SpaceX currently controls nearly every rocket launch on Earth. With 98% of global orbital launches under its belt, Elon Musk’s company has become the de facto gatekeeper to space. While China is still a contender, other countries and private companies have fallen dramatically behind, and this isn’t likely to change anytime soon, with Trump’s proposed budget cuts for NASA gutting the agency even more.
But this isn’t just about bragging rights or accomplishing great things. The stakes here include who controls satellite communications, who can explore space affordably, and even who sets the rules in orbit. And for a Dubai-based startup called Leap 71, leaving that to Musk isn’t an option.
The company has a pretty bold ambition. It wants to use artificial intelligence to design and mass-produce rocket engines faster and cheaper than anyone else can. Their secret weapon is a computational system called Noyron, an AI that Leap 71 says doesn’t just guess like generative models do.
Instead, it has physics and manufacturing constraints directly encoded into its core. The result, the company claims, is a new class of aerospace tools capable of building advanced engines in weeks, not decades. According to an interview with Fast Company, Leap 71 cofounder Lin Kayser says catching up to SpaceX without the use of AI is impossible.
SpaceX has been working for a decade to create the rocket engines that power its missions, and spending 10 years trying to catch up will just leave you further behind. However, unlike traditional engineering, where rocket designs are hand-crafted through long, complex iterations, Noyron generates complete, physics-accurate models on demand. An AI that prints new rocket engines daily would really help push things along.
These designs aren’t just pretty renderings, either, they’re reportedly ready for 3D printing. This approach enables Leap 71 to produce fully functional, high-performance engines without the usual years-long delays or astronomical costs that have been associated with space exploration in the past.
In 2024, the company successfully fired a small engine designed entirely by Noyron and printed as a single piece of copper with internal cooling channels. The engine fired flawlessly.
Leap 71 is now scaling up. It’s developing two next-generation engines: the 200-kilonewton XRA-2E5 and the 2,000-kilonewton XRB-2E6, which rivals SpaceX’s powerful Raptor engine. These engines are expected to reach test readiness by 2026 and 2029, respectively. If successful, it could prove the viability of AI-printed rocket engines for large-scale spaceflight.
Of course, there are obstacles. Large engines require equally large test facilities, of which there are few worldwide. Shipping components across borders is slow and politically tricky. The engines themselves, while rapidly designed, still need to be validated under real-world stress.
Leap 71 has already had to “test blind” by firing engines they couldn’t fully inspect internally. And while every test feeds back into Noyron’s learning loop, that process takes time and money.
But the potential payoff is enormous. Smaller companies and countries with limited space budgets could use Leap 71’s system to generate custom engines on demand. They could select their preferred fuel type, thrust level, and form factor, and then print out the engine for use in a real-world rocket.
The global space economy is growing fast, and with it, the demand for launch systems that are affordable, efficient, and scalable. If AI-printed rocket engines live up to their promise, we could see an explosion of innovation. One that challenges even SpaceX’s head start.
