Interview: Getting The ‘Anywhere Edge’ With SEMPRE
Advancements in military connectivity and growing reliance on networked military hardware have highlighted the importance of ensuring that communications and cloud services remain secure on the modern battlefield, while extreme weather events like Hurricane Helene and cyberattacks on civilian network infrastructure in turn emphasize the need for resiliency in data networks and other critical national infrastructure.
SEMPRE says its “Anywhere Edge” lineup of products offering secure, portable local data networks and resilient communications can help solve these issues, with the company recently closing an $10 million funding round co-led by AE Ventures and 90 Degree North Holdings to help advance its goal of creating a more resilient digital landscape.
To better understand SEMPRE’s products and their use concept, Overt Defense had the opportunity to speak with SEMPRE CEO Rob Spalding.
This interview has been edited for clarity.
To start things off, could you introduce yourself, SEMPRE and the Anywhere Edge lineup?
I’m a former Air Force officer who flew B-2s, and spent a lot of time in China. My last job was as the senior director for strategic planning at the National Security Council at the White House, and the company and products really stem from my time at the White House.
I was looking at digital infrastructure as a key vulnerability for our society and also as an opportunity to solve a lot of the military’s challenges with interoperability. I really thought that 5G as a technology offered some interesting opportunities because it was transitioning from a mostly hardware based product to a mostly software based product, and so I really started to get deep into the technology while I was at the White House.
I wrote a report that got leaked in an Axios article. The report I wrote was about the opportunities and challenges that came with 5G as a technology, but it was portrayed as an attempt to nationalize the 5G network, when in reality it was used to explain to people what the opportunities and challenges were for 5G.
A lot of the coverage that subsequently came out focused more on the challenges, particularly as it pertained to a lot of the security challenges with companies like Huawei, but I was actually a lot more interested in the opportunities 5G presented.
And the opportunities were, I thought, really immense for solving resiliency, security and interoperability. So while I was at the White House, I was introduced to some technology that the Department of Defense was using in aircraft for radar processing, so we started looking at that. I was still in the Air Force then, so we looked into how it could be applied to telecommunications, specifically 5G. And so we looked at encrypting 5G links with that technology. That’s really where SEMPRE gets its start.
When I retired from the Air Force, I said, okay, why don’t we start a company that really focuses on resiliency and security? If we do that, we can solve the DOD’s interoperability problems and we can solve a lot of the problems like we have today with Hurricane Helene, where telecommunication systems go out very easily. The reason they go out is because they’re highly centralized, as the telecom networks have only a few data centers. If you lose access to the data center from the cell tower, or if that data center is taken down like what happened in Nashville in 2020, the network goes down. And we’ve seen that with Hurricane Helene, there were people that needed help, but didn’t have the ability to communicate that they needed help because of network outages.
The idea for the SEMPRE architecture was to take that same exact technology, but instead of centralizing it, we would decentralize it. We take the cellular network and the data centers that run in the cloud a lot of the applications that people use on their smartphones and other devices, like security algorithms on cameras. Let’s take that, let’s shrink it down into a single device and let’s harden that device. Harden it for things like EMP, harden it against tampering so you can detect if somebody physically messed with the device. Make sure it can operate at extreme temperatures, unlike how most data centers are temperature controlled places. Make it autonomous. Make it so that you don’t have to know anything to deploy it, you just plug it in and it works.
Ensure that it is self-healing and can deal with loss of connection, because oftentimes you aren’t going to have a connection back to the Internet, as whether it’s a satellite connection or fiber connection, those connections can be lost. We made it so it pretty much operates on its own, so it continues to deliver services locally even with intermittent connections to a larger network, you can still call people that are on the local radio network.
And then finally, what we’re concerned about from a cybersecurity perspective is the 5G data plane and the control plane often go on the same pathways. We physically separated the control plane, which gave us the ability to take it and put it into space. And because it’s not latency sensitive, it can be a very narrow band link. That’s the way that we designed the overall architecture of these nodes, to be able to come in and control them with a very narrow band link. If you lose that data connection, you can continue to be connected and use that control plane link.
But importantly, if you get hacked, if there’s a cyber attack like a ransomware attack, we can come in through the control plane and shut that down and bring back a clean system.
