9 posts tagged with "Domestic Substitution"

国产 AI 芯片市场对 MLU690 的期待，从 2025 年一路延续到 2026 年。但据公开报道，MLU690 在 2026 年并没有形成大规模出货计划——寒武纪全年晶圆资源优先保障已成熟量产的 MLU580 与 MLU590。摩根士丹利 2026 年 10 月的研报也确认了这一点：MLU690 系列（特别是 MLU692）收入确认延迟，构成三季度业绩不及预期的主因。本文梳理 MLU690 的真实交付进展。

MLU690 交付现状：产能优先给成熟型号​

据公开报道，寒武纪 2026 年的产能分配策略已经明确：全年晶圆资源优先供给成熟量产的 MLU580、MLU590 两款芯片，MLU690 中芯版本暂不规划批量出货，也没有主动向客户推广。核心制约仍是产能瓶颈，即便拿到客户订单也无法按时交付，因此公司索性收缩 MLU690 的市场投放节奏。

摩根士丹利的研报从财务角度印证了同一件事：预计寒武纪 2026 年第三季度收入为 36.5 亿元人民币，低于此前预期的 47.7 亿元。不及预期的主要原因在于 MLU690 系列（特别是 MLU692）的收入确认延迟以及 MLU580 的交付量有限。寒武纪在第三季度已将 MLU692 交付给主要的 CSP（云服务提供商）进行测试，但预计该产品的大部分收入将在第四季度才能确认。

这就是 MLU690 当前的真实状态——不是"卖不动"，而是"没有足够的晶圆去卖"。

型号	2026 年产能地位	交付状态
MLU580	主力，产能倾斜	交付量有限，三季度提价后毛利率仍偏低
MLU590	主力，产能倾斜	上半年内可交付
MLU690 / MLU692	产能优先级靠后	已交付主要 CSP 测试，收入多在四季度确认
MLU790（下一代）	尚未推进	受 690 验证进度影响，节奏更谨慎

据公开报道，MLU690 的合同虽在上半年签订，但交付预计在第三季度。这与摩根士丹利所述"大部分收入在第四季度确认"基本一致。

央国企市场：可见的订单空间​

在消费级与互联网市场之外，央国企与训练场项目是 MLU690 的主要落地方向。据公开报道，2026 年央国企市场的可见商机空间预计在 10 万至 20 万卡，其中上半年预计可落地订单约为 5 万卡。这 5 万卡中，有 3 万卡将用于与京能合作的国家训练场项目，另外 2 万卡来自其他央国企计算项目扩容（如浙江台州、昆山的项目）以及上海、安徽合肥等地的训练场和电信运营商项目。

值得注意的是，央国企订单的竞争格局并不宽松。据公开报道，在央国企市场寒武纪并非绝对的第二供应商，市场地位大约能排进前五；在信创领域华为是绝对的第一供应商，占据超过 60% 的市场份额。对寒武纪构成最大竞争压力的是平头哥和昆仑芯——这两家公司背靠阿里云和百度云，具备算力消纳能力，能通过云服务输出算力，而寒武纪无法做出此类承诺。

超节点布局：停滞状态​

超节点是 2026 年国产算力的主战场，但寒武纪在这条线上基本缺席。据公开报道，寒武纪自己没有做超节点，也没有明确的自建超节点计划。当前主要是给一些做超节点的厂商提供产品做验证和适配，2026 年还没有客户明确要求寒武纪去部署超节点，而且产能也不支持这么做。

现有产能仅能勉强满足字节的八卡基础服务器订单，字节自身也没有超节点部署计划，相关计划更偏 2027 年。腾讯和阿里虽有超节点方向的规划，但寒武纪没有办法保证供货，所以没有真正推进到客户部署层面。目前更多是与 OEM 厂商做平台适配，比如阿里的盘古、浪潮、新华三等平台已经做过适配，但没有量产安排。

还有一个被忽视的连锁反应：MLU690 的延迟量产会进一步拖累下一代 MLU790 的流片进度。MLU690 本身承担 790 前置大规模商用验证任务，缺少足量客户真实场景数据迭代，下一代芯片的研发与流片节奏会持续延后。

