Elon Musk announced a strategic move: SpaceX engineering data will be used to train the next version of Grok. According to Musk, this decision will directly enhance the model's engineering capabilities, opening new horizons for xAI.

On July 21, the billionaire reported on platform X that SpaceX will transfer its engineering developments to train a new model with a 2 trillion parameter architecture. This is nearly double the 1.5 trillion parameters in the V9 architecture on which the current public release of Grok 4.5 is based.

Limitations and Data Scale

An important nuance: some data related to defense and space technologies falls under U.S. export control (ITAR). This information will not be included in the dataset — SpaceX operates strictly within the law. Nevertheless, the array of data on rocket and satellite developments accumulated over the years of the company's rapid growth will become a unique source for Grok. Recall that this year SpaceX successfully conducted a record public offering of SPCX shares and raised fresh capital for engineering and AI projects.

Developer Opinions Divided

Musk responded to a post from a developer who was simultaneously testing Grok Build and OpenAI's Codex. According to the test author's assessment, Grok Build currently outperforms Claude on complex programming tasks, although Codex also holds its ground confidently. However, this assessment came after a series of setbacks: previously, reviewers pointed to contradictory results from Grok 4.5, even though Musk compared the model to the Opus class — Anthropic's top-tier Claude lineup.

SpaceX Expands AI Projects

The data transfer coincided with investments of $60 billion in developer tools — including the acquisition of Cursor's parent company a few weeks after its IPO. Separately, Musk set a goal to bring SpaceX's revenue to one trillion dollars by 2030.

In parallel, xAI is expanding Grok's capabilities. On the chatbot's official page, Grok for Excel was announced: users will be able to build financial models, analyze market data, and work with charts based on Grok 4.5. This shows that xAI is building new products on the same model that will soon begin training on SpaceX engineering data. Exclusive information about rockets and satellites is being directly integrated into Grok's development plans.

Analytical Conclusion: xAI is gaining a set of engineering data that competitors like OpenAI and Anthropic lack. Whether the new 2 trillion parameter model will catch up with Claude and Codex will be shown by real tests, not Elon Musk's statements. However, the very fact of integrating space developments into an AI model is a powerful signal to the market: the boundaries between the aerospace and AI industries are blurring before our eyes, and xAI is gaining a unique competitive advantage.