With recent advancements and discussions about AI, one conversation from the 2024 Space Capital Summit stands out—our panel on "Big Geospatial Data and AI." The discussion, featuring Rachael Martin, Maven Office Director at the National Geospatial-Intelligence Agency (NGA); Nathan Kundtz, Founder and CEO of Rendered.ai; and Michael Sheetz, CNBC Space Reporter, delved into the evolving intersection of artificial intelligence and geospatial data, offering key insights into the challenges and opportunities shaping the future of this industry.
The conversation was framed around a fundamental question: how do we manage and extract value from the exponentially growing volume of geospatial data collected from space? As we continue to deploy more satellites and sensors, the sheer amount of information being brought down to Earth is staggering. Rachael Martin underscored this challenge, explaining that "the volume of data has surpassed the capacity of human analysis," necessitating advanced AI tools to process and make sense of it.
Nathan Kundtz provided a compelling perspective from the private sector. His company, Rendered.ai, is tackling the issue by leveraging synthetic data to train AI models before real-world geospatial data even becomes available. "Imagine trying to build an AI system for satellites when the training data doesn’t yet exist," Nathan noted. "We enable companies to simulate data at scale, helping them stay ahead of the curve."
One of the key takeaways from the discussion was the need to differentiate between various forms of geospatial AI. As Rachael explained, geospatial AI can be divided into two primary categories: image-based AI, which focuses on object detection and recognition, and database-driven AI, which synthesizes vast amounts of geospatially tagged information to uncover patterns and insights. "Each type of geospatial data requires a different AI approach," she explained. "And as we increase the number of sensors in space, our challenge isn’t just more data—it’s more diverse data."
Michael Sheetz, who moderated the discussion, did an excellent job framing the conversation within the larger industry trends. He pointed out that, for years, we’ve operated under the assumption that more satellites would automatically lead to better insights. But as we’re now realizing, more data doesn’t necessarily translate into better decision-making—at least not without the right AI tools to extract meaning from it.
The discussion also touched on the role of synthetic data in AI training. Nathan shared an example of how Rendered.ai worked with a satellite company aiming to detect methane plumes. "You can’t just wait for a satellite to capture the right image to train your model," he explained. "We created large-scale simulations to train the AI ahead of time, allowing it to recognize methane plumes in real-world data as soon as it became available."
One of the most compelling moments came when Rachael discussed the strategic importance of AI in national security applications. "Our adversaries aren’t making it easy for us," she noted. "We have to anticipate threats before they emerge, and that requires AI models capable of predicting patterns based on incomplete or imperfect data."
The panel also explored the evolving relationship between government and the private sector. In recent years, we’ve seen a shift in how defense and intelligence agencies engage with startups and venture-backed companies. As Nathan pointed out, "Eight years ago, AI was seen as something reserved for PhDs in academia. Today, anyone can train a computer vision model in 30 minutes using open-source tools. The bottleneck isn’t the AI itself—it’s the data."
Rachael echoed this sentiment, emphasizing the role of commercial partners in keeping pace with rapid technological changes. "The government doesn’t move as fast as the private sector," she admitted. "We rely on startups to push the boundaries and show us what’s possible. But at the same time, startups need to understand our constraints—especially when it comes to procurement and budgeting timelines."
Looking ahead, the conversation turned to what’s next for geospatial AI. Rachael highlighted the potential of generative AI in enabling non-experts to interact with complex geospatial datasets. "Imagine being able to ask a natural language question—like ‘show me all ships that left San Diego flying a specific flag in the past six months’—and getting an immediate, AI-generated response."
But as Nathan pointed out, with these advances come new challenges. "Generative AI is non-deterministic, meaning we can’t always predict its outputs with certainty," he cautioned. "That’s a big issue when dealing with intelligence and national security applications. The next frontier is figuring out how to evaluate and validate these models effectively.
The panel concluded with a discussion on what needs to improve in the collaboration between the private sector and government agencies. Rachael’s message to startups was clear: patience is key. "We are getting better," she assured the audience. "Bringing new technology into the Department of Defense is hard. But if you can solve our problems, you’ll find that your solutions have applications far beyond just the government."
As I left the discussion, I was struck by how far we’ve come—and how much further we still have to go. AI is no longer just a futuristic concept for geospatial intelligence; it’s a necessity. But making the most of this technology will require deeper collaboration between investors, startups, and government agencies. At Space Capital, we’re committed to fostering these connections and ensuring that the next generation of geospatial AI companies can thrive.
The 2024 Space Capital Summit provided an invaluable forum for these conversations, and I look forward to seeing how our industry continues to evolve in the years ahead. The advancements we discussed today are just the beginning. As we continue investing in AI-powered geospatial technologies, we must also invest in the partnerships, policies, and infrastructure necessary to maximize their impact.
For more insights, watch the other panel discussions from the 2024 Space Capital Summit on our YouTube channel.
