech investors.
As such, Singh has observed and shared the industry’s excitement about robotics, although he remains skeptical about how quickly the innovation will become a common feature of the built world.
Last week, Commercial Observer spoke to Singh about robotics in real estate, China’s use of industrial robots, and why JLL Spark hasn’t taken the plunge into investing in robotics companies.
This interview has been edited for length and clarity.
Commercial Observer: How robotics is affecting real estate today?
Raj Singh: If we ignore specific use cases in things like distribution centers and warehouses, the answer is almost nothing. Robotics is still relatively early and in the pilot phase, trying to figure out what it can do, with the possibility of sub-exceptions for cleaning.
You can see some cleaning robots in buildings that have a lot of floor space – hospitals, airports or whatever – where the robots can easily move around and clean. This is probably the only use case that has significant deployment at this point.
So what does robotics offer in the future? What has changed?
Context is important here. The macro context is that the real estate industry cannot find the people it needs to do what it wants to do in construction. We know there are nearly half a million unfilled jobs in the United States on construction sites. Then, inside the building, think about facility management. This is a relatively low paying job. There is high turnover and absenteeism in the job.
So, we have a problem, which is that we don’t have the manpower we need. Robotics appears to be a potential solution to that. That’s why we’re all interested in robotics.
There is also an interest in robotics because potentially using robots in these environments may be more cost-effective than using people, people who cannot be found today. I don’t know when robots will become mainstream, but the feeling is that they will. The problem is that, up until now, robots have been a one-trick pony: You have a certain robot; able to do certain things; and that thing is not a project. Humans do jobs, and in a job there are many different tasks. Robots tend to be able to perform one task, such as plastering a wall, or creating a layout, or cleaning the floor. No robot can perform the many tasks that humans can.
What has changed is the advent of physical AI, which is essentially taking a robot and creating some form of AI brain, allowing the robot to perform various tasks. That now makes it attractive.
Let’s take a facility management setup. I may want people or projects to include restocking inventory, cleaning the kitchen, or resetting the meeting room to be ready for the next meeting. Until now, I would never have been able to do that with a robot. With physical AI, I could potentially do that.
But the reason why we are still in the pilot phase is that for the robot that we use, in order for it to be activated, there are several things that need to happen. The equivalent of a large language model (LLM) that the robot will use must be created. This model is different from the LLM, because the LLM is trained mostly on the written word, and the internet is a huge source of the written word. There is no such thing as a database of physical actions, what a robot needs to do, so we have to create one.
Sometimes this is called a world model. There are various methodologies such as vision, action learning and visual action-language models that allow robots to see how a task is performed and then be able to perform that task and also add that logic to other tasks. All this is happening now.
One more thing that still needs to happen – and when it happens is not clear – is agility. Then we will be able to use robots for the things we want them to do. Change is happening very, very quickly. But it’s a fool’s game to say exactly when it will be in place.
We hear a lot about how advanced robotics is in China and how it is being used throughout the manufacturing chain. Are the robots still doing one job at a time, or are they doing something they can’t do?
Industrial robotics is different from what I’m talking about here. Indeed, China is ahead of everyone else in the deployment of industrial robots. But that’s not what we’re talking about here. I’m also right that China has advanced in terms of the type of robots that we would consider using in these cases.
However, to the best of my knowledge, they have not solved the problems just discussed in terms of world models and dexterity. But I would say that the most advanced robotics companies are Chinese companies. Partly because it has become a national priority there, so it is funded like that, and partly because there is a large local Chinese ecosystem that allows the company to grow very quickly because all the relevant technologies and parts are available and concentrated in a small area.
So we don’t have that level of sophistication, but we’re definitely trying to get there. I think it is an exaggeration to say that they are out there doing these sorts of tasks. He didn’t. But are they leading in terms of solution development? They can be.
Taking a broad view of the industry, can it be applied to the construction industry – which is huge – or is the sector still unable to meet the needs of the supply of robots?
There is a disconnect. Industrial robots cannot be taken off the factory floor and placed on a construction site. You need to create a robot suitable for a construction site. The factory floor is a predictable and inhumane place. Construction sites tend to be outside in varying weather conditions with floors that are not always smooth, and people.
So, for construction, we are still looking for a robot that can do many things at the same time. A lot of companies I’ve talked to are experimenting with robots, but even the Chinese robots they’re experimenting with aren’t ready for prime time. As such, there is an opportunity for US or European robotics companies to counter these results without having to face overwhelming competition from China.
Then, if we want to go into more detail, the business model that robotics companies are thinking of may have problems because, from my perspective, I want to make a bigger profit by using robots. If robotics companies want to take all the margins for themselves, then I think using robots doesn’t necessarily help.
Given all of these challenges, are there any robotics companies that JLL Spark is looking at or has invested in?
We have not invested in a robotics company until now. We are looking at a robot company, but I’m not sure that we are enough experts in robots to choose a winner. I wonder if the ecosystem is ready for us to use these companies.
My interest in investing is investing in software and services that surround robots, because another reason robotics isn’t ready is that the technology stack inside the building doesn’t know what robots are. So when your building management system releases a work order, a human can easily take the work order, know what to do and do it. The robot does not know what the work order is doing.
We need some way to translate the stack of technologies and outputs that exist today, such that a robot or multiple robots of different types can be correct and solve the problem.
We also need software to be able to coordinate or manage the use of different types of robots in the context of people who are also there. I have no illusions that we will replace much of the workforce anytime soon with the use of robots. It will be a mixed environment where there are people doing some things, robots doing some things, and situations where people and robots have to work together. To do this, we have to work hard to come up with a solution.
Philip Russo can be reached at prusso@commercialobserver.com.