United States: Elon Musk has told employees of his artificial intelligence startup xAI that the global AI race will be won not by the biggest company but by the fastest one, outlining a major restructuring plan, new product verticals and a long-term vision where AI agents could run companies. Speaking at a recent company-wide meeting later shared publicly on X, Musk emphasised speed, execution and organisational focus as critical factors in staying ahead of competitors in advanced AI development.
The internal meeting came at a time of transition for xAI, with multiple early leaders stepping away from the company. Musk acknowledged the departures but said the company would now focus on acceleration, tighter structure and ambitious technical goals.
Leadership exits and organisational reset
During the meeting, Musk addressed the recent exit of two xAI co-founders, Jimmy Ba and Tony Wu. Their departure adds to a broader pattern of turnover among early founding members since xAI was launched in 2023. Several original contributors, including Kyle Kosic, Igor Babuschkin and Christian Szegedy, have also left over the past two years, while Greg Yang has reportedly reduced his role following a health diagnosis.
Musk thanked those who contributed in the early stages and said such transitions are not unusual in fast-scaling technology ventures. He indicated that the next phase for xAI would centre on sharper execution and clearly defined verticals rather than a broad, loosely structured research approach.
He told employees that momentum and pace of delivery now matter more than headcount or size. According to Musk, acceleration in model capability and product deployment will determine leadership in AI.
Four verticals to drive xAI growth
As part of the restructuring, Musk announced that xAI will now operate across four primary verticals.
The first vertical is Grok, the company’s chatbot and voice platform, which is already integrated with X and positioned as a consumer-facing conversational AI system. The second vertical focuses on coding tools, aimed at software development assistance and automated programming support.
The third vertical is Imagine, xAI’s video-generation and visual AI initiative, which is expected to expand into more advanced media synthesis and understanding. Musk indicated that video — especially real-time video interpretation and generation — will be central to the next wave of AI systems.
The fourth and most forward-looking vertical is a new division called Macrohard. This group is tasked with building AI agents capable of operating with higher autonomy, including managing workflows and potentially running business processes.
Macrohard is being led by founding member Toby Pohlen and represents Musk’s push toward AI systems that move beyond chat responses into operational decision-making roles.
AI agents that can run companies
Musk said future AI systems should not just answer questions or generate text and images but act as autonomous agents capable of executing complex tasks. He described a future where AI agents could oversee business functions, coordinate operations and make decisions within defined frameworks.
This direction reflects a broader industry shift toward “agentic AI” — systems designed to plan, act and adapt across multi-step objectives instead of handling single prompts.
According to Musk, building such agents requires integrated advances in reasoning, multimodal understanding and real-time data processing. He suggested that companies that achieve reliable autonomous agents first could gain a decisive competitive advantage.
Emphasis on speed and technical velocity
A recurring theme in Musk’s remarks was velocity. He told staff that progress speed — how quickly teams build, test and deploy — will determine which AI lab leads the field.
He praised internal teams for rapid development cycles, highlighting the Grok voice capability as an example. Division head Aman Madaan said the voice system was built from scratch within roughly six months in response to rising demand and competitive pressure.
Other leadership assignments include Manuel Kroiss overseeing coding initiatives and Guodong Zhang managing the Imagine video product along with supporting programming efforts.
Musk predicted that a large share of future AI computing demand will be tied to video understanding and generation, calling it a key battleground area.
Space, computing power and long-term infrastructure
Musk also connected xAI’s future to large-scale computing infrastructure. He referenced the recent merger of xAI with SpaceX, which reportedly places the combined entity at a valuation of about $1.25 trillion. The integration is expected to strengthen access to capital, hardware and infrastructure required for frontier AI training.
He reiterated that advanced AI models demand enormous computing power and energy resources. In a futuristic extension of that idea, Musk spoke about the possibility of building a manufacturing facility on the Moon to produce AI satellites that could expand compute capacity in orbit.
He suggested that space-based infrastructure, combined with novel launch systems, could eventually support much larger AI workloads than Earth-bound facilities alone. While no timeline or engineering roadmap was presented, the concept underlines Musk’s view that AI progress will be tightly linked to breakthroughs in infrastructure and energy.
Conclusion
Musk’s message to xAI employees signals a shift toward tighter structure, faster execution and more autonomous AI systems. With four focused verticals, new leadership roles and an emphasis on agent-based AI, xAI is positioning itself for rapid iteration rather than incremental scale. As competition intensifies across the AI sector, Musk’s strategy centres on one principle: whoever moves fastest will lead.