Meta's ambitious bet to build its own AI chips through the acquisition of Rivos has hit a significant roadblock.
The core issue, reported by The Information, is that the 2025 acquisition of chip startup Rivos isn't delivering on its promise. Internal friction and a shifting strategy have stalled the integration, preventing Meta from accelerating its in-house silicon development. This is a major concern because custom chips represent Meta's most promising path to taming the astronomical costs associated with running advanced AI models.
So, how did a strategic acquisition turn into a source of friction? The situation arose from a combination of immense pressure and organizational complexity.
First, Meta is investing a colossal sum in its AI future, with a capex plan of up to $145 billion for 2026 alone. This level of spending creates enormous pressure to reduce the per-unit cost of AI computation. Developing proprietary chips is the most effective way to achieve these long-term savings and justify such a massive investment to shareholders.
Second, Meta's internal chip design efforts have already faced significant hurdles. The company previously scrapped an advanced training chip codenamed "Olympus," signaling design and execution challenges. This setback likely made acquiring a promising RISC-V startup like Rivos seem like an ideal shortcut to bolster its capabilities and get back on track.
However, Rivos was integrated into an already crowded and complicated ecosystem. Meta maintains deep, multi-year partnerships with external suppliers like Broadcom and AMD, while still relying heavily on Nvidia. Amidst this complex web of dependencies and shifting internal priorities, such as pausing a collaboration with Samsung, the Rivos team appears to have struggled to find a clear purpose for its technology. This lack of a defined role seems to be a key driver of the reported internal conflicts.
Ultimately, instead of fast-tracking Meta's ambitions, the Rivos acquisition has so far added another layer of complexity. It weakens Meta's negotiating power with its powerful suppliers and raises doubts about its ability to control its spiraling AI costs—a critical narrative for investors.
- Capex (Capital Expenditure): Funds used by a company to acquire, upgrade, and maintain physical assets such as data centers, servers, and other equipment.
- RISC-V: An open-source instruction set architecture (ISA) used to design processors. It serves as a free and customizable alternative to proprietary architectures from companies like ARM.
- ASIC (Application-Specific Integrated Circuit): A microchip designed for a special application, such as an AI accelerator, rather than for general-purpose use.