Elon Musk is responding to a series of resignations from xAI, among them two additional co-founders who departed this week, pushing the total to six out of the initial twelve.
During an all-staff gathering held on Tuesday evening, Musk implied that the departures stemmed from compatibility issues rather than performance. “Having achieved a particular magnitude, we are structuring the organization to operate with greater efficacy at this level,” he stated, as reported by The New York Times. “Indeed, in such circumstances, certain individuals are more appropriate for a company’s nascent phases and less so for its subsequent growth periods.”
On Wednesday afternoon via X, he elaborated, clarifying that these exits were not self-initiated. “xAI underwent a restructuring a few days prior to enhance execution pace,” Musk penned. “When an enterprise expands, particularly at the rapid rate of xAI, its framework must transform akin to any biological entity. Regrettably, this necessitated severing ties with certain personnel.”
He further stated that the firm is “recruiting vigorously” and concluded with a characteristically Muskian proposition: “Become part of xAI if the concept of mass accelerators on the Moon captivates you.”
The relinquishing of half one’s co-founders within a comparatively brief timeframe sparks inquiries, and Musk’s statements appear orchestrated to manage the discourse, presenting the departures as imperative instead of an issue for the organization.
Overall, a minimum of 11 engineers, encompassing the two co-founders, have openly declared their resignation from xAI over the last week — although a couple of those departures seem to have transpired several weeks earlier.
Three of the outgoing personnel have indicated their intention to commence a fresh endeavor alongside other ex-xAI engineers, despite the absence of information concerning this novel undertaking. Other individuals have alluded to a yearning for increased self-governance and more compact groups to develop cutting-edge technology with greater expedition, citing the foreseen spike in AI output.
Techcrunch event
Boston, MA
|
June 23, 2026
Yuhuai (Tony) Wu, an xAI co-founder and head of reasoning, stated in a message publicizing his departure: “My subsequent phase is now due. This is a period replete with opportunities: a compact group equipped with AIs possesses the capability to achieve monumental feats and re-establish the bounds of possibility.”
Shayan Salehian, who contributed to product infrastructure and model conduct after training at xAI and formerly served at Twitter/X, disclosed last week his intention to depart for “a fresh beginning.”
Vahid Kazemi, who experienced a short period contributing to machine learning, declared on Tuesday that he had resigned some weeks prior, appending: “In my opinion, all AI laboratories are constructing precisely the same product, and it lacks excitement… Therefore, I am embarking on a fresh project.”
Roland Gavrilescu, an ex-xAI engineer, departed in November to establish Nuraline, an enterprise developing “proactively deployed AI agents,” but announced again on Tuesday his exit from the firm to construct “a novel venture alongside others who quit xAI.”
These exits coincide with a period of considerable contention for xAI. The enterprise is encountering supervisory examination subsequent to Grok’s generation of non-consensual explicit deepfakes featuring women and minors that circulated on X — last week, French authorities conducted a raid on X’s premises as part of an inquiry. The corporation is additionally progressing toward a scheduled initial public offering later in the current year, having been lawfully procured by SpaceX in the preceding week.
Musk is likewise confronting individual contention following the release of documents by the Justice Department revealing prolonged dialogues with convicted rapist and sex trafficker Jeffrey Epstein. The electronic messages depict Musk deliberating a trip to Epstein’s isle on two distinct instances, in 2012 and 2013. Epstein was initially found guilty of soliciting a minor for prostitution in 2008.
xAI sustains a workforce exceeding 1,000 individuals, thus the resignations are improbable to impact the firm’s immediate proficiencies. Nevertheless, the swift succession of recent exits had gained considerable momentum online, with users playfully declaring on X their own “departure from xAI” despite never having been employed there — an indication of the speed with which the story of a “widespread departure” escalated on Musk’s social platform.
Yet, involuntary co-founder resignations seldom denote unproblematic growth. While Musk portrays the restructuring as deliberate, the reality that numerous engineers trailed the co-founders in leaving — and that a minimum of three are jointly launching a novel initiative — implies that the exits might additionally reveal more profound discords. Within cutting-edge AI, where expertise is rare and renown holds weight, xAI’s capacity to draw in and keep premier researchers will be scrutinized as it contends with OpenAI, Anthropic, and Google.
TechCrunch has contacted xAI seeking further details.
