Meta Platforms is set to begin manufacturing its first-generation in-house artificial intelligence chip, codenamed “Iris,” this September. According to an internal memo, the move is a core part of the tech giant’s strategy to dramatically scale its computing infrastructure and reduce its heavy reliance on external suppliers like Nvidia and AMD.
Quick Facts
- ‘Iris’ chip production begins in September.
- Targeting 7 gigawatts of computing power in 2026.
- Plan to double capacity to 14 gigawatts by 2027.
Moving Beyond Nvidia
For years, Meta has depended on high-performance GPUs from Nvidia and AMD to power the AI behind its Facebook and Instagram platforms. This new strategy, however, signals a decisive shift toward greater independence. By designing its own silicon tailored for its specific workloads, Meta aims to lower its massive computing costs and gain more control over its hardware roadmap.
The “Iris” chip is the first of a four-generation project for Meta Training and Inference Accelerators (MTIA). While Meta designed the chip in-house, it is collaborating with Broadcom and will use Taiwan Semiconductor Manufacturing Co (TSMC) for production. This follows a broader industry trend where “hyperscalers” like Google, Amazon, and now Meta are developing custom chips to optimize performance and cost for their large-scale AI operations.
An Aggressive Hardware Roadmap
According to the memo, the “Iris” chip completed testing in just six weeks with no major issues found, indicating strong momentum for a project that has been in development for over five years.
Meta is not stopping with Iris. The company plans to release a new AI processor approximately every six months through 2027—a much faster cadence than the typical yearly cycle in the chip industry. This aggressive timeline underscores the urgency and scale of Meta’s AI ambitions as it competes directly with OpenAI and Anthropic.
Implications for the MENA AI Ecosystem
While Meta’s chip production is a global move, it has significant ripple effects for the MENA region’s burgeoning tech scene. As Gulf nations invest heavily in sovereign AI capabilities and data center infrastructure, the strategies of Big Tech players set a competitive benchmark.
For MENA-based startups and enterprises relying on cloud services, increased competition in the hardware layer could eventually lead to more competitive pricing from cloud providers. Furthermore, Meta’s push for a self-sufficient supply chain highlights the strategic importance of hardware in the global AI race, a lesson for regional leaders aiming to build resilient and powerful tech ecosystems.
Securing the AI Supply Chain
To support this massive expansion, which includes a planned spend of up to $145 billion on AI infrastructure this year, Meta has locked in multi-year supply agreements. These deals include memory chips from Samsung Electronics, flash storage from Sandisk, and fiber-optic equipment from Sumitomo Electric.
Securing these components is critical amid what some analysts call “chipflation,” where a surge in demand for AI-related hardware has caused prices to rise sharply. These long-term agreements are essential for Meta to hit its target of 7 gigawatts of computing infrastructure this year and its goal of doubling that to 14 gigawatts in 2027.
About Meta
Meta builds technologies that help people connect, find communities, and grow businesses. When Facebook launched in 2004, it changed the way people connect. Apps like Messenger, Instagram, and WhatsApp further empowered billions around the world. Now, Meta is moving beyond 2D screens toward immersive experiences like augmented and virtual reality to help build the next evolution in social technology.
Source: Zawya
