Takeaways by Saasverse AI
- Meta | Acquisition | Rivos Inc. | AI Chip Development (RISC-V Architecture).
- Strategically driven to reduce reliance on Nvidia and boost Meta’s MTIA chip series.
- Rivos specializes in advanced RISC-V chips with TB-scale memory, tailored for AI workloads.
Meta Platforms, Inc. has acquired Rivos Inc., a California-based startup specializing in RISC-V architecture chip designs, marking a significant step forward in Meta's ambition to develop proprietary AI chips. The announcement was confirmed by Yee Jiun Song, Meta’s Vice President of Engineering, on LinkedIn. This move underscores Meta’s commitment to enhancing its Meta Training and Inference Accelerator (MTIA) series, a line of chips designed to power next-generation AI models and workloads.
Rivos, headquartered in Santa Clara, is recognized for its innovative approach to chip design, leveraging the open-source RISC-V architecture. The company develops chips that integrate GPUs with multiple CPUs, utilizing the RVA23 Profile technology. These chips are designed to efficiently handle complex AI models and data-intensive tasks, thanks to their vector extension capabilities and high-performance architecture. Notably, Rivos’ chips will offer “terabyte-scale” memory configurations, featuring DRAM and high-bandwidth memory modules, which are critical for AI and advanced analytics applications.
Prior to the acquisition, Rivos was reportedly seeking $300–400 million in funding, with a valuation exceeding $2 billion. By integrating Rivos into its ecosystem, Meta aims to significantly reduce its dependence on third-party hardware providers, such as Nvidia, which has traditionally dominated the AI accelerator market. This strategic acquisition also aligns with Meta’s broader efforts to scale its AI infrastructure. Earlier today, Meta announced a $14.2 billion compute-power supply agreement with CoreWeave, securing critical resources to develop and run advanced AI models for the next six years.
“ From a strategic perspective, this acquisition is a pivotal move for Meta as it doubles down on chip autonomy. By adopting RISC-V architecture, Meta gains not only cost advantages by avoiding hefty licensing fees associated with proprietary architectures like Arm but also complete control over its silicon technology stack. This freedom is instrumental for scaling AI workloads while driving down costs. Furthermore, Rivos’ cutting-edge capabilities in high-bandwidth memory and vector processing align perfectly with Meta’s needs for training large-scale AI models. ” Saasverse Analyst comments
Saasverse Insights
Meta's acquisition of Rivos highlights a growing trend among tech giants: the pursuit of proprietary chip design to secure independence in AI hardware. As AI models become increasingly complex, the demand for specialized, high-performance chips is skyrocketing. The choice of RISC-V architecture reflects a broader industry shift towards open-source innovation, enabling companies to sidestep traditional licensing constraints while fostering a more flexible and customizable development environment.
For the broader AI, SaaS, and Cloud ecosystem, this deal signals an intensifying focus on chip-level innovation as a differentiator. With RISC-V paving the way for greater accessibility and modularity in chip design, there are immense opportunities for startups and enterprises to build tailored AI solutions. However, as competition for compute power intensifies, companies must navigate risks around supply chain dependencies and resource allocation.
In the next three to five years, as AI adoption scales across industries, compute resources will emerge as a critical competitive asset. For market players, the ability to innovate in localized and vertical-specific AI infrastructure services could unlock significant growth potential, especially as enterprises seek to optimize costs and performance in a rapidly evolving technological landscape.