研判​

寒武纪当前的处境可以用一句话概括：需求不缺，晶圆不够。据公开报道，摩根士丹利预计寒武纪 2027 年向 CSP 客户的出货量将接近 35 万颗，主要得益于 MLU692 的产能爬坡和下一代 MLU580 产品的推出；代工方面主要依赖中芯国际的 N+2/N+1 产能。HBM3e 仍是关键供应瓶颈，但供应链调研显示寒武纪拥有一定库存，并有望随国内产能爬坡获得长鑫存储的一部分 HBM3e 产量。

财务侧的压力同样真实：HBM 成本上升对单位经济效益造成压力，MLU580 即使三季度提价，毛利率仍将保持在相对较低水平。摩根士丹利维持增持评级，目标价从 1408 元上调至 1700 元，但明确将三季度称为"业绩低谷"。

对客户而言，MLU690 在 2026 年不是可规模化规划的选项；2027 年随着产能爬坡与 790 产品线就位，它才会真正进入放量期。这一年多的窗口期，寒武纪要证明的是晶圆产能能否转化为可持续的交付能力。

国产化替代最常见的失败模式不是"卡买不到"，而是"代码跑不动、精度对不齐、性能差五倍"。迁移本质是一个软件工程问题，用清单化的方式管理，成功率会高得多。

一、三条生态路线先分清​

昇腾（CANN 加 torch_npu）：生态最完整的国产路线。PyTorch 官方支持度较好，主流模型（LLaMA、Qwen、DeepSeek 系列）的适配案例最多，社区踩坑记录丰富。硬件侧 Atlas 800T A2（8 卡 910B）训练机约 75 万至 200 万元区间，910C 8 卡服务器渠道价约 110 万至 120 万元，推理侧 Atlas 300I Duo 单卡约 2 万至 4 万元。

寒武纪（Neuware）：训练卡 MLU 系列、推理卡加边缘盒子的组合，推理场景渗透率提升明显，国产推理卡渠道价已进入 3 万至 5 万元区间。

海光（DTK，ROCm 系）：DTK 从 ROCm 分支演化，对 CUDA 代码的兼容性是三条路线中相对友好的，GPGPU 架构使通用算子迁移成本较低，适合已有大量自定义 CUDA 算子的团队。

二、六阶段迁移清单​

阶段	工作内容	典型工作量	常见坑
1. 资产盘点	扫描代码中的自定义 CUDA 算子、专有库依赖（NCCL、TensorRT、FlashAttention）	1 至 2 周	低估第三方闭源库依赖
2. 环境搭建	驱动、固件、CANN 或对应 SDK、框架版本对齐	1 至 2 周	版本矩阵组合错误，驱动与 SDK 不匹配
3. 框架适配	安装 torch_npu 或对应插件，跑通开源模型 baseline	2 至 4 周	依赖库里隐藏的 CUDA 专属调用
4. 自定义算子迁移	用等价 API 重写或请求原厂支持	2 至 8 周，取决于算子数量	算子性能回退数倍，需原厂深度调优
5. 精度对齐	FP16 或 BF16 下逐层对比 loss 与输出	1 至 3 周	舍入差异累积导致 loss 曲线发散
6. 性能调优与压测	通信库替换、算子融合、混合精度策略	2 至 6 周	集群线性加速比不达标，多因网络配置

整体而言，一个"标准开源模型加常规训练脚本"的项目迁移到昇腾，顺利情况下约 6 至 10 周；有大量自定义算子和闭源依赖的项目，周期可能拉长到半年，且强依赖原厂驻场支持——签约时务必把驻场支持写进合同。

三、最容易翻车的五个环节​

1. NCCL 依赖。 分布式训练脚本里的 NCCL 调用要替换为昇腾 HCCL 或对应通信库，超参与拓扑配置需要重新调试，多机通信效率是国产集群性能差距的最大来源。

2. 量化推理工具链。 从 TensorRT 到昇腾 MindIE 或各家的量化工具，量化校准数据集和精度评估都要重做，生产环境的精度回归测试不可省略。

3. 生态外围工具。 DeepSpeed、vLLM、Transformer Engine 等常用库在国产平台上的版本支持滞后，很多团队卡在"模型能跑，加速库不兼容"这一步。

