Why Armilar invested in Neuraspace…
… and a few insights on Space Traffic Management
TL;DR
- We are now going through a new Space era driven by entrepreneurs
- The Space economy is poised to grow into a 1 trillion dollar economy
- But the exponential growth of satellites and debris makes it very dangerous to operate in Earth orbit: critical services could be disrupted and constellations plans made unfundable
- Neuraspace aims to solve this problem by providing satellite operators an AI-powered solution that prevents collisions between satellites and debris
The new Space race is on… and it’s like nothing we’ve seen before
We are now at Space 4.0 and it’s an extraordinary time for founders and investors. For context:
- Space 1.0 refers to the era before humans had the ability to launch satellites and relates mostly to astronomy
- Space 2.0 begins with the emergence of rockets and satellites, and covers the first Space race period, which was fueled by the cold war
- Space 3.0 invokes the decades-long era of the International Space Station (and, in general, of international cooperation), Space Shuttle, broadcasting communication satellites, Global Navigation Satellite Systems (GNSS), amazing scientific and technological advancements and the emergence of new spacefaring nations
- Space 4.0 (which VCs call SpaceTech and the Space community calls NewSpace) is all about entrepreneurs taking over this “traditional” sector and dramatically expanding its market size
Perhaps with the exception of broadcasting communication satellites, the vast majority of Space activities until a few years ago were driven by scientific, geopolitical, and military goals.
How things have changed since SpaceX (and other trailblazers) began disrupting the status quo (intelligently supported by NASA and other agencies) and brought entrepreneurs front and center to the new business-driven Space era… It was the birth of Space 4.0.
To get some sense of “how much” things have changed, consider the exponential growth in the number of satellites launched — driven by commercial small satellites placed in Low Earth Orbit (LEO).
And more recent data continue to impress: in H1 2022 alone, more than 1,100 satellites were launched.
Other relevant and forward-looking insight is the sheer quantity of startups being formed and the massive investments directed to them.
From this base of hundreds of founders receiving very meaningful funding, we will certainly witness many avenues of growth in the industry.
It is very clear (and quantifiable) that in less than a decade, Space has evolved into one of the hottest areas for entrepreneurs and investors. And entrepreneurs are doing what they do best: move fast and build things. Space 4.0 is here.
As exciting as SpaceTech may seem in 2022, this is really just the beginning… according to multiple sources, the Space economy is poised to grow from just under USD 400b in 2021 to USD 1t in 2040. If this growth forecast turns out to be approximately correct, we are looking at hundreds of billions of dollars of value being created in the next couple of decades (from an already sizeable, multi-billion dollar, industry).
How has Space 4.0 come to be and why do we believe that now is the time to get involved?
We have identified four structural shifts that we have seen play out in other industries where we have backed incredibly successful founders (notably in IT and software):
1. Once-in-a-generation structural cost reduction
The cost to launch satellites into orbit has been radically reduced. By 2018, SpaceX’s Falcon Heavy’s cost to insert 1kg into LEO was only ~10% of the equivalent average cost of legacy launchers of about a decade earlier. Some estimates point to continued cost reduction down to a few hundred dollars, driven by even more reutilization, and by new materials and processes
2. Disruptive and scalable technologies
Satellite and ground equipment technologies have evolved significantly and are also benefiting from general advancements in both hardware and software. Miniaturization, standardization, and industrialization mean that satellites (and their related ground equipment and receiving infrastructure) can now do a lot more with a lot less
3. New business models
We’re seeing an evolution from rocket, satellite, and ground equipment manufacturing and selling into essentially what many call Space-as-a-Service. This shift is opening up Space to players who could not own, manage or even access this infrastructure and to new “consumers” who will use Space-based solutions when they would never contemplate doing so before (case in point: the new iPhone 14 satellite communication features)
4. Talent magnet
Engineers and scientists flock to Space activities for passion. Perhaps more so than in most other professional fields. The Space sector has never really had a difficult time attracting technical talent. What’s different this time is that Space is also alluring non-Space native entrepreneurs who look at Space as another sector where they can build successful businesses: this was mostly absent until a few years ago
With all these tailwinds and excitement, there’s nothing standing in the way of the one trillion dollar Space economy…
… or is there?
