Exo-Space Co-founder Jeremy Allam Discusses Space Data

In 2018, nearly the entire town of Paradise, California was destroyed by a wildfire. NASA officials and fire chiefs have said that one key factor in minimizing the impact of any fire is how quickly a fire is identified. Current state-of-the-art satellites take five hours from the time that a picture is captured to when it is accessible to the end user, informing them of a fire’s location. The problem is that in those five hours, up to 40,000 acres of forest can be burned. That five hour time delay is what the industry refers to as “data latency.” And data latency can really mean the difference between a catastrophic fire and one that is quickly contained. 

It turns out this problem of data latency is not only present in fire fighting. If you look halfway across the world in the oceans of Southeast Asia, you see the same issues with organizations that are trying to locate illegal fishing vessels. They’re often not able to intercept these illegal fishing boats because the photograph information they need is arriving too late. 

And that is where Exo-Space comes in. We’ve developed an image processing device, which we call FeatherEdge, that mounts directly onto a satellite and processes images coming from the camera to identify objects or events of interest from space in near-real time.

My last job at NASA was what inspired me to take the leap. Since I was very young, I wanted to work on spacecraft at NASA. After graduate school, I was able to work at NASA JPL as a software developer for satellites. I loved working there, but three months into the job I realized that what I really wanted to do was focus all my time and energy on creating a business around next generation satellites.

Space is a really exciting field to be in right now. We’re in a new era of space technology where the cost of launches has been reduced by almost a factor of 100 compared to what it was 10 years ago. micro-electronics have become widely available, which is opening up new market potential for for-profit space companies to solve really hard problems. I think it took me achieving a lifelong goal to realize that I actually wanted to start a company.

I have a master’s degree in aerospace engineering and worked at NASA JPL as a satellite software developer. Prior to that, I was the lead software developer for a satellite project at the Space Engineering Research Center, so I have a lot of experience leading teams and also developing space technology. Basically, the opportunity to lead engineering teams while also developing my technical skills on the software and space side has really helped in starting this company.

Currently, the space computing industry is early stage and fragmented — with each company focusing on a different aspect of space computing. I think what’s going to happen in the future is a lot of these fragmented, specialized space-computing companies will begin to consolidate. We’ll start to see larger companies emerge that are able to serve most needs in the commercial information-as-a-service space market. This consolidation will also lead to the massive tech giants getting involved in orbital cloud computing and much more general-purpose data processing in space, which currently is a $1 trillion global market. 

Regardless of how the market settles, most experts agree that in the next 10 years there’s going to be a million times more data generated in space every single day. That means the current ground station and earth-link pipelines aren’t going to be able to support that insane amount of data. So this lends itself to the view that if you want to operate in space, you’re going to need to do a lot of computation on orbit and only send down the relevant data. Customers are already interested in knowing when a satellite passes over a certain area, what type of crop is there or what is the overall health of the crops, or how many ships are in a harbour on any given day. It doesn’t make sense to send down a ton (gigabyte or more) of data, just to process it on the ground to find out that there are nine cargo ships approaching the San Francisco Bay. So, to be able to enhance insight-driven decision making in the future means that we need to reduce the amount of time and, therefore, the amount of data being sent down from satellites.

There are two other co-founders on our team, in addition to three interns. I originally met my co-founder Marcel when we were in graduate school. He has a similar background to me with a master’s in aerospace engineering, but his focus is on mechanical engineering. He was the lead structural design engineer for a satellite project at the Space Engineering Research Center, and then he joined Exo-Space after that. We’ve been friends for a long time, and since my software experience and his hardware experience are so complimentary, it just sort of naturally came about that we started this company together. 

Our third co-founder, Mark, actually started as a subcontractor for us. We hired him to help do some mechanical design work for a contract we had been awarded at the time. Throughout that process, Mark proved to be exceptionally skilled on the technical and business side, so we offered him the co-founder position six months after he started. Mark’s background is in business development as well as engineering. So, altogether, we have a well-rounded team of technical and business founders.

There is some competition, but as I mentioned earlier, the market is so new right now that most of the companies that we would consider competition are operating in the space data market — but not specifically in our domain of expertise. For example, there’s one company that focuses on general-purpose edge computing, and there’s another company in Europe that focuses on machine vision but for very specific use cases. We’re sort of in between those two companies, but regardless, the way we differentiate ourselves is in two keys ways: 1) the processor we use is the fastest machine vision processor on the market, which reiterates how focused we are on reducing space data latency, and 2) our FeatherEdge platform is designed to integrate with any satellite system, no matter how big or small. Since our team has a lot of technical experience in the space industry we were able to design a system that is lightweight, low power, and which easily integrates onto other satellites.

Our business model is pretty simple: we send our device into space, and we allow customers to rent time on our device by paying a monthly subscription fee. The subscription fee allows customers to update their detection model on orbit whenever they want, and they can get data down whenever they want. Basically, it allows them to monitor their areas of interest on their own terms. In addition to renting the time, we also develop custom detection models for our customers. If a customer doesn’t want to build their own detection model but they still want to be able to identify something from space, we can develop a custom detection model for them (for a one-time fee) and deploy it to our satellite.

We really want to engage with the space community. Typically, space technology companies don’t do crowdfunding rounds because the required capital to get started is so high. But we see a lot of early investors as a huge asset, and we want space enthusiasts on our side because they provide such a strong network and positive conversation. The money we raise in this round will go directly towards paying for our first launch into space.

We have a solid technical team, and we’re backed by a group of expert advisors with experience in the government and commercial space industry. We have some very exciting announcements coming soon and for a detailed list of our traction, progress, and pitch please refer to our offering on the Spaced Ventures website.

In the next five to 10 years, I think there will be a consolidation of the companies operating in the space data market. Our goal is to use the momentum we’ve gained in the machine vision vertical to position ourselves as a leader as we move into more general purpose space-computing applications. The hardware and software architectures that we are developing in this machine vision stage have been designed to scale for multi-user applications of orbital cloud computing. We’ll be able to do things like predictive maintenance of satellite constellations or autonomous rendezvous and docking of satellites in orbit. The space data market is a $7 billion dollar market, but the global cloud computing market is a $1 trillion dollar market and likely to be where most of the space computing companies today end up. 

So in the next five to 10 years, we are set on becoming a cloud-computing expert in the space domain. Our main priority as a company is to give maximum return to our stakeholders. When we get further along, we will consider a merger or acquisition if it makes sense and allows for the highest return for investors. At this point, we are solving problems for our pilot program customers, and we see a lot of potential for growth in the near future. We’re very focused on building a strong company around great tech and a solid financial plan. This technology has the potential to grow into a strategic option for the largest competitors in our space, like Amazon, IBM, or Microsoft, and we are committed to growing over the long term in a way that will give us more options in our exit strategy.