The devices that we are starting to introduce to the market are the SEMPRE-T and the SEMPRE-F. The SEMPRE-F, F for fixed, is an up to 50-foot tall, hurricane-proof tower.
The SEMPRE-T is the transportable one. Importantly, the T can be moved around by a single person like a wheelbarrow and set up within about five minutes, then you have your data center and you have your network up and operating. But it can also be deployed at the base of a tower or on a rooftop to really create that network and data center in any location, at any time, anywhere, including very hostile locations like the battlefield. If, for instance, you deploy it in the battlefield and you lose it or you get overrun, somebody is able to capture it and they try to open it up, the node can detect that and erase any encryption keys so they can’t get any of the data that’s inside.
How long has your product line been worked on? In addition to the SEMPRE-T, there’s also the F and the M, which I see is described as “coming soon”.
We have spent six years developing them. The SEMPRE-F has been operating pretty much continuously for the last three and a half years. Because the SEMPRE-T is more transportable, we don’t operate it continuously, but it’s been out for about two years. We’re on Generation 5 of the technology, and what we’ve been doing with each generation is basically integrating all of the different components and features that I previously talked about. More importantly, what we’ve been doing on the software side is really getting all of that software to run on smaller and smaller and smaller devices. We can run an entire 5G network, radio network and all that, on a single, actually half a server right now. That allows us to put a lot of processing power within the T and the F, letting you run AI applications or other things.
The last two years have really been focused on manufacturability and increasing the amount of network volume that’s devoted to applications by decreasing the amount of volume that the network and all the management systems take.
You’ve mentioned EMP hardening and resistance to extreme weather like hurricanes. Is there anything else you can share about how the SEMPRE products are hardened to handle other hostile environments like electronic warfare or physical damage from shrapnel, gunfire and the likes?
Obviously these things can be destroyed, they’re not as tough as a tank as it would be impractical from a deployment perspective. We have some capability to defend it from small arms fire, but if you’re going to drop a bomb on it, you’re going to blow it up. It’s not indestructible, but we believe that decentralized infrastructure provides the most unbreakable edge because it allows you to not have everything in a single location. A lot of the time, what military units have is multiples of a thing to deal with something getting lost or damaged – we have another one if that happens. Ultimately, if somebody’s really concerned about those types of physical attacks, then you would have more than one.
But there are other attacks like, for example, we were doing a demonstration out at Whiteman Air Force Base where somebody took a backhoe and dug up the fiber optic cables to the local cell towers and all the cell towers went out. That doesn’t affect us because we have the entire network in the node. If somebody digs up or cuts your fiber connection or breaks your satellite connection, you don’t stop working.
I think the biggest problem that you have with this type of infrastructure in terms of surviving a natural disaster is flooding. That is the one that really typically ends up getting everything. One of the benefits of shrinking the size, weight, and power of the node so that you can elevate it means that you can, and this is our intent for these nodes, get them up on the rooftops above the waterline. Our intent is really to have these powered by solar panels so that if you do lose grid power, they continue to function, they continue to provide services until the centralized network is restored.
We really emphasize the low size, weight, and power, the ability to run on renewable energy, the ability to be completely off grid, and the ability for it to deal with things like EMP, extreme temperatures, or extreme weather like winds. Rooftops are a great place because they tend to stay intact through a storm, it’s the trees and power lines that go down. If you can get this installed on a rooftop, it ends up being a pretty good place to have equipment. And then our goal is, if you do have a situation like that, your communications and your applications don’t go away. If you need to get services, you have the ability to do that. Those are the kinds of things that we anticipate in terms of how our products would be physically attacked or how they would get damaged through natural disasters.
When it comes to electronic warfare, the difference between what we are doing with the 5G radio and what, say, the Ericssons and Nokias of the world are doing is that we actually give access to that raw electromagnetic data coming off the radio, so you can run different applications on top of that radio architecture. So if you want to turn your 5G radio into a radar, if you want to be able to detect signals, these are all part of applications and features that will be coming in future models, as people develop software algorithms that will let you use the radio in that way, including to resist jamming. It’s really designed to be able to use radio frequencies in ways that a typical 5G commercial radio would not.
