Elon Musk used an all-hands meeting to relaunch xAI as a distinctly different enterprise: four sharply delineated product units reporting directly to him, and a remit that now stretches from coding assistants to a moon-based factory for AI satellites. The reorganisation follows the quiet departure of several co‑founders and is being presented as a necessary step as the firm scales from scrappy start‑up into a component of a much larger SpaceX-led conglomerate.
Under the new structure xAI is split into Grok (the core model and voice), Grok Code (a coding specialist unit), Grok Imagine (image and video generation), and a provocatively named Macrohard team tasked with building digital agents capable of running companies. Leadership assignments are explicit: Aman Madaan stays on for the main model and voice work, Zhang Guodong and Manuel Kroiss head the code effort with a six‑month mandate to compete for enterprise business, and ex‑DeepMind engineer Toby Pohlen leads Macrohard, which aims to simulate end‑to‑end digital business output.
Musk also teased a near-term model upgrade — Grok 4.2 as a ‘small’ advance before medium and large versions — and expanded the remit of Grokipedia, a Grok-powered encyclopedia positioned as an alternative to Wikipedia. The company claims rapid progress in image and video generation and has publicly set aggressive benchmarks for the coding arm, explicitly naming Anthropic and other competitors as targets. The tone of the meeting combined technical bragging with managerial bluntness: growth, Musk argued, inevitably means some early members must move on.
That bluntness follows a steady trickle of talent departures: co‑founders and senior researchers have left in recent months, and some employees cite a punishing hours culture and unease about the SpaceX merger. SpaceX completed an all‑stock acquisition of xAI, creating a combined entity valued at roughly $1.25 trillion and folding the AI lab into Musk's broader space and social‑media ambitions. For many early staff the bargain of a fast, autonomous AI lab has become a part of a larger, more cumbersome enterprise with different incentives.
The strategic case Musk articulated hinges on compute scale and a literal expansion into space. xAI describes itself as the first to run training on ten thousand and then a hundred thousand H100 equivalents, and is planning to scale toward a million H100‑equivalent GPUs. Infrastructure already cited includes a Memphis cluster with hundreds of thousands of Grace Blackwell chips and a planned “Macroharder” cluster adding hundreds of thousands of GB300 chips. Linked to SpaceX, Musk offered an audacious plan: build factories on the Moon to manufacture AI satellites and use a mass‑driver to fling them into space, thereby unlocking orders of magnitude more deployable compute.
Beyond models and chips, Musk mapped xAI into X (formerly Twitter), seeking to turn the social network into a super‑app with end‑to‑end encrypted messaging (XChat), payments (X Money) and embedded AI services. The stated intent is to give users daily reasons to engage — communications, AI assistants and finance in one product — and to push monthly active users well beyond the current roughly 600 million toward a target exceeding one billion daily active users.
The move is strategically coherent but fraught. Coupling AI ambitions to SpaceX’s launch and infrastructure plan creates an organising narrative — compute at planetary and extra‑planetary scale — but raises execution and talent risks. The engineering and capital hurdles to push meaningful compute infrastructure off Earth, the management challenges of integrating an aggressive AI lab into a trillion‑dollar aerospace group, and the regulatory and geopolitical sensitivities around advanced chips and data flows all threaten to slow the rapid cadence Musk prizes. Still, if any company can command the industrial backing to attempt such scale, Musk's vertically integrated vision is a plausible, if unconventional, route to a dominant position.
For international observers the episode highlights a familiar trade‑off in cutting‑edge tech: unilateral speed and grand vision can produce headline‑grabbing ambition and market momentum, but they often come at the cost of institutional stability and predictable governance. xAI 2.0 now reads less like a narrow research lab and more like a strategic pillar inside a sprawling Musk empire — one whose success will depend as much on engineering and capital as on the ability to retain creative talent and navigate emerging regulatory boundaries.