The AI race has kicked off a new gold rush for training data, and companies are leaving no stone unturned in collecting as much as they can.
According to a Bloomberg report, the Elon Musk-led SpaceX is currently pursuing customer and operational data from defunct and struggling startups. The discussions are still informal and may not end up in confirmed deals, sources say. However, if they go through, the company will gain access to vital AI training data at affordable rates.
The approach is similar to what Google did with Spirit Airlines. To train its Gemini models, the search engine giant purchased Spirit’s business data for $10 million from a bankruptcy auction last month. The move immediately triggered privacy concerns as the acquired data included employee conversations from Microsoft Teams as well as millions of other company emails.
For SpaceX, the data acquisition could be about training Grok AI models, which it owns through its subsidiary xAI. The company already uses your X posts to train Grok models unless you opt out, but these models also need high-quality real-world information to improve their outputs and use cases. Currently, Grok lags behind competing models from Anthropic, OpenAI, and even Meta across multiple metrics. Given the circumstances, it’s no surprise that SpaceX spent nearly $15.8 billion in the previous quarter on AI developments.
xAI’s current flagship model is Grok 4.6, and the company plans to release Grok 4.7 this month, with Grok 5 launching later this year. Musk has high expectations for Grok 5.
“With Grok 5, which should be out before the end of this year, we will be incorporating the entire corpus of SpaceX data,” he said during SpaceX’s first earnings call last month. “Basically, all the data that SpaceX has ever produced, which is a tremendous amount over the course of a quarter century, will be incorporated into Grok training. We think this will make Grok by far the best engineer.”