Conversation
Highlights

How Icarus Robotics Secured NASA Deployment in Their First Year

Astronaut time costs $130,000 per hour. A significant portion doesn’t go to breakthrough cancer therapeutics or materials science experiments. It goes to moving cargo bags, cleaning equipment, and routine maintenance—work that doesn’t require years of PhD-level training.

In a recent episode of BUILDERS, Ethan Barajas, CEO and Co-Founder of Icarus Robotics, shared how his company secured a deployment partnership with NASA and Voyager Space for the International Space Station in 2027—despite being just over a year old. The strategy wasn’t about superior robotics technology. It was about timing market entry to a structural economic shift in how space stations are funded.

The Economic Shift That Unlocked the Market

For decades, NASA funded the International Space Station. The $130,000 per hour cost of astronaut labor was absorbed by government budgets. If you proposed robotic labor replacement during that era, Ethan explains, “people would laugh at you and say hey, NASA’s got the bill on that one.”

At the turn of this decade, the economics fundamentally changed. Commercial space stations began preparing to launch: Axiom, Vast, Orbital Reef, and Star Lab are scheduled for 2027 and 2028. These aren’t government agencies with open-ended budgets—they’re commercial operators who now bear the $130,000 per hour labor cost directly.

“You have this big switch from a government backed international space station where that $130,000 an hour is footed by NASA and at the turn of this decade now that’s footed by a commercial company,” Ethan said. Labor efficiency transformed from optimization to immediate P&L imperative. “These commercial space stations are launching know next year in 2027 and the year after another company…they have to foot that bill.”

Stacking Convergent Infrastructure Bets

Icarus made three simultaneous wagers. If all three materialized, they would create compounding option value rather than sequential validation hurdles.

First: teleoperated robotics architecture. Second: optical communications delivering the sub-100ms latency required for Earth-based control. Third: the commercial station market transition itself.

Historically, ISS communications relied on S-band radio relays—routing signals 22,000 miles to GEO satellites and back. “You get about 800 millisecond latency, pretty bad,” Ethan noted. But technologies like Starlink promised sub-100ms performance: “It’s probably better than what we’re talking to each other right now.”

When Polaris Dawn validated optical comms in orbit, all three bets converged. “We see the Polaris dawn mission and we see that technology come to rise. And that was a really big unlock,” Ethan said. Low latency enabled teleoperation, teleoperation generates training data, training data builds toward autonomy, and commercial stations provide the economic model to monetize robotic labor.

Architecting the Data Moat Before Deployment

Icarus structured their initial product not just to deliver operational value but to generate a proprietary dataset no competitor could replicate without orbital deployment.

Their teleoperation approach mirrors terrestrial robotics scaling patterns. “The way that you’ll see terrestrial robotics companies train autonomy and tasks is they teleoperate the robot, they collect their in distribution data with cameras and video of the actual physics that’s happening,” Ethan explained.

This data trains high-level primitives: “Move cargo bag from node A to node B or carry out X task. And autonomy grows with deployment time.” By their 2027 ISS deployment, they’ll accumulate operational physics data in microgravity that no competitor can access. The ISS isn’t just their first customer—it’s the corpus generator for intelligent robotics across all space applications.

Embedding Into Infrastructure Design Cycles

Rather than responding to finalized RFPs, Icarus engaged commercial station developers during design phases—when modifications cost nothing and switching costs can be embedded structurally.

“We’ve had conversations with some of these commercial stations to say, hey, actually, if we changed this hatch closeout to have tabs that are 2cm larger, it’d be easier for a robot to actually change. Or if we put a fiducial here, the robot could localize itself,” Ethan shared.

These modifications are trivial in CAD, impossible post-manufacture. By positioning as design partners rather than vendors, they’re influencing station architecture before procurement processes begin. “We’re working with them in a collaborative sense to design for a robotic architecture rather than purely a human one.”

