Elon Musk has done something you almost never see in a high‑stakes AI lab: he now stands alone at the top, with every single co‑founder of xAI reportedly gone just as the company is being rebuilt and folded under SpaceX. For anyone watching the race to build frontier models, this is more than just executive gossip. It raises hard questions about governance, talent retention and whether xAI is a serious long‑term competitor to OpenAI, Anthropic and Google — or simply another instrument in Musk’s personal empire. This piece looks at what the shake‑up really means.
The news in brief
According to TechCrunch, citing reporting from Business Insider, the last two remaining co‑founders of Elon Musk’s AI startup xAI have now left the company.
Business Insider first reported that Manuel Kroiss, who led xAI’s pre‑training team and reported directly to Musk, told contacts he was leaving. Shortly afterwards, it reported that Ross Nordeen, described as Musk’s key operations lieutenant and previously involved in planning layoffs at Twitter after the 2022 acquisition, had also departed.
Earlier in March, nine of xAI’s original eleven co‑founders were already reported to have left. Musk recently acknowledged on X that xAI “was not built right” initially and said it is being rebuilt from the ground up.
The exits come just after xAI was acquired by SpaceX, bringing SpaceX, xAI and X (formerly Twitter) under one corporate umbrella, at a time when SpaceX is reportedly preparing for a public listing. TechCrunch notes that xAI did not respond to a request for comment at the time of publication.
Why this matters
Founder churn happens at startups. A total founder wipe‑out at a still‑young frontier‑model lab is something else entirely.
For Musk, this consolidation removes any internal counterweights. With Kroiss and Nordeen gone, there are no remaining co‑founders who can challenge strategic decisions, slow down risky bets or make the case for a different technical roadmap. xAI becomes even more of a one‑man project, culturally and operationally.
That has three immediate consequences:
Talent attraction becomes harder. Top AI researchers already have their pick of employers: OpenAI, Anthropic, Google DeepMind, Meta, plus well‑funded independents. Joining a lab where every founding leader has recently left — and where the owner publicly says the company is being rebuilt from scratch — looks risky for anyone who values organizational stability.
Governance questions get louder. While OpenAI, Anthropic and others face criticism over governance, they at least pretend to separate the interests of their most powerful backers from the day‑to‑day lab. By contrast, xAI is now nested inside Musk’s private conglomerate with minimal visible checks and balances. That may be efficient; it will also worry regulators, enterprise customers and potential SpaceX investors.
Execution risk spikes. Kroiss ran pre‑training, the beating heart of any frontier‑model effort. Losing that expertise during a self‑declared “rebuild from the foundations up” is a huge operational shock. Unless xAI already has a second line of leaders ready to take over, roadmap slippage is almost guaranteed.
The winners here are Musk — who gains more direct control — and rival labs, which may soon be able to hire disillusioned former xAI staff. The losers are anyone who wanted xAI to emerge as a strong, independent alternative to the current US AI duopoly.
The bigger picture
This episode fits a familiar Musk pattern: acquire or found a company, move fast, blow up the org chart, centralise power, then rebuild around a smaller inner circle. We saw versions of this at Twitter/X with mass layoffs and rapid pivots, and at Tesla with brutal restructurings around key product deadlines.
AI, however, is not cars or social networks. Frontier‑model development behaves more like a multi‑year scientific programme than a consumer‑app sprint. Institutional memory, stable teams and carefully planned iterations on data, infrastructure and safety tooling matter enormously.
Across the industry, the opposite trend is visible:
- OpenAI has locked itself into a deep, multi‑year partnership with Microsoft, exchanging equity‑like economics for long‑term compute and capital.
- Anthropic has deliberately split its cap table between several big tech partners to avoid single‑corporate control, even as it raises billions.
- Google DeepMind’s strength comes from long‑lived research groups and infrastructure that survived multiple internal restructurings at Google.
Even independent labs are increasingly tying themselves to cloud giants because training frontier models at scale now costs billions of dollars in compute, talent and data infrastructure.
