Musk has just revealed that xAI's entire codebase was stolen.

Today, xAI filed a lawsuit against a former employee, alleging theft of trade secrets.

According to xAI, this employee has since joined OpenAI.

Following previous incidents of former researchers threatening Meta and new employees stealing secrets from xAI, internet users are questioning why OpenAI seems to be consistently involved in such controversies.

While xAI has not named OpenAI as a defendant in this case, this incident of Altman poaching talent from Musk has certainly created complications.

To date, neither the defendant employee nor OpenAI has provided any comment.

Notably, just before the theft incident occurred, this former employee cashed out his xAI equity holdings, accumulating nearly $7 million.

The defendant, who happens to be one of the most closely watched Chinese professionals in Silicon Valley's AI talent war, adds another layer to this controversy.

**Former Employee Stole Entire xAI Codebase**

The former Chinese employee in question is Xuechen Li. According to the lawsuit filed by xAI in the U.S. District Court for the Northern District of California, xAI has brought four charges against him.

The charges include breach of confidentiality agreement, misappropriation of trade secrets, violation of California computer data regulations, and fraud. The company seeks damages and injunctive relief preventing him from joining competitors like OpenAI, as well as demanding that Xuechen Li surrender all related devices and accounts.

However, all four charges relate to the single act of Li's theft of confidential information.

Former Amazon Web Services and Google executive Bindu Reddy posted on social media that Xuechen Li stole Grok's entire codebase. Musk responded below, clarifying that it wasn't just Grok, but the entire xAI codebase, and that Xuechen Li has admitted to it with logs serving as evidence.

According to xAI's disclosed timeline, Xuechen Li resigned from xAI on July 28. Three days prior to his resignation, he uploaded xAI-related data to his personal systems.

Before leaving xAI, he also cashed out $2.2 million worth of his shares, adding to the $4.7 million he had cashed out earlier in the summer, totaling nearly $7 million.

After resigning, Xuechen Li signed a document stating he had returned company property, deleted all copies, and would maintain confidentiality of information.

Xuechen Li took extensive measures to cover his tracks, including deleting browser history and system logs, renaming and compressing files.

On August 11, during a routine security software log review, xAI discovered signs of data exfiltration and immediately sent a letter to Xuechen Li, pointing out they had discovered his illegal exfiltration of company confidential data and demanding he return and delete all related information.

After receiving the letter, instead of cooperating, Xuechen Li immediately changed the passwords to key accounts used to store the stolen data, attempting to prevent the company from accessing and recovering this information.

On August 14, Xuechen Li admitted to his misconduct both verbally and in handwritten documentation, with his attorney present throughout the process.

However, after Xuechen Li's August 14 admission, xAI discovered additional stolen materials he had not disclosed.

On August 18, Xuechen Li signed an "authorization access agreement," providing account credentials that he promised not to change before September 15, and agreed to allow xAI's forensic experts to access his personal devices and accounts to investigate and delete the company's confidential information.

However, when xAI attempted to use the credentials provided by Xuechen Li, they found he had not provided passwords to key accounts and had concealed the existence of multiple other accounts that might also contain xAI's confidential information.

When xAI questioned these omitted accounts and passwords, Xuechen Li's attorney responded that he "couldn't remember" the passwords, despite these accounts being recently used.

Reportedly, Xuechen Li had already received an offer from OpenAI before leaving and joined OpenAI on August 19.

According to xAI's lawsuit, the trade secrets Xuechen Li allegedly stole include "cutting-edge artificial intelligence technology that outperforms ChatGPT and other competitors," technology that "could save OpenAI and other competitors billions of dollars in R&D costs and years of effort."

**Who is Xuechen Li?**

Xuechen Li was already a veteran employee at xAI.

He graduated from Stanford last year with a PhD in Computer Science and joined xAI in February of that year, having previously interned at Google and Microsoft.

At that time, xAI had only been established for about six months, making Xuechen Li one of xAI's early team of approximately 20 engineers. He was responsible for developing and training Grok and had access to the entire technology stack.

Currently, Xuechen Li's LinkedIn profile only shows that he has left xAI, though he has not yet added OpenAI to his resume.

His latest post was on July 11, after the release of Grok 4, where he expressed admiration for xAI's significant achievements in such a short time.

This was only half a month before the theft incident that xAI disclosed.

Additionally, Xuechen Li was one of the core contributors to the Alpaca series of large language models.

Such background has left internet users puzzled about why Xuechen Li would take such enormous risks to commit theft.

As for OpenAI's role in this incident, whether the theft was Xuechen Li's personal action or orchestrated by OpenAI, remains to be seen...