Elon Musk Sets $1 Trillion Revenue Goal for SpaceX by 2030 as Starlink and AI Compute Drive Growth
Elon Musk used SpaceX’s first earnings call as a public company to deliver a bold message: the company is aiming for $1 trillion in annual revenue by 2030, and he made it clear that he was referring to actual revenue, not annual recurring revenue.
The call followed a strong second-quarter 2026 earnings report that showed SpaceX growing rapidly across its space, connectivity, and AI infrastructure businesses. The company reported revenue of $7.8 billion for the quarter, beating analyst expectations of $6.93 billion and marking a 92% year-over-year increase.
Starlink remained the biggest contributor, generating $4.3 billion in revenue, up 66% from the same period last year. SpaceXAI, the company’s fast-expanding artificial intelligence and compute infrastructure unit, brought in $2.6 billion. The traditional space segment added $962 million, growing 29% year over year.
Despite the impressive revenue surge, SpaceX is still not profitable on a net basis. The company posted a net loss of $541 million, though that was significantly better than the expected loss of nearly $1.94 billion. EBITDA came in at $3.5 billion, far above estimates of $2.03 billion and up 191% year over year. Earnings per share landed at a loss of $0.09, better than the expected loss of $0.26.
While the earnings numbers were strong, the biggest headlines came from Musk’s comments on SpaceX’s AI ambitions, satellite computing plans, mobile connectivity strategy, and Starship launch goals.
One of the most important announcements was SpaceX’s commitment to using NVIDIA GPUs for its future AI infrastructure. Musk said the company will rely on NVIDIA’s Vera Rubin platform to power upcoming data centers, signaling a major push into large-scale AI compute.
SpaceX’s AI1 satellites are expected to use NVIDIA’s Space-1 Vera Rubin module, which includes four Rubin GPUs, two Vera CPUs, and a large memory configuration. These satellites are designed to support up to 150 kilowatts of peak compute payload. They will also include liquid radiators, meteorite shielding, centralized compute systems, and deployable solar arrays.
The satellites are set to be manufactured at SpaceX’s Gigasat facility in Texas, giving the company a dedicated production base for its next generation of AI-focused space hardware.
Musk said SpaceX plans to bring roughly 2 gigawatts of compute capacity online by the end of 2026. By the end of 2027, the company is targeting a much larger footprint, with Musk describing the goal as “closer to 10 than 5” gigawatts of compute. The broader power and cooling load could reach around 20 gigawatts.
That scale would place SpaceX among the most aggressive players in the AI infrastructure race, especially as demand for compute continues to climb across cloud services, AI model training, and inference workloads.
SpaceXAI has already signed multiple cloud computing agreements, and according to the company’s finance chief, those deals have a payback period of less than one year. The company reportedly contracted $6.7 billion worth of cloud deals in the first weeks of the third quarter alone.
A key part of the strategy is renting out unused compute capacity. Musk noted that only about 10% of SpaceXAI’s compute resources are currently used to train newer versions of Grok, the AI model associated with Musk’s broader technology ecosystem. The remaining capacity is either used for inference or made available to customers through cloud contracts.
Musk also highlighted what he sees as one of the biggest obstacles facing the entire AI industry: memory supply.
According to Musk, memory production is increasing by roughly 20% per year, which would normally be considered exceptional growth for a mature industry. However, he argued that AI demand is rising far faster, possibly around 200% per year or more. In his view, that mismatch between supply and demand means memory prices are more likely to rise than fall.
“Limiting factor currently is memory,” Musk said, explaining that the speed of demand growth is far outpacing the industry’s ability to increase supply.
This comment reflects a growing concern across the AI sector. GPUs receive much of the attention, but high-bandwidth memory and other advanced memory components have become critical bottlenecks for scaling AI systems. If SpaceXAI is building gigawatt-scale compute capacity, access to memory could become just as important as access to GPUs.
Musk’s revenue outlook was another major moment from the call. He said SpaceX is now aiming to reach $1 trillion in revenue by 2030, one year earlier than previously expected. He also said the company expects to exceed $100 billion in annual recurring revenue by December 2026.
The distinction is important. Annual recurring revenue measures contracted or repeatable revenue streams, often used by cloud and software companies. Musk emphasized that the $1 trillion target refers to full company revenue, suggesting SpaceX expects massive growth not only from Starlink subscriptions, but also from AI compute services, cloud deals, satellite infrastructure, mobile connectivity, launch services, and future Starship operations.
Starlink Mobile was another major focus. Musk made clear that SpaceX wants to challenge the largest U.S. wireless carriers with a satellite-powered mobile service. However, SpaceX still faces major hurdles, especially when it comes to access to terrestrial spectrum in dense urban areas.
Because the major carriers have not agreed to a mobile virtual network operator arrangement with SpaceX, the company is exploring another route. Musk’s plan involves using small, low-power cellular base stations placed on existing Starlink dishes. This femtocell-style approach could help SpaceX expand mobile coverage without immediately needing to build a traditional nationwide tower network or spend heavily on spectrum licenses.
If successful, this could reshape the mobile connectivity market by allowing Starlink users to access cellular-style service through satellite-linked equipment. The approach could be especially disruptive in rural and underserved areas, where traditional wireless networks are often expensive to expand.
Musk also addressed SpaceX’s launch ambitions, setting an aggressive target for Starship. He said the company is aiming for a flight cadence of one Starship launch per day by 2027. That would represent a dramatic leap in reusable rocket operations and would be central to SpaceX’s long-term plans for satellite deployment, lunar missions, Mars transportation, and large-scale space infrastructure.
The company is also targeting a return to the Moon, with crewed lunar activity planned for 2028.
Taken together, SpaceX’s latest earnings call painted a picture of a company trying to become much more than a rocket launch provider. Starlink is already a major revenue engine, SpaceXAI is emerging as a serious cloud and AI compute business, and Starship remains the backbone of Musk’s long-term space ambitions.
The company is still operating at a net loss, but its revenue growth, improving margins, expanding cloud contracts, and ambitious infrastructure plans suggest investors are looking closely at its path toward profitability.
Musk’s $1 trillion revenue target may sound extreme, but the strategy behind it is becoming clearer: combine satellite internet, mobile connectivity, AI compute, cloud infrastructure, reusable rockets, and space-based data centers into one vertically integrated technology platform.
If SpaceX can execute on even a portion of that roadmap, the company could become one of the most influential players in both the space economy and the artificial intelligence infrastructure race by the end of the decade.