Mapping Horizontal Expansion During Customer Discovery

While ISS labor provided the wedge, Ethan conducted hundreds of customer discovery calls to systematically map adjacent markets requiring similar capabilities. “Really mapping what it looks like and where the immediate needs and where some of the hair and fire things were.”

The opportunity extends far beyond station housekeeping: “We have multi thousand agent satellite constellations now. We have people talking about data centers in space, we have people talking about so many of these infrastructure projects that need maintenance, that need robotics to actually change out physical parts.”

Each market—satellite constellation servicing, space infrastructure maintenance, eventual cislunar operations—requires overlapping capabilities: microgravity operation, autonomous navigation, physical manipulation. “Once you kind of use the ISS as a wedge, we can actually use that intelligence in robotics to service other parts of the industry and expand out horizontally from there.”

This mapping transformed their pitch from point solution to platform play, fundamentally changing investor appetite and market sizing conversations.

Exploiting the Flight Heritage Technology Gap

NASA operates under Flight Heritage constraints—a database of every component that’s flown to space. “If it’s there, you use it,” Ethan said. The result: robotics currently operating on the ISS run on chips and boards that “stop production in the early 2000s. Stop production.”

The technology gap between terrestrial and space robotics spans two decades. A smartphone today has capabilities that would have been impossible for ISS robotics in the early 2000s. “If you think about like even a smartphone and what your phone can do now versus what it could have done in the early 2000s and the advancements we’ve made in robotics, that’s a really amazing and exciting time.”

The shift from NASA-backed to commercial-backed operations eliminates Flight Heritage as a binding constraint. “That pivot from NASA backed to commercial backed allows us to leverage these new technologies and put them into space.” Commercial operators can deploy modern technology without decades-old qualification processes.

Deploying Distribution for Technical Talent Acquisition

Deep tech companies competing against FAANG compensation need different talent strategies. Ethan recognized distribution serves dual purposes: customer acquisition and engineering recruitment.

He observed how ChatGPT’s launch “drew some of the best minds to work on it” because distribution made hard problems visible to engineers who self-select for technical difficulty. “Getting really intelligent people to work on really hard problems. People want to work on hard problems.”

Icarus uses media, podcasts, and thought leadership not primarily for brand awareness but to make engineering challenges visible to talent that prioritizes problem difficulty over equity compensation. For technical founders, distribution ROI should be measured through engineering hire quality and time-to-close, not vanity metrics.

The Economic Inevitability Framing

Ethan doesn’t pitch robotics capabilities—he presents labor economics that make automation structurally inevitable. The space industry has had “700 astronauts in all of history. There’s just about 100 active right now.” Yet the industry plans to deploy data centers in space, massive constellations, and scaled infrastructure.

“We’re going to use a space shuttle and tether them and have them go dock to the Hubble telescope and float out there with a wrench and fix it. It’s not the most scalable thing in the world.” The procurement conversation shifts from “should we buy robots?” to “how do we avoid being uncompetitive without automation?”

“When you paint it in this very obvious picture with the tailwinds from the entire industry, it becomes very clear that robots have to be on orbit. Not removing humans, but making humans most effective at what they’re doing.”

Looking Ahead

Ethan sees robotics as foundational infrastructure for space expansion. “For all of these things to occur, robotics will be there supporting every step of the way in every part of from low Earth orbit to CIS lunar space to Martian missions, robotics will be there.”

The 2027 ISS deployment represents the wedge. As commercial stations multiply, satellite constellations scale to thousands of agents, and space infrastructure projects materialize, the robotic labor corpus compounds in value.

“To be able to move that from what we’ve seen the early 2000s and only NASA and JAXA and big institutions tackling it to now commercial companies and you know, be some of the pioneers there is pretty dang exciting,” Ethan reflected.

Icarus Robotics demonstrates how technical founders can compress validation timelines by aligning product roadmaps with structural market transitions. By stacking convergent infrastructure bets, architecting data moats before deployment, and inserting into design cycles rather than procurement processes, they’ve positioned themselves at the center of commercial space operations—all within their first year.