OpenAI Planned to Split Robotics and Consumer Hardware Departments
According to The Wall Street Journal, OpenAI CEO Sam Altman pushed for the separation of the robotics and consumer hardware departments at the end of last year, allowing them to finance and operate independently to avoid dragging down the core AI business.
This plan was ultimately rejected, mainly because the new entities would still need to be consolidated into the parent company's financial reports, preventing a true separation. Currently, both departments are operating independently internally and report directly to Altman.
OpenAI is facing pressure to focus: its core business is lagging behind Anthropic, some revenue targets have not been met, and the video tool Sora has been cut to free up computing power. The company is shifting towards developing "super applications" aimed at developers and enterprise users. A restart of the split may reference Google's Alphabet model in the future.
The hardware department acquired Jony Ive's company io and its 55 employees for $6.5 billion in stock last May, with new devices expected to ship no earlier than February 2027.
Source: Public Information
ABAB AI Insight
Sam Altman has long approved a large number of non-chatbot projects. This split discussion continues OpenAI's transition from "full-stack AGI exploration" to a focus on core business before an IPO. Earlier, projects like Sora were cut to concentrate computing power, reflecting growing internal concerns about the long-term cash burn of hardware and robotics.
In terms of capital strategy, Altman attempted to attract external financing through independent entities to reduce the burden on the parent company. However, accounting consolidation issues led to the shelving of the plan. The strategic motive is to clean up non-core assets for a potential IPO while maintaining control over hardware (especially Jony Ive's team) for future consumer-grade AGI devices.
Similar to Google's establishment of Alphabet in 2015 to separate its core search business from Moonshot projects, or Meta's early attempts at independent hardware departments, top AI labs are currently in a mid-to-late stage of transitioning from aggressive diversification to core business slimming before an IPO, significantly alleviating valuation pressure on parent companies with clear financial paths.
Essentially, this is about capital concentration: the split attempt aims to shift the capital demands of non-core hardware/robotics from the parent company to independent financing. The mechanism is to optimize finances before an IPO by shedding high cash-burn businesses, shifting pricing power from broad exploration to core large models and super applications, accelerating the concentration of industry capital towards a few players focused on AGI infrastructure and enterprise scenarios.
The harder the split is to achieve, the greater the IPO pressure becomes; financial consolidation is the real rule that determines survival.
The longer hardware burns cash, the more the core business needs to slim down; focus is always the only answer before going public.
The earlier the acquisition of Jony Ive's team, the later the new devices ship, making consumer hardware a long-term bet for AI companies.