The problem with real estate scarcity and occupancy on Earth’s orbits
Useful Earth orbits are a relatively scarce resource: premium Space real estate of sorts. LEO in particular is getting crowded with satellites and debris, which make in-orbit operations dangerous.
The number of satellites and debris will continue to grow
The satellite population on Earth’s orbits is predicted to grow from less than 6,000 today to perhaps close to 100,000 in the years to come.
The European Space Agency (ESA) estimates that there are over 100 million objects of up to 1cm orbiting the Earth. More than 36,000 of these are bigger than 10cm. Many of these objects would likely destroy (or at least, significantly impair) a satellite if they hit it: which is rather intuitive if one takes into account that debris typically travel at about 10 times the muzzle velocity of a bullet fired by an M16 rifle.
Furthermore, the debris population has been continuously growing: for the past 20 years, there have been, on average, more than 10 “fragmentation events” per year: including those resulting from collisions, explosions, and even missiles that destroy satellites (the latest event was in 2021 and it created more than 1,000 new debris, which ultimately doubled the risk of orbital debris penetrating the ISS).
Importantly, one of the biggest problems with debris tracking is that we can’t really monitor all of them with current technology and infrastructure. As of today, possibly only slightly more than 30,000 debris objects are actively monitored (and it’s impossible to avoid colliding with what we can’t “see”).
Houston, we have a Space Traffic Management (STM) problem
While the urgency of the problem is new (as satellite constellations are only now being launched at breakneck speed), the theoretical knowledge of the threat of Space debris is not a recent topic. In 1978, Kessler (a NASA scientist) published a seminal paper that highlighted the perils of LEO exploitation (and consequent debris formation) and a potential cascade of fragmentation events that could eventually lead to a debris belt around the Earth, seriously threatening the future of Space exploration. This concept evolved into what is widely known as Kessler syndrome: in simple terms, it’s a domino effect resulting in an ever-growing number of debris, which is triggered once a certain number of fragmentation events take place. This sounds catastrophic (and it could be), but it is not inevitable. If we solve the most immediate problems in a smart way and agree on sustainable practices, we should be ok.
Should companies or Governments solve this problem?
Considering that LEO is basically “open for any operator to use”, one may reasonably argue that there are no direct and short-term incentives for each operator to act in a way that is conducive to a sustainable exploitation of this shared and limited resource. This circumstance could be tantamount to a “tragedy of the commons”, for which, some could say, regulation is the only answer.
We don’t believe the answer lies in a Government(s)-centric approach, but we do consider that some basic regulation (simple “rules of the road”) is required, namely at four levels:
- Satellite in-orbit insurance
- Compulsory collision avoidance capabilities
- More aggressive resource management plans, including end-of-life satellite de-orbiting
- No more anti-satellite missile tests (and the like)
While these (and other) regulatory changes are important, they will neither be sufficient nor will they come in due time. The healthy pressure created by SpaceTech founders to deploy tens of thousands of Space assets (to feed into their data pipelines and business models) is incompatible with what could be (optimistically) a decade-long international effort to bring regulation and Government(s)-centric solutions up to speed.
In practical terms, the Space community has three options:
- Significantly reduce the entrepreneurial SpaceTech momentum and radically limit the number of new satellites being placed into LEO until regulations and institutional entities solve the problem
- Ignore the problem and just hope that nothing goes wrong
- Build on SpaceTech’s entrepreneurial drive and embrace commercial solutions that solve the most immediate problems, while building regulatory frameworks that are in line with the expected technological innovation that will come from these solutions
Not hard to guess what we think should be done…
Regardless of regulation, or mitigating solutions, there will always be a large number of objects in Earth orbit, and the most pragmatic way to address today’s real and immediate problem, we believe, is to provide each satellite operator (and, collectively, the users of Earth’s orbits) with cost-effective ways for them to safely run their operations taking into account this increased orbital congestion.
So, assuming it is reasonable to agree that commercial solutions are the way to go, what is the problem with today’s approaches? Why do we need innovation and startups in the STM segment?
