Maia
Microsoft's family of in-house AI inference chips, intended to provide energy-efficient inference compute for Microsoft's own products and services and to reduce reliance on external GPUs.
Timeline (last 90 days)
The Maia project trains separate models for different rating bands (such as 1100, 1500, and 1900 Elo) to capture the typical style and weaknesses of players at each level.
OpenAI models, Microsoft's in-house MAI models, and mainstream open-source model families all run on Maia chips, covering the United States, Europe, and Australia.
Microsoft plans to scale the Maia family to multi-gigawatt deployments starting in 2027 and make it the primary inference system for Microsoft products and services.
Maia系列并不对外零售，而是作为Azure内部基础设施的一部分存在。
微软在2023年Ignite大会上正式发布了自研AI加速器品牌Maia，首代产品为Maia 100。
Maia系列走ASIC式专用加速路线，与NVIDIA GPU的通用路线不同，牺牲部分通用性换取单位功耗下的推理效率。
Amazon has Trainium and Inferentia, Google has TPU, and Microsoft's Maia has already iterated to its third generation.
Major vendors reduce per-inference costs through Mixture of Experts (MoE) models, model distillation, quantization techniques, and custom AI chips (such as Google's TPU and Microsoft's Maia).
All Facts (9)
TPU的成功开启了AI专用芯片的行业浪潮，此后亚马逊推出了Trainium/Inferentia，微软开发了Maia
90%UnverifiedThe Maia project trains separate models for different rating bands (such as 1100, 1500, and 1900 Elo) to capture the typical style and weaknesses of players at each level.
50%UnverifiedOpenAI models, Microsoft's in-house MAI models, and mainstream open-source model families all run on Maia chips, covering the United States, Europe, and Australia.
50%UnverifiedMaia系列走ASIC式专用加速路线，与NVIDIA GPU的通用路线不同，牺牲部分通用性换取单位功耗下的推理效率。
50%UnverifiedMaia系列并不对外零售，而是作为Azure内部基础设施的一部分存在。
50%UnverifiedAmazon has Trainium and Inferentia, Google has TPU, and Microsoft's Maia has already iterated to its third generation.
50%UnverifiedMajor vendors reduce per-inference costs through Mixture of Experts (MoE) models, model distillation, quantization techniques, and custom AI chips (such as Google's TPU and Microsoft's Maia).
50%UnverifiedMicrosoft plans to scale the Maia family to multi-gigawatt deployments starting in 2027 and make it the primary inference system for Microsoft products and services.
50%Unverified微软在2023年Ignite大会上正式发布了自研AI加速器品牌Maia，首代产品为Maia 100。
50%