No Nvidia, No AMD, No Intel, No ARM: Meta plans inference-led RISC-y future without friends as 1700w superchip emerges with 30 PFLOPs performance and
half Terabyte (yes 512GB) HBM
Date:
Thu, 19 Mar 2026 21:05:00 +0000
Description:
Meta develops MTIA custom chips and a 1700W superchip to run GenAI inference efficiently without relying on mainstream silicon vendors.
FULL STORY ======================================================================Copy link Facebook X Whatsapp Reddit Pinterest Flipboard Threads Email Share this article 0 Join the conversation Follow us Add us as a preferred source on Google Newsletter Tech Radar Get the TechRadar Newsletter Sign up for
breaking news, reviews, opinion, top tech deals, and more. Contact me with news and offers from other Future brands Receive email from us on behalf of our trusted partners or sponsors By submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over. You are
now subscribed Your newsletter sign-up was successful An account already exists for this email address, please log in. Subscribe to our newsletter Metas 1700W superchip delivers 30 PFLOPs and 512GB of HBM memory MTIA 450 and 500 prioritize inference over pre-training workloads Future MTIA generations will support GenAI inference and ranking workloads Meta is advancing its AI infrastructure with a portfolio of custom MTIA chips designed specifically
for inference workloads across its apps.
The company is developing a 1700W superchip capable of 30 PFLOPs and 512GB of HBM, integrated within the same MTIA infrastructure to handle inference tasks at scale. Interestingly, it is achieving this feat without any of its friends
no Nvidia , AMD , Intel , or ARM . Article continues below You may like 'No one deploys AI at Metas scale': Meta signs up Nvidia to power its next big AI projects so what exactly do Mark Zuckerberg and Jensen Huang have planned? AMD and Meta reveal massive GPU deployment news 6GW of Instinct hardware set to massively boost Facebook owner's computing stack Meta sets up 'top-level' Compute initiative to make sure its AI data centers get all the power they need - Zuckerberg promises 'tens of gigawatts this decade, and hundreds of gigawatts or more over time' Meta scales inference with extensive MTIA deployment According to Meta , hundreds of thousands of MTIA chips are
already deployed in production, supporting ranking, recommendations, and ad-serving workloads.
These chips are part of a full-stack system optimized for Metas specific requirements, achieving higher compute efficiency than general-purpose hardware for its intended workloads.
Unlike other hyperscalers such as Google , AWS, Microsoft , and Apple , Meta is pursuing a fully custom silicon strategy.
This design prioritizes efficiency over general-purpose use, allowing inference to run more cost-effectively than on mainstream GPUs or CPUs. Are you a pro? Subscribe to our newsletter Sign up to the TechRadar Pro
newsletter to get all the top news, opinion, features and guidance your business needs to succeed! Contact me with news and offers from other Future brands Receive email from us on behalf of our trusted partners or sponsors By submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over.
It maintains compatibility with industry-standard software such as PyTorch, vLLM, and Triton.
Metas MTIA roadmap anticipates four new generations of chips over the next
two years, including MTIA 300, currently in production for ranking and recommendations.
Future generations MTIA 400, 450, and 500 will expand support for GenAI inference workloads, with designs capable of fitting into existing rack infrastructure. What to read next The AI data centers of 2036 wont be filled with GPUs: FuriosaAIs CEO on the future of silicon Microsoft unveils Maia
200, its 'powerhouse' accelerator looking to unlock the power of large-scale AI 'Just remember, you have to be ahead for all time to come': Microsoft CEO Satya Nadella reveals AI chip shopping spree - says it will still buy Nvidia, AMD chips, despite recent launch of in-house Maia 200 hardware
Meta emphasizes rapid, iterative development, releasing new chips roughly every six months through modular and reusable designs.
The modular design allows new chips to drop into existing rack systems, reducing deployment friction and accelerating time to production.
The approach allows the company to adopt emerging AI techniques and hardware improvements faster than competitors, who typically cycle one to two years
per generation.
Unlike most mainstream AI chips that prioritize large-scale GenAI
pre-training and later adapt for inference, Metas MTIA 450 and 500 focus
first on inference workloads.
The chips can also support other tasks, including ranking and recommendations training or GenAI training, but their design keeps them tuned to anticipated growth in inference demand.
Metas system-level design aligns with Open Compute Project standards,
enabling frictionless deployment in data centers while maintaining high compute efficiency.
The company acknowledges that no single chip can handle the full spectrum of its AI workloads.
This is why it is deploying multiple MTIA generations alongside complementary silicon from other vendors.
The strategy aims to balance flexibility and performance while accelerating innovation toward personal superintelligence. Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
======================================================================
Link to news story:
https://www.techradar.com/pro/no-nvidia-no-amd-no-intel-no-arm-meta-plans-infe rence-led-risc-y-future-without-friends-as-1700w-superchip-emerges-with-30-pfl ops-performance-and-half-terabyte-yes-512gb-hbm
--- Mystic BBS v1.12 A49 (Linux/64)
* Origin: tqwNet Technology News (1337:1/100)