NEW YORK, Jan 31, 2026, 08:29 (EST)
- Starlink revised its privacy policy, now permitting customer data to be used for training AI models unless users choose to opt out
- The policy permits sharing data with service providers and “third-party collaborators”
- SpaceX is in discussions about merging with Elon Musk’s xAI as the company prepares for a potential IPO
SpaceX’s Starlink has revised its privacy policy to permit the use of customer data for training machine learning and AI models, unless users choose to opt out. The update also gives Starlink the ability to share this data with service providers and “third-party collaborators.” Yahoo
This change comes as SpaceX considers an initial public offering—their first public share sale—later this year and negotiates a merger with Elon Musk’s AI firm xAI, raising questions about the data Starlink gathers and its potential uses.
AI systems rely on data to identify patterns. Starlink reports serving over 9 million users, making its policy update relevant to a vast number of customers.
Starlink’s Global Privacy Policy, last updated on Jan. 15, states the company may use personal information “to train our machine learning or artificial intelligence models” as outlined in the document. Starlink
The sharing section states that Starlink can share personal data with affiliates, service providers, and “third-party collaborators.” Unless customers opt out, their information might be used for AI model training, including by those collaborators for their own separate purposes. Starlink doesn’t disclose who these collaborators are, instead advising users to opt out via their account settings.
The policy outlines the types of data it gathers, covering location info characterized as broad rather than exact, payment card information, contact details, and IP addresses. It also covers technical stats like throughput, connectivity, and WiFi performance, along with “communication information” that may involve audio or visual content, user-uploaded files, and inferred data.
The document doesn’t specify which parts of that data will feed into AI training. It’s also vague on how much training happens within Starlink compared to what’s handled by external partners.
“It definitely raises an eyebrow and would worry me if I relied on Starlink,” said Anupam Chander, a Georgetown University professor specializing in technology law. “There are often valid reasons for using your data, but here, there’s no clear boundary on how it might be applied.” Reuters
Musk’s xAI was valued at $230 billion following its latest funding round, the report noted. The company is working on its Grok chatbot, which uses large language model (LLM) technology for text generation, and also owns X.
Starlink operates a constellation of over 9,000 satellites and claims to have more than 9 million users. Partnering with xAI might grant the AI firm access to fresh data sets, including communications data, while enabling SpaceX to expand its AI-powered offerings.
Amazon’s Project Kuiper, OneWeb, and Viasat are all ramping up their satellite broadband services. But Starlink’s size gives it a chance to shape industry standards when it comes to data usage policies.
Broad sharing language tends to draw scrutiny, and if customers start opting out, the amount of data available for training could shrink significantly. SpaceX has yet to comment on the matter.
