On May 6, local time, Elon Musk announced on social media that xAI will no longer exist as an independent company and will be integrated into SpaceX, becoming its AI product line and renamed SpaceXAI.
Musk’s statement was straightforward: “xAI will no longer exist as an independent company, it will simply be SpaceXAI, the AI product of SpaceX.” This statement directly responds to market speculation about xAI’s independent status and marks the end of a nearly three-month integration process, with Musk’s “space + AI” vertical integration strategy entering a new phase.
Musk’s statement was in response to a post by Tesla investor @SawyerMerritt about SpaceX’s agreement with Anthropic.
Just hours earlier, SpaceXAI had partnered with AI company Anthropic, allowing it to use SpaceXAI’s supercomputer Colossus 1’s 300-megawatt computing resources, and the two companies expressed their intention to cooperate on the development of gigawatt-scale orbital AI computing power.
Musk also stated that SpaceX will provide computing resources to other AI companies, on the condition that they use their own models to benefit all humanity.
The roots of xAI’s creation lie in Musk’s complete split with OpenAI. In 2015, Musk co-founded OpenAI with Sam Altman, Peter Thiel, and others, with the initial intention of creating an open-source, non-profit AI institution to counterbalance tech giants like Google.
In the early days, Musk not only invested funds but also deeply participated in strategic decision-making, but conflicts soon arose over control and development philosophy. In February 2018, Musk left OpenAI’s board of directors and terminated funding. After that, OpenAI gradually leaned towards Microsoft, accepting its massive investment and transitioning to a closed-source, for-profit model, which Musk believed betrayed the original intention.
In July 2023, Musk officially announced the establishment of xAI. This AI company, founded by Musk, gathered core AI talents from top institutions such as OpenAI, Google DeepMind, and Microsoft Research, with an initial positioning as a public-benefit enterprise, and its core mission was to “understand the essence of the universe.”
Over two years, xAI rapidly completed its technological and capital accumulation. In March 2024, it launched its first large model, Grok-1, and later iterated to Grok-4, with multimodal reasoning capabilities; in September of the same year, it built the Colossus training system equipped with 100,000 NVIDIA H100 graphics cards, and in early 2026, it was further upgraded to Colossus 2, with 550,000 NVIDIA GB200 GPUs, becoming the world’s first single-body AI training cluster with gigawatt-level computing power.
In terms of capital, xAI completed multiple rounds of financing, with a total financing amount exceeding $42 billion, and its valuation rose to $230 billion after the latest round of financing in January 2026. Investors included NVIDIA, Cisco, and other leading tech companies.
As early as February this year, SpaceX announced that it would acquire xAI through an all-stock transaction, with the transaction valuing SpaceX at around $1 trillion and xAI at $250 billion, and the combined entity’s valuation reaching $1.25 trillion, setting a record for the largest corporate acquisition in history.
Regarding the core strategic logic of this integration, Musk has repeatedly publicly stated that the core goal is the ultimate layout of “space computing power.” He explicitly pointed out that the Earth’s energy and heat dissipation limits will soon constrain AI computing power development, and predicted that the minimum cost of generative AI computing will shift to space within two to three years. In Musk’s plan, SpaceXAI’s core task is not just to compete with OpenAI and Google in the large model market, but to become the “space brain” of SpaceX’s space exploration.
This integration also completed the “data – model – computing power – communication – transportation” closed loop of Musk’s core assets.
On the data side, the social media platform X generates over 500 million real-time human behavior data per day, providing continuous training support for the Grok model; on the model side, SpaceXAI will focus on four core businesses: Grok conversational AI, programming, image and video generation, and digital simulation; on the computing power side, it will build the Colossus series of supercomputers and layout space computing power clusters; on the communication side, Starlink will provide global network coverage and data transmission; and on the transportation side, SpaceX rockets will be responsible for sending computing devices into space, forming an unreplicable competitive barrier.
However, the integrated SpaceXAI still faces many challenges. On the one hand, xAI has seen the departure of its founding team, and Musk has rarely acknowledged that xAI is “significantly behind its competitors.” How to quickly improve technical competitiveness and catch up with industry giants like OpenAI and Anthropic after integration is a core problem facing SpaceXAI. On the other hand, the layout of space computing power faces multiple challenges, including high technical difficulty, high investment costs, and strict policy regulation.
SpaceX plans to launch an IPO in the middle of this year, with a valuation target of $1.75 trillion, and the integration of SpaceXAI will be an important support point for its valuation. The commercialization prospects of space computing power and AI technology will directly affect the capital market’s valuation expectations for SpaceX. At the same time, xAI shareholders will become SpaceX shareholders, sharing the benefits of the integration of space and AI development, but also bearing the development risks after integration.
The feud between Musk and OpenAI has not ended, but has escalated into a century-long lawsuit. In February 2024, Musk formally sued OpenAI, CEO Altman, and President Brockman in the California Superior Court, accusing them of breaching non-profit commitments and violating trust obligations, and seeking $150 billion in damages.
On April 28, 2026, local time, the case was officially heard in the US Northern California Federal District Court, entering the witness testimony phase, with Musk, Altman, Brockman, and Microsoft CEO Nadella all testifying, and the trial is expected to last three weeks.
This lawsuit is seen as the most influential AI ethics and business dispute case in Silicon Valley history, and the outcome may reshape the global AI industry rules, directly affecting OpenAI’s planned IPO process this year.
As a leading company in the global AI industry, OpenAI aims to complete its IPO in the fourth quarter of 2026, with a valuation target of $852 billion, planning to raise over $60 billion, and if successful, it will become the highest-valued unicorn company in the global AI sector and one of the largest IPOs in tech history.
