Elon Musk says xAI is starting over — here’s what that means 🚀
Elon Musk recently announced that xAI is being rebuilt “from the foundations up,” after a wave of departures and organizational changes at the deep learning lab. What began as an ambitious rival to OpenAI and Anthropic is now undergoing a personnel overhaul and a renewed focus on the products that drive revenue — especially coding tools and larger agent ambitions like Macrohard. This post breaks down the who, why, and what to watch next. 🤖
What happened: departures, reorganizations, and public scrutiny
Of the original 11 xAI cofounders, only two now remain actively involved. This week cofounders Zihang Dai and Guodong Zhang left after disagreements over product competitiveness, particularly around AI coding tools. A month earlier, 11 senior engineers — including two other cofounders — departed during a broader reorganization Musk described as necessary to scale the business.
Reports say SpaceX and Tesla executives have been brought in to review staff and make cuts where needed. Musk framed the shakeup as intentional: xAI “was not built right the first time around,” and now the team is reworking the foundations to compete more effectively with established labs.
Why coding tools matter (and why xAI is under pressure)
Coding assistants are a major revenue channel for AI companies. Products like Anthropic’s Claude Code and OpenAI’s Codex demonstrate that developer productivity tools can convert to subscriptions and enterprise deals. xAI’s Grok LLM had an early user surge partly because of lax content moderation allowing provocative outputs, but that growth doesn’t translate directly into sustainable revenue.
Musk criticized xAI’s coding tool output as failing to keep pace with rivals and called an all-hands to map a recovery plan, aiming to be more competitive by mid-year. The practical implication: building high-quality coding assistants is now a business imperative, not just a product milestone.
Hiring moves and internal fixes
Musk has been personally involved in reviewing past job rejections and tapping talent from other firms. Notable hires include Andrew Milich and Jason Ginsberg from Cursor, a company focused on AI coding tools. Their move signals two things:
- xAI is prioritizing people with hands-on product engineering experience in coding tools.
- Access to frontier models and compute (rather than relying on third-party models) remains a competitive advantage for labs like xAI.
Musk also said he and Baris Akis are revisiting previously rejected applicants to find overlooked talent — an unusual, hands-on recruiting approach for a CEO of multiple companies. LinkedIn headcounts place xAI at roughly 5,000 employees, versus 7,500+ at OpenAI and 4,700+ at Anthropic, underscoring the scale challenge.
Macrohard and the broader agent vision
Beyond coding tools, Musk is pushing a grander idea: Macrohard, an xAI project aimed at building an agent that can perform any white-collar task on a computer. The concept pairs xAI’s language model with a Tesla-developed agent dubbed “Digital Optimus,” suggesting xAI would instruct a Tesla-operated agent to execute tasks.
Macrohard has already seen leadership churn — Toby Pohlen, tapped to lead the project in February, left soon after — and Business Insider reported the initiative is currently on pause. Still, Musk says Macrohard is a joint effort with Tesla, tying xAI more tightly to his other companies and hinting at cross-company synergies (and pressure).
How xAI compares to competitors
- OpenAI and Anthropic: Both have proven commercial products in coding and enterprise offerings; xAI is racing to catch up.
- Cursor: A specialized coding tool provider whose engineers recently joined xAI — a sign that product talent is moving toward labs with in-house models and compute.
- Perplexity and others: Companies exploring “digital proxy” agent concepts that overlap with Macrohard’s vision, showing the idea isn’t unique but is competitive.
What’s at stake: business, reputation, and a looming IPO narrative
xAI is part of Musk’s wider corporate web, now linked to SpaceX. With a potential public offering of SpaceX shares on the horizon, Musk faces pressure to show that xAI is more than a cash-burning experiment. A struggling AI arm could muddy investor sentiment, so delivering tangible product traction on Grok and related tools matters more than ever.
What to watch next 🔭
- Progress on xAI’s coding tools: measurable product improvements and new enterprise deals.
- Further hires from competitive AI tool companies — to see if xAI can attract and retain product-focused engineers.
- Macrohard’s status and any concrete collaboration announcements with Tesla or SpaceX.
- User adoption and monetization of Grok: are early engagement metrics converting to sustainable revenue?
Bottom line: Musk has cast xAI’s current turmoil as deliberate rebuilding. The lab’s fate will hinge on its ability to rapidly improve revenue-generating products (especially coding tools), stabilize leadership, and make Macrohard more than an aspirational project. If xAI can combine stronger talent with direct access to models and compute, it still has a shot — but the clock is ticking. ⏳