4. 运维体系。 现有监控（DCGM 类工具）、故障诊断、驱动升级流程全部要重建，这部分工作量经常被漏算。

5. 人才储备。 国产平台的调优经验集中在原厂和头部客户，招聘有昇腾或寒武纪实战经验的工程师成本高，提前安排团队参加原厂认证培训更实际。

四、结尾建议​

给准备迁移的团队三条建议：第一，先迁推理、后迁训练——推理负载对生态的依赖更浅，Atlas 300I Duo 这类推理卡性价比高，能用最小代价验证国产路线的可行性；第二，选硬件前先选生态，优先挑你的核心模型已有官方适配案例的平台，能省一半工作量；第三，保留双栈能力，迁移期让 CUDA 与国产平台并行运行至少一个季度，用同一份测试集做持续对比，确认精度与性能达标后再下线旧集群。国产化替代是一场马拉松，清单化管理、分阶段验收，才能避免"买回来跑不起来"的困境。

国产替代不是"把英伟达换成国产卡"一句话能概括的决策。合规压力、负载类型、迁移成本、供货确定性、预算五个维度都要过一遍，本文用站内芯片卡数据把每一步落到具体型号上。

一、替代决策五问​

动手选型前，先回答五个问题：

1. 合规要求：是政策强制的信创场景（政府、国企、金融），还是成本驱动的自愿替代？前者看《安全可靠测评》等资质，后者看性价比。
2. 负载类型：训练还是推理？Prefill 还是 Decode？不同负载的答案完全不同——昇腾 950PR 是推理 Prefill 卡，拿它做 FP16 密集训练就是选型错误（站内卡明确提示：FP16 训练应选 910C，其 800 TFLOPS 对 950 系列的 486–547 TFLOPS 保持领先）。
3. 生态迁移成本：现有 CUDA 代码库有多少算子依赖？团队有没有非 NVIDIA 平台经验？迁移成本有时比卡价更高。
4. 供货确定性：产能爬坡与 HBM 供应决定交付周期，头部型号已经在涨价排队。
5. 预算口径：国产卡价格已从"替代品折价"转向"供给稀缺溢价"，预算要按动态报价做。

二、国产主要选项对比​

以下数据全部取自本站芯片卡：

型号	架构	FP16/BF16 算力	低精度	显存 / 带宽	互联	软件栈	参考价
昇腾 910C	Da Vinci 双芯	BF16 800 TFLOPS	不支持 FP8/FP4	128GB HBM2e / 3.2 TB/s	HCCS 784 GB/s + UB 392 GB/s	CANN	渠道价超 ¥11 万
昇腾 950DT	Da Vinci v5	BF16 547 TFLOPS	FP8 1,034 TFLOPS、FP4 2,007 TFLOPS	144GB 自研 HiZQ 2.0 / 4 TB/s	灵衢 2 TB/s	CANN Next	渠道指示价超 ¥25 万
寒武纪 思元690	MLUarch 双 die	FP16 700+ TFLOPS	INT4 2,800+ TOPS	196GB HBM3 / 3.35 TB/s	MLU-Link 890 Gbps	NeuWare	约 ¥17.5 万
摩尔线程 MTT S5000	MUSA 平湖（第四代）	BF16/FP16 400 TFLOPS	FP8 稠密 1,000 TFLOPS	80GB / 1.6 TB/s	MTLink 784 GB/s	MUSA / MUSIFY	官方指导价 ¥5.5 万
天数智芯 天垓150	ivcore11	FP16 192 TFLOPS	INT8 384 TOPS	64GB HBM2e / 1.2 TB/s	PCIe 4.0	IXUCA	未公开
燧原 云燧T21	邃思2.0	FP16 未公开	TF32 160 TFLOPS	HBM2E（容量未公开）	自研多卡互联	燧原智算软件栈	约 ¥7 万

怎么读这张表：训练优先看昇腾 910C/950DT 与思元690（FP16 算力与显存容量领先，且都有千卡/超节点级方案——思元690 已落地 256 卡超节点，950DT 的 Atlas 950 SuperPoD 1024 卡版本已在 WAIC 2026 真机展示）；FP8 推理看 MTT S5000（FP8 稠密 1,000 TFLOPS 是国产卡中的亮点，且已通过《安全可靠测评》2026 年第 2 号）；视频与低功耗推理看昆仑芯 R200（150W、INT8 256 TOPS、108 路视频解码）；信创训练集群可评估云燧T21（OAM 标准模组，8000+ 卡集群已验证）。

