Parasail 将规模化部署 Positron Titan 服务器,单机可容纳 32 万亿参数模型
Parasail 要把 Positron 的 Titan 服务器推上线,单台能装 32 万亿参数模型,靠 DRAM 替代 HBM 降内存成本。做超大模型推理部署的团队可以看看这条不走 NVIDIA 的路线,5MW 起步规模不小。
推理云服务商 Parasail 宣布将率先规模化部署 Positron 的下一代 Titan 服务器,首批规模为 5MW。Parasail 是一家通过统一 API 运营多种芯片类型的推理云平台,此前已接入多类加速硬件。
Titan 的核心思路是用廉价的大容量 DRAM 取代稀缺且昂贵的高带宽内存(HBM)来承载模型权重,使单台服务器能装下高达 32 万亿参数的模型。这一架构以"每瓦内存容量"作为核心指标,适合多万亿参数模型的长上下文推理场景,代价是内存带宽低于 HBM。
对推理服务市场而言,这代表了一条与 NVIDIA GPU + HBM 路线不同的路线:用商品化 DRAM 降低大参数模型的内存成本,可能改变 MoE 类稀疏大模型的部署经济性。对需要部署超大参数模型的企业来说,新增了一个 GPU 之外的选项。
Memory capacity per watt is becoming one of the most useful ways to judge AI hardware.
Positron Titan runs multi-trillion-parameter models on commodity DRAM, and @parasail_io is the first inference cloud putting it online at scale.
Parasail, an inference cloud that already runs several chip types behind one API, just announced it will be the first to deploy Positron’s next-gen Titan servers at scale, starting with 5MW.
Titan keeps giant models in cheap smartphone-style memory instead of scarce high-bandwidth memory (HBM), so one server can hold a model of up to 32 trillion parameters.
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Positron, Parasail, AI chips, HBM.
Source attribution: curated from rohanpaul_ai (@rohanpaul_ai).