[DigitalToday reporter Jinju Hong (홍진주)] SpaceX is shifting its business structure from a rocket development company to an infrastructure company centered on Starlink and artificial intelligence (AI). It has broadened its business beyond launch vehicle development to communications networks, data processing and automation in production and inspections, and has set a long-term goal of building space infrastructure connecting the moon and Mars.

IT outlet Ars Technica reported on Nov. 9 local time that Kiko Dontchev, SpaceX's vice president of launch, said at the recent UP.Summit in Florida that the company's business direction is rapidly changing to focus on Starlink, data and AI.

Dontchev called SpaceX "still a rocket company, but now fundamentally an infrastructure company," saying, "We are building infrastructure on Earth and in low Earth orbit, and will soon do so on the moon and ultimately on Mars." The idea is to go beyond developing launch vehicles and sending them to space and evolve into a communications and data infrastructure operator using space.

AI is one of the main drivers of the change. After acquiring Elon Musk's other company xAI earlier this year, SpaceX is actively introducing AI tools into internal work. Their use is expanding beyond engineering to production, field operations and quality inspections.

Dontchev stressed that "the pace of AI adoption has become extremely fast this year," adding that the acceleration in engineering execution was unprecedented compared with what he has seen in 16 years at the company.

SpaceX is using the AI assistant Grok to detect foreign substances mixed into aerospace components. Foreign substances can cause serious damage to engines or spacecraft parts, requiring precise inspections. Dontchev said human accuracy was about 94 percent when checking the same images, while Grok's detection accuracy reached 99.9 percent.

AI automation is also being applied to visual inspections of hardware such as Starlink customer terminals. SpaceX is using AI not as a simple work aid but as a core operating system to improve productivity and inspection accuracy. Dontchev explained that as technicians and field managers use AI routinely, capability gaps among staff are narrowing and the pace of work growth is speeding up.

Its launch vehicle strategy is also set for a major shift. Falcon 9, SpaceX's core launch vehicle over the past 10 years, is gradually moving toward retirement, and Starship, developed with the goal of full reusability, is expected to replace it.

Dontchev pointed out that Falcon 9 is "not a fully reusable launch vehicle" and that it "throws away the second stage on every flight." Starship, by contrast, aims to recover and reuse both the first-stage booster and the second-stage spacecraft. With SpaceX pursuing thousands of launches a year and, in the long term, about 10,000 launches, it has judged that it will be difficult to meet that goal if it must build a new second stage each time.

Starship's competitiveness is also evident in communications capacity per launch, not just payload. Starship's first orbital flight in late September carried large Starlink V3 satellites. SpaceX said that a single mission carried 10 times more bandwidth than the Starlink satellite payload carried by Falcon 9.

Launching from Florida can further raise efficiency by using a route more favorable for certain orbits. Dontchev said that, based on Florida launches, a single Starship flight could carry up to 20 times more bandwidth than a single Falcon 9 launch.

SpaceX is also expanding launch infrastructure for this. Pad infrastructure for the first Starship launch tower to be built at Kennedy Space Center's Launch Complex 39A is nearing completion, and SpaceX is also developing two additional launch pads in Florida.

Dontchev said the 39A pad was virtually complete and that SpaceX plans to make its first attempt this December to transport the booster and spacecraft by barge to the Cape. If preparations proceed smoothly, it also plans to pursue a first flight by year-end.

How it will accommodate existing commercial launch customers during the transition to Starship remains a task. SpaceX is optimizing Starship for its own needs, but it did not specify how it will handle launch demand from private customers that have used Falcon 9. It has been suggested that, over the next few years, Starship launch capacity could be allocated first to building SpaceX's own satellite network and to missions for the U.S. National Aeronautics and Space Administration (NASA) and government customers.

Ultimately, SpaceX's changes go beyond a simple rocket replacement. It is a strategy to boost launch cost and transport efficiency with Starship, expand space communications networks through Starlink, and automate production and operations with AI. Its move to shift from a company that builds rockets to one that directly builds and operates space-based communications and data infrastructure is gaining momentum.