Home
Meta to Deploy Four Generations of Self-Developed AI Chips by 2027 in Computing Power Race

Meta is accelerating its push to reduce reliance on external hardware suppliers as it strives to lead the costly global AI race. Reports indicate that the social media giant plans to have four generations of its own AI chips deployed by the end of 2027. This aggressive roadmap is designed to deliver custom hardware that powers its rapidly expanding AI business, while also cutting long-term dependence on outside vendors like NVIDIA.
Meta's in-house chip lineup now follows a clear iterative roadmap. The MTIA 300, built for content ranking and recommendation model training, is already in mass production. The MTIA 400, codenamed "Iris," has cleared lab testing and is entering deployment. The more advanced MTIA 450 ("Alce") and MTIA 500 ("Astrid") are scheduled for launch in the first and second halves of 2027, respectively. This rapid R&D cadence underscores Meta's commitment to keeping hardware evolution in step with AI algorithm advances.
While Meta is pouring billions into building its own chip team and acquiring startups like Rivos to expand its talent pool, its strategy is not entirely closed off. Senior leadership has made clear that the company follows a dual-track approach: it continues to be one of the world's largest GPU buyers, signing major deals with NVIDIA and AMD to secure baseline computing power, while also deploying custom chips that strip away unnecessary features for general use, thereby achieving higher efficiency in specialized tasks such as Instagram feed ranking and generative AI inference.
This tight integration of software and hardware is becoming a key competitive advantage for top tech companies. Even though chip development typically takes two years and involves significant engineering challenges, Meta is confident that custom chips can reduce long-term operating costs by eliminating non-essential features. Given the unexpectedly high demand for computing power, Meta is reviewing and refining the technical roadmaps for each generation of its MTIA chips, aiming for a dynamic balance between in-house development and external procurement to maintain its edge in generative AI.
Related article
Five Departments Launch AI Plus Education Plan to Develop AI MOOCs Across All Academic Stages
On April 10, the Ministry of Education, the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Science and Technology, and the National Data Administration jointly released the "Action Pla
ByteDance’s Seed launches global campus drive, offering virtual shares to win top large model talent
In the competitive landscape of large language models, securing top-tier talent remains the most critical strategic asset.On April 1st, ByteDance announced the launch of its Seed global campus recruitment initiative, part of its large model talent de
Suno to Watermark Songs Amid Legal Battles
Suno, the platform enabling users to generate AI-created music, has unveiled new features to label platform-produced tracks, restrict downloads, and update community standards to curb unauthorized replicas. These updates arrive as Suno confronts mult
Related Special Topic Recommendations
Comments (0)
0/500

Meta is accelerating its push to reduce reliance on external hardware suppliers as it strives to lead the costly global AI race. Reports indicate that the social media giant plans to have four generations of its own AI chips deployed by the end of 2027. This aggressive roadmap is designed to deliver custom hardware that powers its rapidly expanding AI business, while also cutting long-term dependence on outside vendors like NVIDIA.
Meta's in-house chip lineup now follows a clear iterative roadmap. The MTIA 300, built for content ranking and recommendation model training, is already in mass production. The MTIA 400, codenamed "Iris," has cleared lab testing and is entering deployment. The more advanced MTIA 450 ("Alce") and MTIA 500 ("Astrid") are scheduled for launch in the first and second halves of 2027, respectively. This rapid R&D cadence underscores Meta's commitment to keeping hardware evolution in step with AI algorithm advances.
While Meta is pouring billions into building its own chip team and acquiring startups like Rivos to expand its talent pool, its strategy is not entirely closed off. Senior leadership has made clear that the company follows a dual-track approach: it continues to be one of the world's largest GPU buyers, signing major deals with NVIDIA and AMD to secure baseline computing power, while also deploying custom chips that strip away unnecessary features for general use, thereby achieving higher efficiency in specialized tasks such as Instagram feed ranking and generative AI inference.
This tight integration of software and hardware is becoming a key competitive advantage for top tech companies. Even though chip development typically takes two years and involves significant engineering challenges, Meta is confident that custom chips can reduce long-term operating costs by eliminating non-essential features. Given the unexpectedly high demand for computing power, Meta is reviewing and refining the technical roadmaps for each generation of its MTIA chips, aiming for a dynamic balance between in-house development and external procurement to maintain its edge in generative AI.
Five Departments Launch AI Plus Education Plan to Develop AI MOOCs Across All Academic Stages
On April 10, the Ministry of Education, the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Science and Technology, and the National Data Administration jointly released the "Action Pla
ByteDance’s Seed launches global campus drive, offering virtual shares to win top large model talent
In the competitive landscape of large language models, securing top-tier talent remains the most critical strategic asset.On April 1st, ByteDance announced the launch of its Seed global campus recruitment initiative, part of its large model talent de
Suno to Watermark Songs Amid Legal Battles
Suno, the platform enabling users to generate AI-created music, has unveiled new features to label platform-produced tracks, restrict downloads, and update community standards to curb unauthorized replicas. These updates arrive as Suno confronts mult