Current STM processes and the cost of keeping satellites safe
Here is a (very) simplified STM workflow:
- The whole process starts when a satellite operator receives a “warning”: technically, a Conjunction Data Message (CDM) provided by a space situational awareness provider, notably by the US Department of Defense 8th Space Defense Squadron. This CDM is built from data collected by Earth-based sensors (in the future could include in-orbit sensors), such as radars, and is based on analyses that project the orbital paths of objects it observes and determines whether a crossing of paths between two objects is likely to occur
- How many CDMs an operator receives per satellite per year depends on the orbit and physical characteristics of the satellite and whether the satellite and the objects it could collide with are tracked. Based on several assumptions (including orbital altitude and inclination), ESA estimated that, in 2021, a satellite in LEO could receive something like 40 alerts (taking data for h=550km and i=53º)
- Once an alert passes through the first screening — based on the operator’s “internal systems and processes” (which, for multiple operators, is fundamentally a manual screening procedure) — it is “escalated” and brought into a deeper (typically, manual and very costly) analysis. CDMs trigger actions based on several parameters, including the probability of collision, which some operators may consider at a threshold of ~10⁻⁴
- After this analysis, only a fraction of the alerts actually result in a collision avoidance maneuver (if the satellite has that capability)
- ESA, for example, reported performing “on average” two collision avoidance maneuvers per satellite per year for LEO; with some satellites operating in more debris-intense orbits performing twice as many
The operational cost per executed maneuver depends on a satellite’s characteristics and orbit, as well as on multiple other variables, including the operator’s: i) procedures, ii) cost structure (which is also a function of the number of satellites in a constellation), iii) insurance policy (if any), iv) risk appetite (and SLAs with clients), as well as v) regulatory constraints and liabilities. An OECD study indicates that 5–10% of the overall mission cost of a Geostationary Orbit (GEO) satellite is dedicated to the full range of protective and debris mitigation measures (which go beyond tracking and maneuvering). It is also noted in the report that this percentage would increase significantly for satellites in LEO, where debris and satellite numbers are significantly higher. Furthermore, it is reasonable to consider that perhaps half of a mission’s cost beyond design (i.e., operational costs not planned for) could be attributed to collision avoidance.
On top of the operational cost, we need to contemplate the potentially much, much more significant opportunity cost of false positives (maneuvers that were unnecessary and reduced the satellite’s lifetime, disrupted business and resulted in lost revenue, and eventually broke SLAs) or false negatives (basically choosing not to maneuver and colliding with debris or another satellite).
Cost is only part of the problem; the bigger issue is that current STM solutions will not scale to the new Space 4.0 paradigm, for two motives:
- The number of alerts will increase exponentially. This will happen for two reasons. First, the number of objects a satellite could collide with will increase dramatically (see the comments above about debris and satellite population growth). Second, more precise radars and other sensors will detect many more small objects which are currently not catalogued and will therefore increase the “visible” conjunctions
- Current methods of generating alerts are not accurate enough, yielding too many false positives (in part driven by operators’ rational conservative approach of avoiding a collision even if the probability is “low”). These false positives happen at two levels. First, only a fraction of alerts actually results in maneuvering decisions after detailed and resource-consuming analyses. Second, many of the actually executed maneuvers would likely not be necessary anyway
Following these two points (increased number of alerts and too many false positives), there is no way that current heavily-manual STM tools and procedures (for analysis, maneuver planning and execution, and inter-operator communication) will be able to cope with a much more dynamic environment.
Enter Neuraspace
Delivering in-orbit safe operations and business continuity
At its core, Neuraspace offers satellite operators peace of mind so that their leaders and staff can focus on their business, which is delivering value to customers by providing data and added-value services stemming from their Space assets.
Neuraspace’s platform abstracts the growing complexities of in-orbit collision avoidance and delivers satellite operators value by:
- Enabling operators to maintain their assets safe in an environment that is increasingly much more prone to collisions and for which current tools and procedures are inadequate
- Significantly reducing costs related to alert management by delivering much more accurate risk assessments
- Automating critical and time-consuming tasks, such as detailed analyses and execution of maneuvers
Data. Artificial intelligence. Automation.
The front-end of Neuraspace’s system is a virtual assistant to satellite operations managers, providing them with a single pane of glass for their assets’ collision risk posture and evasion strategies.
Under the hood, Neuraspace has developed a sophisticated data fusion strategy that feeds into its proprietary machine learning models to provide more accurate collision risk assessments and actionable insights.
