xAI announces new company structure: Top executives and their roles

xAI implements a new company structure to support rapid growth and focused product development. The AI firm has organized its structure into four core product teams and multiple supporting infrastructure layers, as presented in the recent all-hands meeting.

Core Product Teams

• Grok Main & Voice — Leads the core Grok model and voice features. Aman serves as lead, with Tyler and Jake H. supporting.
• Coding — Develops programming and code-related capabilities. Mako leads the team, supported by Guodong.
• Imagine — Focuses on image and visual generation. Guodong leads, with Haotian and Chaittu contributing.
• Macrohard — Builds advanced agents and simulations, including computer use and company-scale digital operations. Toby leads, assisted by John M.

Supporting Infrastructure Layers

• API & Core Product Infrastructure — Handles foundational APIs and product backbone. Toby and Jaime oversee this area.
• Expert Tutors & Grokikipedia — Manages domain expert evaluations and the internal knowledge base (Grokikipedia). Diego and Jake T. lead this effort.
• ML & Data Infrastructure — Supports machine learning pipelines and data systems. Mako, Alex B., Arshdeep, Lianmin, Saeed, and Yunlong contribute here.
• Compute & Networking Infrastructure — Manages high-performance compute clusters and networks. Heiner and Spencer handle this layer.
• Physical Infrastructure — Oversees data centers and physical facilities. Dan and Brent lead this segment.

xAI company structure (Source – xAI)

The company presents this clear, segmented setup to enable faster execution and specialization across product and infrastructure needs.

The reorganization follows recent departures among early members, including several cofounders. Elon Musk addresses this in the meeting, stating that company growth requires structural evolution, which leads to parting ways with some individuals. Reports indicate half the founding team has left, with some departures voluntary and others prompted. The changes aim to streamline operations for faster progress toward ambitious goals.

At the employee-only meeting, Musk revealed several remarkable advances that it achieved in only 30 months, including leading voice, image, and video generation volumes that surpass competitors combined.

Compute scales aggressively, with clusters approaching 1 million H100-equivalent GPUs. Hiring continues actively, and ambitious long-term visions integrate with SpaceX efforts, such as orbital data centers and lunar mass drivers to support massive future compute.