Elon Musk has declared that artificial intelligence will become SpaceX's dominant business, projecting that AI revenue will exceed all other company segments combined within months. In an all-hands meeting with employees on August 11, Musk outlined an ambitious vision for the aerospace company's future, one centered not on rockets or satellites, but on computing infrastructure and large language models (LLMs), which are AI systems trained on vast amounts of text data to generate human-like responses.

How Is SpaceX Transforming Into an AI Company?

SpaceX's pivot toward artificial intelligence accelerated dramatically after the company acquired xAI, Elon Musk's AI startup, in February 2026 for $250 billion in an all-stock transaction. This deal gave SpaceX access to Grok, xAI's frontier large language model, and to massive computing hardware resources including the Colossus supercomputing facilities, which contain over 1 million Nvidia H100 graphics processing units (GPUs), the specialized chips used to train and run AI models.

The company is now leveraging these assets to become a major player in the AI infrastructure market. SpaceX has already signed high-profile contracts with major tech companies seeking access to computing power. Anthropic, the AI safety company behind Claude, is renting roughly 325,000 Nvidia GPUs from SpaceX's Colossus data centers for $1.25 billion per month, while Google has a similar deal worth $920 million monthly. These contracts alone could generate approximately $26 billion in annual revenue under the best-case scenario.

What Are SpaceX's Specific AI Growth Targets?

Musk's projections are strikingly aggressive. He told employees that "probably, almost certainly, next month, in September, AI revenue will be larger than all of SpaceX's other business segments combined," and that "in the fourth quarter it will greatly exceed revenue from the other segments". He also stated that AI could account for 99 percent of SpaceX's total company value within the next four to five years.

To support this growth, SpaceX is rapidly expanding its computing infrastructure. The company currently operates data centers in Memphis and Mississippi with a combined power capacity of 1.4 gigawatts (GW), a unit measuring electrical power. Musk said this capacity will grow to more than 2 GW by the end of 2026, with a target of 10 GW by the end of 2027. For context, these power levels are enormous; a single gigawatt can power roughly 750,000 homes.

• Grok 4.6 Performance: The latest version of Grok, released on August 12, scored 61 on the Artificial Analysis Intelligence Index, a widely used benchmark for comparing AI models, up 5 points from the previous version and 23 points from Grok 4.3. According to Samsung Securities analyst Lee Young-jin, this performance "has risen to a level that threatens GPT-5.6-Sol," placing Grok among the frontier models alongside competitors Anthropic and OpenAI.
• Grok's Unique Training Approach: Musk emphasized that Grok will be trained on SpaceX's internal information and will "learn from" employees, who he called "the parents of the AI." He explained that "the AI will inherit your thoughts, ideas and beliefs," positioning the model as uniquely shaped by SpaceX's engineering culture and values.
• Long-Term Space-Based Computing: SpaceX is pursuing an ambitious plan to build AI computing infrastructure in space itself, using solar energy to power orbital data centers. The company is aiming for an initial demonstration of orbital AI computing in 2027.

Musk also reframed Starlink, SpaceX's satellite internet service, as part of the AI infrastructure strategy. "Starlink, too, will expand from a satellite internet service into AI infrastructure connecting billions of AI devices, self-driving cars and humanoid robots," he said. This suggests that Starlink's future role extends far beyond providing internet to remote areas; instead, it would serve as a global network backbone for AI-powered devices and autonomous systems.

What Are the Risks and Capital Requirements?

While the AI opportunity is substantial, SpaceX faces significant capital demands. The company is working with Tesla, Musk's electric vehicle manufacturer, to build a massive semiconductor manufacturing facility called Terafab, designed to eventually produce 1 terawatt (TW) of AI compute capacity per year, more than the current global supply. The first phase of Terafab requires $16.8 billion in capital investment, with total capex potentially reaching $119 billion if all planned phases are completed.

SpaceX's stock performance reflects investor caution about these ambitious plans. After its initial public offering in mid-2026, SpaceX shares have fallen 41 percent from their all-time high of roughly $226, and now trade below their opening price. The company's price-to-sales ratio has declined from roughly 116 in June to 61 today, though this still appears elevated compared to the S&P 500's average of 3.8.

"Assessments of its coding performance are mixed, but it has risen to a level that threatens GPT-5.6-Sol. This is the moment it fully joins the frontier race alongside Anthropic and OpenAI," stated Lee Young-jin, analyst at Samsung Securities.

Lee Young-jin, Analyst at Samsung Securities

The success or failure of Terafab will likely determine whether Musk's vision becomes reality. If the facility succeeds in producing chips at scale, SpaceX and Tesla could control a significant portion of the AI hardware market. If it fails or faces delays, the companies will have committed enormous capital to a project that may not deliver returns for years. Additionally, the companies currently renting SpaceX's computing capacity, including Google and Anthropic, are designing their own custom chips, which could eventually reduce their dependence on external providers and compress margins across the industry.

Despite these challenges, Musk's message to SpaceX employees was unambiguous: the future belongs to AI and robots, and SpaceX must win in both hardware and software to remain relevant. Whether the company can execute on this vision while managing the capital intensity of its ambitions remains an open question for investors and industry observers.