In addition to collecting data from existing sources, Neuraspace is developing its own proprietary Space object catalogue and merging different data types (e.g., space situational awareness data, satellite digital twins, etc.). Neuraspace feeds this “enhanced proprietary data set”, into its proprietary models which deliver more accurate and, importantly, explainable results (operators don’t have to trust a black box; they can trace the underlying data and assumptions that lead to a certain risk assessment).
A core feature of Neuraspace’s solution is to automatically, and preemptively, suggest multiple evasive maneuvers, which contemplate not only the current need to prevent a collision, but also address the consequences for the satellite’s (and constellation’s) operational requirements. This will give visibility to satellite operators not only of each instance of collision avoidance but also of the impact on their businesses.
Time to grow
Neuraspace was founded in late 2020 and the team did not waste any time. In this short period of time, they have secured contracts with ESA, developed an MVP, rolled out their product, and are currently undertaking pilots with several notable commercial satellite operators.
In addition to our Seed investment earlier this year, Neuraspace has been awarded a grant from the Portuguese Government, which significantly boosted the company’s ability to expand its data infrastructure, continue building product features, enhance its go-to-market and reinforce customer support.
While the company did hit the ground running, we are all well aware that there is a lot of work to be done, obstacles to be overcome, risks to be avoided and mitigated, and (truth be told) respectable competitors to be taken into account. As in any moonshot investment, even if the potential prize is high, the odds of success are typically not.
Still, Neuraspace’s stellar team is nothing but fully committed to being the global leader in STM, as they believe that their solution can deliver the highest and fastest value to satellite operators. The team’s enthusiasm and drive to be a key contributor to the sustainable exploration of Space are contagious.
So what is it about the team that excites us so much?
A unicorn builder, a renowned Space expert, and a passionate and talented group
One of the privileges of being around for more than twenty years as a firm is that we get to partner with talented entrepreneurs spanning multiple industries, geographies, and generations. And some of them just can’t stand still, despite their extraordinary successes. It’s the case of Nuno Sebastião (co-founder and CEO of Feedzai), who founded and invested in Neuraspace.
We initially partnered with Nuno and his Feedzai’s co-founders, Paulo Marques and Pedro Bizarro, when we led Feedzai’s first funding round back in 2011. Feedzai is now a leader in financial fraud detection and mitigation. It’s been an extraordinary adventure to witness Feedzai grow into the unicorn that it is today (and it keeps growing…).
Prior to co-founding Feedzai, Nuno worked at ESA, more specifically at the Agency’s operations center in Germany, and he has kept in touch with many of his former colleagues. In one of those interactions, Nuno realized that satellite operators were facing a problem similar to what he had seen before (and is now solving with Feedzai). The fundamental conceptual challenges that operators face to undertake safe in-orbit operations, it turns out, are analogous to those faced by financial institutions with regards to preventing fraud:
- A deluge of data coming in from multiple sources that need to be handled and analyzed much more efficiently
- Risk assessment models that are neither complete nor accurate enough
- Split-second decisions needed to be taken, which have measurable business impact
- Critical manual tasks that need to be automated for scale
Nuno thought “why not solve this problem (again)”?
To execute on his vision for Neuraspace, Nuno partnered with Chiara Manfletti to lead the company’s operations. Chiara took on the role of COO of Neuraspace following a successful career as a researcher at the German Aerospace Center (DLR), as an advisor to ESA’s Director General, and as the first president of the Portuguese Space Agency. Chiara is now growing a team of incredibly talented aerospace engineers, data scientists, product developers, business development and sales managers.
Fueled by Chiara’s leadership and deep sector expertise, and by Nuno’s precious mentoring and experience, the team is powering ahead.
When Nuno invited us to be Neuraspace’s Seed investor, we got to work and did a deep dive into the subject. Realizing the magnitude and urgency of the problem, the company’s unique solution, and the stellar team being built, we obviously jumped on this amazing ride.
Above and beyond
We see an enormous amount of talent getting into SpaceTech and we’ll continue engaging with the Space community. While we didn’t build a SpaceTech investment thesis, we will be paying particular attention to data-driven and software-based: i) enabling infrastructure and ii) downstream solutions.
Authored by Rodolfo Condessa, Armilar Venture Partners. Thank you Chiara Manfletti, Pedro Ribeiro Santos and Pedro Castel-Branco for reviewing the text.