三、软件栈迁移成本评估​

迁移成本是国产替代里最容易低估的一项，各家生态进展差距明显：

• CANN（华为）：进展最快。CANN Next 提供 CUDA 兼容层，约 80% 的标准 PyTorch 推理代码仅需配置修改；2026 年 10 月 DeepSeek 与华为联合发布面向昇腾 950 的开源编程工具与库。生态拐点已现——站内昇腾 960 卡口径显示，CANN 外部开发者占比升至 61%，首次超越内部团队，月活跃开发者突破 5,200 人；PyTorch 官网也已将昇腾收录为中国算力平台。
• NeuWare（寒武纪）：云边端统一一套代码是国内独一份，PyTorch/TensorFlow 适配版齐全，MagicMind 是全球首批商用 MLIR 图编译推理框架；短板在千卡以上分布式训练的工程成熟度，单机 8 卡与百卡小集群表现优异。
• MUSIFY（摩尔线程）：算子覆盖率约 70%–80%（CUDA 为 99%+），复杂 LLM 需手工优化，但 Day-0 适配能力快速进步——2026 年 6–7 月连续完成 MiniMax M3、智谱 GLM-5.2、Qwen3.5 和美团万亿参数 LongCat-2.0 的适配。
• IXUCA（天数智芯）：高度兼容 CUDA 生态，官方口径迁移时间下降 50% 以上，已适配 200+ 模型，覆盖 LLM、CV、NLP、语音等领域。

四、供货现实与价格走势​

供需关系是当前国产算力市场的第一现实：

• 需求侧：华为副董事长徐直军在 HC2026 访谈中的表态是"能产多少供多少"——需求远未见顶，第一约束已从市场转移到产能。
• 产能侧：站内 950DT 卡引用的行业口径显示，中芯国际产能可支撑年产超 100 万颗昇腾芯片，但国产 HBM 产能仅够约 25 万–30 万颗成品加速器，缺口约六成——瓶颈在 HBM 而非逻辑制程。
• 价格侧：HBM 成本攀升推高全线上调——950DT 渠道指示价超 ¥25 万（较两个月前 +20%–50%，Bloomberg 称三个月涨约 60%，已对标 B200）、950PR 突破 ¥8 万（+30%）、910C 超 ¥11 万（+22%）、寒武纪 690 等同期上调 20%–30%。预算必须按动态报价做，且注意以上均为渠道指示价，实际成交随采购规模与捆绑条件浮动。

五、混合部署策略与风险清单​

务实的企业路径是混合部署：NVIDIA 存量卡继续承担成熟训练负载与 CUDA 专有工作负载，国产卡承接增量需求——通常从推理切入（生态迁移成本最低），训练负载在 CANN/MUSA 成熟度验证后逐步迁移。昇腾 910C 集群已验证万亿参数大模型全参数后训练能力（站内卡记载的 DeepSeek V4 Pro 案例），训练迁移的技术可行性正在快速提高。

同时把风险清单贴在采购文档里：

1. 生态风险：算子覆盖不全导致的性能回退与排期延误，先做 POC 再批量采购。
2. 价格风险：涨价周期中锁价窗口短暂，长单尽量锁量锁价。
3. 供货风险：HBM 约束短期无解，关键项目准备备选型号。
4. 人才风险：非 CUDA 平台的工程师储备需要提前 6–12 个月建设。
5. 残值风险：国产卡二手市场流动性弱于 NVIDIA，折旧年限要单独评估。

小结​

国产替代选型的正确姿势是：用五问框架定位自己的真实约束，按负载类型从对比表里筛型号（训练看昇腾与思元690、FP8 推理看 MTT S5000、低功耗视频看昆仑芯），把软件迁移成本算进总预算，在涨价与供货的现实下混合部署、增量替换。更多型号的全字段对比见站内完整对比表，各卡详细规格请进入对应芯片卡页面。

2026 年，美国对华 AI 芯片出口管制完成了从"推定拒绝"到"逐案审查"的历史性转向，同时附加关税、产能与安全三重枷锁。管制形态从"按芯片型号划线"演变为"按买家总部划线"，全球算力供应链随之重新绘制版图。

一年之内：政策的密集摇摆​

时间	事件	影响
2025 年 4 月	H20 出口需申请许可	英伟达计提 45 亿美元相关损失
2025 年 5 月	AI 扩散规则被暂停	全球分级许可框架搁置
2026 年 1 月	H200 转为逐案审查	TPP 低于 21000 且 DRAM 带宽低于 6500GB/s 的芯片可申请许可
2026 年 1 月	同步加征 25% 关税	相关芯片出口附带"收入分成"
2026 年 5 月	D:5 总部指南生效	管制对象从产品延伸到买家总部所在地
2026 年 8 月	英伟达财报指引	不假设任何来自中国的数据中心计算收入