Musk is trying something different: funding xAI off the balance sheet and valuation of SpaceX, a space and satellite company whose cash flows and data assets (especially Starlink) may indeed be useful for AI. Strategically, that could make sense: AI‑optimised satellite networking, autonomous operations and onboard inference are real opportunities.
But the co‑founder exodus suggests a cultural mismatch between the long‑horizon, research‑heavy nature of frontier AI and Musk’s preferred mode of blitzkrieg management. Historically, founder purges at tech companies — think Apple in the 1980s or Uber in the late 2010s — have usually been imposed by boards trying to rein in a dominant personality. At xAI, the dynamic appears inverted: the dominant personality has outlasted the rest.
That raises the question: is xAI being built as a durable institution, or as another highly centralised vehicle for one person’s strategic and political goals?
The European and regional angle
For European users and companies, the xAI shake‑up matters less because of Grok’s meme‑heavy answers on X, and more because of what it says about governance in the AI stack you may eventually rely on.
The EU AI Act, GDPR and the Digital Services Act all push in the same direction: clearer accountability, documentation of risks, and meaningful separation between platforms, data and high‑risk AI systems. A lab fully embedded in a single individual’s private conglomerate, with little visible independent oversight, is going to face intense scrutiny if it wants to sell into regulated sectors or EU governments.
European AI champions such as Mistral (France), Aleph Alpha (Germany) or smaller sovereign‑AI initiatives in the Nordics and CEE are positioning themselves explicitly around trust, compliance and openness. Their pitch to enterprises in Frankfurt, Paris or Milan is: we give you cutting‑edge models without opaque US‑style corporate control.
Against that backdrop, xAI’s founder clean‑out and absorption into SpaceX may make it a harder sell for European corporates and public‑sector buyers who already hesitate to depend on X as a strategic channel. Even for individual users in Europe, where privacy expectations are higher and Musk’s clashes with EU regulators over X are fresh, the idea of one man controlling rockets, satellites, social discourse and now a frontier‑model lab will reinforce calls for strict enforcement of EU rules.
At the same time, the departing xAI founders and senior engineers are exactly the kind of talent European labs and startups would love to attract — if they can offer compelling research freedom and governance in return.
Looking ahead
Over the next 12–18 months, expect three things.
First, a rebranding and technical reboot of xAI under the SpaceX umbrella. Musk will likely position xAI not just as a chatbot provider for X, but as the intelligence layer across Starlink, rocket operations and maybe even Tesla‑adjacent projects. Watch for announcements about new large‑scale training runs, massive GPU clusters and tight integration between satellite data and AI models.
Second, a high‑profile hiring spree. To counter the narrative of a hollowed‑out lab, Musk will need visible, credible AI leaders — ideally with pedigrees from Google DeepMind, OpenAI or top academic labs. How many such people are willing to work in a highly centralised, high‑drama environment will be a key test of xAI’s future.
Third, greater transparency via finance. If SpaceX does move towards an IPO or quasi‑public listing, investor disclosures could reveal xAI’s financial weight, R&D spend and commercial traction. That, in turn, will clarify whether xAI is a serious profit driver or mainly a strategic asset bundled into Musk’s story about a vertically integrated tech empire spanning space, connectivity and AI.
Unanswered questions abound: Will xAI continue as a relatively independent brand or disappear into SpaceX product lines? Will X remain the primary consumer interface, or will xAI target enterprises and governments directly? And crucially, will regulators in the US and EU view this three‑way integration of rockets, social media and AI as acceptable, or as a concentration of power that demands new safeguards?
The bottom line
The complete departure of xAI’s co‑founders, right as the lab is folded into SpaceX and “rebuilt from the foundations up,” is a stark signal: this is no longer a conventional startup, but an extension of Elon Musk’s personal strategy. That may yield bold, tightly integrated products — or it may scare away the very talent and partners a frontier‑model lab needs to matter. The real question for readers is simple: would you trust your business, or your democracy, to an AI stack built this way?
