Rise of the ‘twin’

For years, researchers have built AI-powered models to simulate virtual scenarios of the planet. The concept, known as Earth’s “digital twin,” leverages a wealth of real-time data gleaned from satellites, drones and other means, to encompass everything from security issues and climate dynamics to the very systems that drive global supply chains. Reams of data are then captured, processed and parsed to identify subtle, real-time shifts, all in the name of optimizing decisions through simulation.

Consider, for instance, three potential scenarios: A hurricane begins to form in the Atlantic Ocean, consumer behavior starts to change around the holidays, or Chinese troops begin assembling along the Taiwan Strait. Digital twin technology could then be used to predict the path of the storm, analyze shifts in shopping patterns and foot traffic, or forecast the outcomes of a military escalation in everything from the effects on regional security to its impacts on global markets with the kind of predictive insights that decision-makers could once have only dreamed.

In a recent interview I had with Dan Smoot, CEO of Maxar Intelligence — a Colorado-based space technology company that specializes in geospatial intelligence — he said that the ability to see the Earth from those so-called “ant colony” perspectives, and run simulations against them, is precisely “where the pressure in the marketplace is going.”

What “both our U.S. government and international governments really care about is really understanding the dynamics of that changing world, and how they can make rapid decision making around that what used to be,” he told me on the Space Minds podcast.

For Maxar, which has a client list of not only commercial partners but also defense and intelligence agencies, surveillance includes imaging and processing, as well as analysis that can be incorporated into those emerging digital landscapes.

The company’s largest contract is a $3.2 billion, 10-year agreement with the National Reconnaissance Office (NRO), awarded alongside California-based Planet Labs, to provide satellite imagery, as well as shortwave infrared, nighttime and non-Earth imaging — a service that relates to space situational awareness. Under that deal, the NRO can also buy “point collection” services, where commercial satellites are used to capture images of a specific location. And yet how those imaging services are positioned, especially by federal customers, are also subject to changing priorities.

For instance, following a rare on-camera dispute in the Oval Office between President Donald Trump and Ukrainian President Volodymyr Zelenskyy, the Trump administration suspended U.S. satellite imagery support for Ukraine, only to reinstate it two weeks later amidst talk of a potential ceasefire in its ongoing war with Russia.

Meanwhile, Maxar, which had detected the earlier Russian build-up of troops along the Ukrainian border near the start of the war in 2021, has since played pivotal roles in monitoring conflict zones. With roots that date back to the 1960s, it was rebranded in 2017 after acquiring the Colorado-based space imagery and geospatial provider, DigitalGlobe. The company’s services have since evolved, but one of the more major recent shifts, Smoot explained, is the sheer speed with which its services are in demand.

What used to be “weeks and months to be able to assess information” can now come within a “90-minute time frame,” he said. “That’s where I think that you’re seeing this compilation of data and where the geospatial intelligence really starts to drive that kind of decision making,” particularly as digital twin technologies are increasingly buttressed by three-dimensional representations of the Earth.

“3D is absolutely where everything’s moving to,” Smoot added, referencing advanced methods such as stereo imaging, employed by his company, as well as those of Planet Labs and Google.

That approach involves capturing several images of the same area from multiple angles, which allows for a virtual reconstruction of detailed 3D features. Buildings and mountains start to appear more intricate and life-like, creating more immerse models relative to traditional two-dimensional mappings. The technique, while effective, also generates vast amounts of data, however, which can lead to bottlenecks in terms of speed and efficiency.

“You have to do it almost from space, because trying to get that information on a low-level kind of ingest takes forever,” Smoot explained, especially given those newfound speed imperatives. Mapping companies “have to be able to provide that information at a rapid pace now. That’s just the world of which we live in.”

This article first appeared in the April 2025 issue of SpaceNews Magazine.