Speaking of the ability to handle damage, how much emphasis is there with current designs on modularity to allow the users to repair minor damage in the field?
It’s not really designed for repair in the field because it’s a sealed volume. It’s designed to protect the data inside, so unless somebody is authorized and they know what they’re doing, we would not see repairs in the field. It’s more likely that there will be spares of the device to replace a damaged one, and then the device would be brought back to service by repairing it at a repair facility.
I’ve previously spoken with the founder of an Ukrainian charity who told me about a place in Ukraine that has a “graveyard” of Starlink terminals cannibalized to repair damaged ones.
Starlink is an impressive system, but it is literally a link. Our systems are meant to have all your data and applications there locally. If somebody has access to that, that means they have access to your data, they have access to your applications. That would be a very, very valuable target from an adversarial perspective.
We don’t want the enemy to get that data, get those applications, so we made it essentially impossible for that thing to be opened up without proper authorization. The node will defend itself even if it’s powered off. It stores its encryption keys for up to 20 years and if you try to get into it without authorization, it will zeroize the keys. You can put it in a powered off state and let it sit, but if anybody tries to get into it in that powered off state, they’re not going to get any of the data. And that’s important because the most important thing to our customers is their data.
You mentioned how you have provided demonstrations of the Anywhere Edge lineup. How production ready are these products?
We’ve just started low rate initial production, and the first production units will come off the production line in January 2025.
Are there any particular products that will be delivered first under low rate initial production?
SEMPRE-T is the one that we see a lot of interest in. Because of the way we designed these, these are really the first EMP hardened, military grade commercial products that are affordable for commercial clients. These are not made to be super expensive and not only for the military.
One of the biggest segments that have expressed interest in these things is commercial real estate construction. They often need connectivity and the cloud to manage a project site. It’s very difficult to deploy infrastructure to support that, but for us, that’s really easy. We bring one of these nodes to a production site and that allows them to set up and manage their project quite easily at a price that’s very, very affordable.
We already have SEMPRE units deployed in the field, with Global Strike Command, which represents the ICBMs and the bombers of the United States Air Force. And you know, we intend to increase those deployments over time. We really thought that Global Strike Command would be the best first customer because they have the hardest requirements of any in the Department of Defense.
Can you share how many test units of the various Anywhere Edge products are already in the field?
There’s five SEMPRE-Ts and one SEMPRE-F deployed.
Are these all with Global Strike Command?
No, we have some partners that are defense prime contractors that have acquired the nodes, as they see the opportunity to bring their services and applications to their customers in a resilient and secure way by bundling that with the applications and services that they sell to their customers. So defense primes like Boeing are essentially taking these nodes, they’re loading their applications on the nodes and then presenting them to their customers for different solutions like maintenance, logistics, operations or planning.
Are they integrating the SEMPRE lineup into their own offerings, or as part of the support package?
An example of what we do here is a demonstration we did for Global Strike Command. Boeing has an AI algorithm that they’ve developed that basically takes drone footage of an aircraft, runs it through the algorithm, and that’s able to give you all the defects it detects with up to 94% accuracy. When an aircraft lands and you’re conducting a post flight inspection, that typically takes eight hours and around 50% accuracy and what that entails is a person walking around the aircraft, visually inspecting it and then writing down the defects they see on a notepad.
What we demonstrated is connecting the drone to our node using 5G, allowing it to see the aircraft, take all of that imagery into the node and then process that in real time using the AI algorithm. You get an improved understanding of the aircraft, it’s a faster, less risky inspection because you don’t have to have people on top of the aircraft walking around, and then it allows you to better manage the fleet. And this is not just for military aircraft, it’s also usable for commercial aircraft.
That’s just one example. The reason they can’t often provide these services in the field is because there are massive amounts of data that need to be collected on the aircraft, the links that they have to send the data back to the cloud are too small, and so they’re not able to realistically do this with the infrastructure that’s currently available. But getting our node there with the connectivity and compute at the edge allows this to happen in real time very, very quickly.
What has user feedback been like for the various trials and demonstrations you’ve held so far?
