Meta Platform is preparing to take a significant step toward reducing its dependence on external chip suppliers by beginning production of its first large-scale in-house artificial intelligence chip, codenamed Iris, from September. The move forms part of the company's broader strategy to strengthen its AI infrastructure as competition intensifies among global technology giants building next-generation AI platforms.
According to industry reports, Meta plans to rapidly expand its computing infrastructure to 14 gigawatts by 2027, nearly doubling its expected capacity over the next year to support increasingly demanding AI workloads powering products across Facebook, Instagram, WhatsApp, and its expanding AI ecosystem.
Meta Accelerates In-House AI Silicon Strategy
The Iris processor is part of Meta's long-term Meta Training and Inference Accelerator (MTIA) roadmap, a four-generation family of AI chips being developed internally to improve both AI model training and inference capabilities.
Unlike relying entirely on third-party graphics processors, Meta is designing its own silicon specifically optimized for its AI models and large-scale data center operations. The company believes custom-built processors can deliver higher efficiency, lower operating costs, and better performance for its unique AI workloads.
The project also represents a major milestone for Meta's semiconductor ambitions after earlier attempts to develop proprietary AI chips progressed slowly over the past several years.
Successful Testing Clears Path for Production
The internal memo revealed that testing of the Iris chip was completed in just six weeks without any major technical issues, clearing the way for production to begin later this year.
Industry observers view the successful validation as an encouraging sign for Meta's chip development program, which aims to become less dependent on external GPU vendors while gaining greater control over AI infrastructure.
The company has partnered with Broadcom for chip design support, while manufacturing will be handled by Taiwan Semiconductor Manufacturing Company (TSMC), the world's leading semiconductor foundry.
Reducing Dependence on NVIDIA and AMD
Although Meta will continue deploying significant volumes of AI GPUs from NVIDIA and AMD, its custom processors are expected to complement those systems rather than replace them immediately.
Internal documents reportedly indicate that integrating the latest commercial GPUs across Meta's massive infrastructure has become increasingly complex and expensive. By introducing internally designed accelerators, the company hopes to optimize deployment timelines while lowering long-term infrastructure costs.
Analysts say owning critical semiconductor technology has become increasingly important as AI companies seek greater independence from external hardware suppliers.
Massive Infrastructure Expansion Underway
Meta is simultaneously investing heavily in expanding its global AI computing capacity.
The company expects to operate approximately 7 gigawatts of computing infrastructure during 2026, including one gigawatt already deployed during the first half of the year and an additional 5.5 gigawatts planned before year-end.
By next year, Meta intends to double total computing capacity to 14 gigawatts, creating one of the world's largest AI infrastructure networks dedicated to training and deploying large language models and generative AI services.
For perspective, one gigawatt of electricity is generally sufficient to power nearly 800,000 homes, highlighting the enormous scale of Meta's AI investments.
Billions Invested in AI Infrastructure
Meta expects to spend up to $145 billion on AI infrastructure this year, making it one of the largest technology infrastructure investments globally.
To support its aggressive expansion plans, the company has secured long-term supply agreements with several technology partners, including Samsung Electronics for memory chips, SanDisk for flash storage solutions, and Sumitomo Electric for fiber-optic networking equipment.
These multi-year agreements are intended to ensure stable component availability amid continued global demand for AI hardware and data center infrastructure.
AI Race Driving Custom Chip Development
Meta joins a growing list of technology companies investing in proprietary AI processors as artificial intelligence becomes central to future digital services.
Custom-designed AI chips allow companies to optimize hardware specifically for their own AI models while reducing dependence on commercial GPU suppliers, improving efficiency and lowering operational costs over time.
As AI adoption accelerates worldwide, ownership of both software and semiconductor technology is increasingly viewed as a strategic competitive advantage, positioning companies like Meta to better control the performance, scalability, and economics of next-generation AI platforms.
𝐒𝐭𝐚𝐲 𝐢𝐧𝐟𝐨𝐫𝐦𝐞𝐝 𝐰𝐢𝐭𝐡 𝐨𝐮𝐫 𝐥𝐚𝐭𝐞𝐬𝐭 𝐮𝐩𝐝𝐚𝐭𝐞𝐬 𝐛𝐲 𝐣𝐨𝐢𝐧𝐢𝐧𝐠 𝐭𝐡𝐞 WhatsApp Channel now! 👈📲
𝑭𝒐𝒍𝒍𝒐𝒘 𝑶𝒖𝒓 𝑺𝒐𝒄𝒊𝒂𝒍 𝑴𝒆𝒅𝒊𝒂 𝑷𝒂𝒈𝒆𝐬 👉 Facebook, LinkedIn, Twitter, Instagram