Chronology of resignation declarations
The subsequent personnel have openly proclaimed their resignations from xAI via X in the past few days:
February 6: Ayush Jaiswal, an engineer, penned: “This marked my final week at xAI. I intend to dedicate a few months to family and experimenting with AI.”
February 7: Shayan Salehian, who contributed to product infrastructure and model conduct after training and was formerly at X, penned: “I departed xAI to initiate a new venture, concluding my more than 7-year period of employment at Twitter, X, and xAI with immense appreciation.” He further stated that collaborating closely with Elon Musk instilled in him “an intense focus on particulars, fervent promptness, and the ability to reason from foundational precepts.”
February 9: Simon Zhai, MTS (member of technical staff), stated: “Today marks my final day at xAI, and I feel profoundly privileged regarding the prospect. It has constituted an extraordinary expedition.”
February 9: Yuhuai (Tony) Wu, co-founder and head of reasoning, penned: “I stepped down from xAI today. My subsequent phase is now due. This is a period replete with opportunities: a compact group equipped with AIs possesses the capability to achieve monumental feats and re-establish the bounds of possibility.”
February 10: Jimmy Ba, co-founder and research/safety head, stated: “Final day at xAI. We are advancing toward an epoch of 100x output with the appropriate instruments. Iterative self-enhancement cycles are likely to become operational within the forthcoming 12 months. It is time to re-evaluate my focus on the broader perspective. 2026 is poised to be wild and probably the most active (and significant)
year for the prospects of humankind.”
February 10: Vahid Kazemi, a Machine Learning PhD, disclosed his departure from xAI “several weeks earlier,” further remarking: “As I see it, all AI research facilities are creating virtually the same product, and it lacks excitement. I believe greater inventiveness is possible. Consequently, I am embarking on a novel venture.”
February 10: Hang Gao, whose work encompassed multi-modal endeavors, including Grok Imagine, posted: “My departure from xAI occurred today.” He characterized his tenure there as “genuinely fulfilling,” highlighting his input into Grok Imagine’s product launches and commending the group’s “modest skill and bold aspirations.”
February 10: Roland Gavrilescu, the engineer who left in November to establish Nuraline, announced: “My time at xAI has concluded. I am creating a novel enterprise alongside other former xAI colleagues. We are recruiting.”
February 10: Chace Lee, a member of the Macrohard foundational group, penned: “Taking a brief reset, then returning to the cutting edge.” (Macrohard represents an exclusive AI software enterprise operating under xAI, conceived to entirely automate software creation, programming, and management through Grok-driven, multi-agent frameworks. Its designation serves as a subtle criticism of Microsoft.)
February 11: Andrew Ma, who had been affiliated with xAI since X was known as Twitter, contributed to enhancements in application and recommendation models, encompassing elements such as “the X video stream, search interface, user profiling, introductory packages, and the primary content feed model.” He conveyed: “I am enthusiastic concerning what lies ahead — my precise next steps remain undetermined (my direct messages are accessible), yet there exists a planet ripe for transformation, and no moment should be squandered. Proceed with vigor, colleagues; maintain your concentration, remain spirited. I eagerly await our reunion on the moon and further, and genuinely, I assert there is no individual on Earth I more implicitly trust to achieve that, for a world awaits triumph.”
February 12: Radhakrishnan (Rad) Venkataramani, whose efforts were directed at Grok’s reasoning and reinforcement learning frameworks, communicated: “The preceding eight months within the RL systems/SWE-RL group, dedicated to advancing our programming model to achieve state-of-the-art status and move towards iterative self-enhancement, will forever stand out as the most unforgettable period of my existence … We currently stand at a pivotal moment where intellect initiates its own rapid advancement, and from this juncture, its progression will become exclusively upward.”
This piece was initially released on February 11 and has since been revised to incorporate further staff exits.
Do you possess a discreet piece of information or classified paperwork? Our focus is on exposing the internal mechanisms of the artificial intelligence sector — encompassing entities that define its trajectory and individuals affected by their choices. Contact Rebecca Bellan at rebecca.bellan@techcrunch.com, Russell Brandom at russell.brandom@techcrunch.com, or Tim Fernholz at tim.fernholz@techcrunch.com. For private correspondence, you may reach them through Signal using rebeccabellan.491, russellbrandom.49, or tim_fernholz.21.
{content}
Source: {feed_title}