新规则并非简单放松：对华出口总量受"不超过同期对美供应总 TPP 一半"的产能配额约束，每批产品须经美国第三方实验室检测验证，云端用途还需提交远程终端用户清单。英伟达甚至修改付款条款，要求中国客户全款预付，把政策风险完全转嫁给买方。

中国市场的受控交织​

尽管约束重重，需求依然旺盛。据公开报道，中国企业累计订购的 H200 级别芯片超过 200 万颗，字节跳动单年度采购金额据报高达 140 亿美元。作为对价，中国工信部引入"配对"要求：企业每进口一定数量海外芯片，须按比例配套采购以昇腾为主的国产处理器，形成"国产芯片承担推理与通用负载、海外芯片保留高价值训练"的混合算力部署模式。中美监管的双重合规压力，成为跨国采购团队的新常态。

与此同时，管制反向催化了国产替代的加速。据公开报道，华为 AI 芯片年收入正逼近 120 亿美元；BIS 甚至发布指南，提示昇腾 910B、910C 等芯片在全球范围内的交易也可能触发美方许可要求——管制的长臂正从硬件延伸到远程算力访问。2025 年 11 月出台、2026 年 11 月到期的"关联规则"暂停令，则让实体清单的覆盖范围悬而未决，进一步放大了政策的不确定性。

供应链的连锁反应​

合规物流让每颗输华芯片先赴美国检测再出口亚洲，交期延长、成本上升。更深层的挤压在制造端：H200 与 Blackwell 共享台积电 CoWoS 先进封装产能，对华订单直接排挤更高毛利产品的排产。英伟达已锁定台积电 2026 年约六成先进封装产能，其余厂商只能在剩余份额中竞争。2026 年 5 月，黄仁勋向媒体表示英伟达"基本已让出中国 AI 芯片市场"——对芯片巨头而言，中国市场的财务贡献正在从增长项退化为风险项。

结语：版图的三条断层线​

2026 年的供应链版图上，三条断层线已然清晰：美国以"逐案审查加关税"维持受控开放；中国以"配对采购加国产放量"构建双轨算力体系；其余国家则在两种体系之间摇摆观望。管制已经无法阻止算力需求的增长，只能改变算力的地理分布与成本结构。对行业观察者而言，比门槛数字更值得跟踪的，是每一次政策摇摆背后，供应链各方重新谈判的定价权。

This article is based on market analysis published by Bloomberg Intelligence and Bernstein in September 2026 and public reporting; market share and revenue figures are analyst estimates, not official disclosures.

On September 28, two Wall Street reports simultaneously painted a dramatic picture: the balance of market share in China's AI compute market completed an almost total flip within 12 months.

1. The Share Flip: 40% to 8%, Ascend to 50%​

Bernstein's latest forecast:

• NVIDIA's share of the China AI chip market will fall from about 40% to about 8% by the end of this year
• Huawei Ascend rises to nearly 50%
• The immediate trigger: China's September ban on new H20 orders from NVIDIA, which shut the last channel for NVIDIA's return to the Chinese market
• ByteDance, Alibaba, and Tencent have all placed large Ascend orders

On the timeline, this was not a sudden shock but the cumulative result of three years of tightening export controls: NVIDIA's share of the China AI chip market once peaked at 95%; as high-end chips were cut off, Chinese hyperscale cloud providers pivoted wholesale to Huawei and domestic alternatives. Even when Washington allowed H200 sales to China in January 2026, the gears of that shift had already meshed — there was no going back.

2. Ascend 950PR: The "Chinese H200" in Analysts' Eyes​

What underpins this share data is product strength itself:

• The Ascend 950PR entered mass production in March this year, with FP4 compute of up to 2 PFLOPS and 128GB of domestic HBM
• Analysts rate it as roughly on par with the NVIDIA H200
• Huawei's AI chip revenue is expected to reach $12 billion this year, up 60% from $7.5 billion in 2025

For the full specifications of the 950 series, see our earlier in-depth articles: Ascend 950PR/950DT Dual-Configuration Analysis and Ascend 960 Official Launch — the 960 was ready three quarters ahead of schedule, confirming a one-generation-per-year cadence.

