Musk says SpaceX’s AI revenue could overtake its other businesses as early as September

“Probably our AI revenue — not probably, definitely — our AI revenue will exceed all other SpaceX revenue probably in September, like next month,” Musk said in his speech.

“And will significantly exceed all other SpaceX revenue in the fourth quarter,” he said. SpaceX is no longer looking at artificial intelligence as a side business. Musk now sees AI infrastructure and computing as a central part of SpaceX’s future.

AI could soon become SpaceX’s biggest business

SpaceX has traditionally made money from launching rockets, building spacecraft and operating Starlink, its satellite internet network. But AI is now growing at a much faster pace.

The company’s second-quarter numbers show how quickly that change is happening. SpaceX reported $7.81 billion in total revenue for the quarter, with $2.56 billion coming from AI. Connectivity products, including Starlink, brought in $4.29 billion, while space products generated $962 million.

That means AI already accounted for a significant part of SpaceX’s business in the second quarter. Musk’s September prediction suggests that he expects this part of the company to grow rapidly over the next few months.

The company also reported $14.1 billion in contracted cloud services agreements, including deals with companies such as Anthropic and Google. These agreements are part of SpaceX’s effort to build a large AI computing business.

If Musk’s prediction comes true, SpaceX could soon have a very different revenue structure from the one most people associate with the company.

From rockets to AI data centres

Musk’s argument is that SpaceX has an advantage that traditional data centre companies do not have: its engineers. SpaceX is investing heavily in AI infrastructure, including large data centres designed to provide the computing power needed to train and run AI models. Musk believes the engineering skills used to build rockets and satellites can be applied to this new business. After spending nearly $16 billion on AI infrastructure in the second quarter, Musk defended the spending during SpaceX’s first earnings call as a public company.

Data centres are a “trivial problem” compared with what SpaceX’s rocket engineers work on every day, Musk said. “We’re taking a small amount of the expertise that we use for rockets and satellites, and applying that to scaling terrestrial data centers,” he said, arguing that even a fraction of SpaceX’s aerospace engineering experience produces “an amazing outcome” when applied to AI infrastructure.

Musk compared SpaceX’s position in the AI infrastructure market to “the New York Yankees going in and playing a Little League team.” The comparison was aimed at established data centre and cloud companies, although Musk did not name any specific competitors. Amazon, Microsoft and Google already operate some of the world’s biggest cloud businesses, while a growing number of AI-focused data centre companies are also trying to build the computing capacity needed for the AI boom.

Musk believes SpaceX can move faster because the company is already used to solving extremely difficult engineering problems. “Rocket science is literally our daily business,” Musk said. He argued that SpaceX engineers spend their days designing rockets that have to survive extreme conditions, while data centres are a much simpler engineering challenge.

SpaceX is spending heavily to build AI capacity

The scale of SpaceX’s AI ambitions can also be seen in its capital spending. The company’s total capital expenditure reached $18.4 billion in the second quarter, compared with $10 billion in the first quarter. SpaceX executives said during the earnings call that spending is expected to remain around the current level for the next two quarters.

The company ended the quarter with 1.4 gigawatts of nameplate computing capacity and expects to cross two gigawatts by the end of the year. But Musk has a much bigger target in mind.

He said SpaceX has a tentative goal of having 20 gigawatts of power and cooling capacity online by the end of 2027. Even if the company faces delays, he said it could still reach close to 15 gigawatts. That would put SpaceX among the biggest players in the global race to build AI computing infrastructure.

Nvidia chips will power the AI push

A major part of this strategy depends on Nvidia’s advanced AI chips. Musk said SpaceX plans to build its AI computing platform exclusively around Nvidia’s Vera Rubin architecture. He described it as the best AI computing system available and said SpaceX has a close relationship with Nvidia.

“Our understanding with Nvidia is that we will receive a very significant percentage of their GPUs next year,” Musk said. Access to large quantities of advanced GPUs could become an important advantage as companies around the world compete for AI computing power. SpaceX is effectively trying to build the infrastructure that allows other companies to train and run their AI models, while also using that infrastructure for its own AI projects. Musk wants Grok trained on SpaceX

SpaceX’s AI ambitions go beyond data centres

Musk also wants the company’s AI models to learn from the enormous amount of information generated inside SpaceX. “We’re going to be training Grok on the sum total of all SpaceX information,” he said. “So in a way, it’ll be trained on you.”

He said SpaceX employees would be “the parents of the AI,” and it would inherit their thoughts, ideas, and beliefs. Musk did not explain exactly how this information would be collected or how employees would be involved in the training process.

That question could become important as companies increasingly look at their employees and internal information as sources of training data for AI systems. Meta, for example, introduced a programme in April that involved using employees’ keystrokes to train its AI models. The programme faced significant backlash and was later paused. The shift towards AI is part of a much larger transformation at SpaceX.

The company acquired xAI, another Musk-owned AI company, and has since pushed aggressively into artificial intelligence. It is also exploring the idea of putting data centres in space, where they could use solar power to run AI computing systems. Musk has therefore begun building a business that stretches far beyond rockets and satellites.

SpaceX still operates Falcon 9 and Starship launch vehicles, Crew Dragon and Cargo Dragon spacecraft and Starlink. But AI is increasingly becoming the centre of its growth plans.

The company has also been expanding its cloud business, signing large contracts with companies that need computing capacity for artificial intelligence. That makes Musk’s September prediction particularly significant. He is not simply saying that AI will become another important business for SpaceX. He is saying it could soon become the company’s largest source of revenue.

The September prediction will be a major test

Musk’s claim that AI revenue will overtake all other SpaceX revenue in September sets a very short timeline for the company’s transformation.

SpaceX has already shown that AI can become a multibillion-dollar business for the company. But becoming its largest revenue source would represent a much bigger shift.

The company is spending enormous amounts of money to build computing infrastructure, secure Nvidia chips and expand its cloud business. It is also developing its own AI capabilities through Grok and exploring data centres in space.

For Musk, the logic is straightforward. SpaceX already has the engineers, infrastructure and capital needed to build at a scale that could allow it to compete with much larger technology companies. Whether that strategy works will depend on how quickly SpaceX can turn its massive AI investment into actual revenue.