One of the things that we demonstrate is the ability to take a push to talk radio network that’s local and connect it in within a few minutes into the node, connect that to a 5G device and then not only connect that locally but connect it to, across the globe with another node that’s connected to a 5G network and push to talk devices. When you look at the challenges that we face from an interoperability standpoint, we can basically bring in that analog radio traffic, convert it to digital. and then allow that network to be integrated with other networks very, very easily.
People are just shocked that we can take a push to talk radio, run that over the network, and take that data and have them talk to different devices on the same network, whether it be a 5G phone or a radio from a completely different manufacturer. We could have a Motorola device talking to a L3Harris device, but we can also enable real time language translation with that. Imagine deploying this in the field and you have an Ukrainian unit talking to somebody only speaking English, you can do real time language translation to not only allow interoperability from a connectivity perspective, you also allow it through the ability to translate languages in real time.
That’s just one example. People’s eyes really light up when they realize, “I have a push to talk radio, but I can talk to 5G phones around the world, I can talk to an aircraft in the air”. This is something that they find really, really amazing, especially with how quickly we can do this.
And then the other thing is the ability to use this for things like counter-unmanned aerial systems, the ability to use AI applications for sensors. We get incredible feedback here. One of our customers said, “you know, this solves 80% of the challenges that we have as an organization right now”.
When you think about it from a military perspective, the military has never been able to adopt these devices and applications because they have never been resilient and secure enough for them to use. So this gives them an ability to embrace a lot of the devices and applications like smartphones, by having them actually be usable in the field, have the applications on them work. They have connectivity. They can use them, rather than just having the devices not connected to a network, no access to the cloud, leaving them really pretty useless in the field.
Now you bring infrastructure that supports connectivity in a resilient and secure way. This allows a lot more collaboration, a lot faster ability to introduce new applications and services. That’s what really gets them excited, we’ve had overwhelmingly positive feedback.
Over time, what we’re going to learn is “how are these things used?” “Are there any issues with regard to reliability?” Because these are the things that we want to make sure that we’re improving every single day.
Speaking of improving every day, what is the pipeline for incorporating user feedback into future improvements like?
Right now, we have people at each of our locations where the nodes are. They’re there interacting with the customer. But we’re partnering with companies like Boeing and T-Mobile, who made an announcement on their capital markets all about our partnership. Going forward, they will be the point men in the field gathering feedback from our end customers.
One of the great things about the T-Mobile partnership is that we’re able to leverage their existing infrastructure to cover broad areas really, really quickly with this technology by connecting to it. We really feel that our partners are going to be a big help in getting that feedback back to us as they work with their customers.
Any closing thoughts?
I think our development has kind of come full circle for us. For the last six years, we’ve really been becoming experts in the full architecture ecosystem for the mobile digital world. And it was really to understand the software, understand how these systems work together, because a SEMPRE node is a system of systems functioning as a device.
And in order to get it to function as a device, that has been a lot of work on the software side and integrating the hardware. I think we’re at the point now where we can actually come back to that original technology, that original compute architecture that offered very, very high performance with very low size, weight and power requirements – now we know how to put them all together. And so the vision really was to have very small devices that run on very little power, that allow you to empower the digital world, no matter where you are, no matter whether you’re connected to the internet or not connected. And I think we’ve come full circle on that vision, we’re ready to deploy that as a fully baked product.
We’re excited to get it out there because we see so much need both in the military for interoperability, but also in the commercial space for resiliency and security.
You have to hope that it doesn’t happen every time, but having Hurricane Helene demonstrate one use case for SEMPRE’s products in such a dramatic fashion is definitely timely.
Well, look at what’s happening today – AT&T outages after Helene, the Microsoft CrowdStrike problem, these are all problems that struck our infrastructure deployed in centralized locations, like airports where Delta was severely impacted. If you remember the Maersk ransomware attack, these are all problems that our nodes are designed to solve right there. They’re taking us away from these highly connected, centralized systems and they’re creating this highly decentralized, segmented architecture with separated control and data planes. That allows us to stop an attack that’s happening in order to protect these systems.
We see our ability to interoperate with the existing infrastructure today, and to be able to stop a cascading failure each time is one of the most important things we can do.