3. The Model Side: Chinese Open Models Take the OpenRouter Token Pie​

While chip shares flipped, the model usage data is equally striking (OpenRouter platform statistics):

Metric	Data
DeepSeek token share (June)	16.3%, surpassing Google, Anthropic, and OpenAI individually to rank first
Combined token share of Chinese open-weight models (May)	About 61% (DeepSeek, Qwen, MiniMax, Tencent Hunyuan, etc.)
Weekly token consumption of Chinese models	About 18 trillion, versus about 5.5 trillion for US models — a gap of more than 3x, with the overtake completed within a year
China-US frontier model performance gap (June)	Narrowed to 6%, a historic low (9% in May)

Compute and models form a positive feedback loop: cards that cannot be bought force domestic compute to scale up; scaled-up domestic compute produces cheap tokens; and cheap tokens let Chinese open-source models swallow the bulk of global inference traffic.

4. But the Stock Market Isn't Buying It​

Bloomberg Intelligence pointed out a contradiction in the same period: the valuation discount of the China Tech 8 relative to the US Magnificent Seven has widened to more than 50%, the widest of the year, and BI believes only a "true AI breakthrough" can close it. Year-to-date stock performance: Alibaba about -7%, Tencent about -23%, while US AI infrastructure leaders broadly gained more than 15%.

There are two readings of this divergence: either the market is underestimating China's AI fundamentals, or earnings execution (Alibaba's EPS missed expectations by 17.3% last quarter, Baidu's by 36%) has dragged down the delivery of the AI narrative. For industry observers, the notable point is: the compute-side data (share, revenue, shipments) has moved ahead, while application-side monetization is still on the way.

5. Three Takeaways for Compute Buyers​

1. For deployments in China, the Ascend ecosystem is already the default option. The cost of migrating from CUDA to CANN is one-time, while supply uncertainty is persistent; after the September ban, there is no compliant path to procuring new NVIDIA cards in the Chinese market.
2. Domestic compute "matching the H200" is a watershed. The implicit assumption of domestic substitution used to be "discounted performance, discounted price"; when single-card performance catches up to the H200 generation, procurement decisions return to pure TCO and supply-security calculations — we recommend using the TCO Calculator to put electricity, depreciation, and utilization into the same table.
3. The token cost curve determines the application landscape. Chinese models' 61% share of OpenRouter tokens shows that the chain of compute self-sufficiency leading to token price drops leading to application prosperity has completed its first leg; the next question is whether inference costs can keep falling.

Summary​

From 40% to 8% is not a slogan but a market reshuffle driven jointly by the ban, product strength, and the model ecosystem. For the Chinese market, the mass-production cadence of Ascend 950/960 and the maturity of the CANN ecosystem will determine whether this share can be held; for the global market, the combination of China's compute "internal circulation + open-source models going overseas" is rewriting the geographic distribution of inference traffic.

(Market share and revenue forecasts come from Bernstein / Bloomberg Intelligence analyst reports; actual figures are subject to each company's financial reports.)

Beyond the official Ascend 960 unveiling and the Zhenwu V900 launch, late September also delivered three ecosystem signals that headline coverage easily drowned out, yet whose structural significance rivals flagship chips: a carrier open-sourcing a model trained on Ascend, an OEM delivering a hundred-chip-class domestic inference system, and a cloud-compute chip startup closing a new funding round. This article breaks down each one.

---

1. China Telecom Open-Sources Xing4.0: Carrier-Grade Endorsement for End-to-End Ascend Training​

On September 23 (as reported), China Telecom open-sourced Xing4.0-29B-A4B — an agentic mixture-of-experts (MoE) model that the company says was trained end-to-end on Huawei Ascend accelerators.

Why it matters:

• Third-party proof that "it can train": until now, evidence that "Ascend can train large models" came mainly from Huawei itself (40+ natively trained models); a carrier running training with its own data, engineering teams, and clusters — and open-sourcing it — is the first scaled training case outside the Huawei ecosystem;
• Agentic positioning: the 29B-A4B (29B total / 4B active) MoE form factor targets agent workloads directly — aligned with the open-source cadence of DeepSeek and Qwen, rather than a small experimental model;
• Open-source spillover: publishing model weights plus the training recipe means Ascend training engineering know-how can be reused by other institutions — the standard playbook for ecosystem diffusion.

Combined with Huawei's disclosure that "CANN has entered routine open-source operations, with external developers at 61%", the Ascend ecosystem is shifting from "vendor-led" to "community-operated".

2. Inspur MetaBrain SD200 Ultra: Measured Throughput Claims for a 128-Chip Domestic Inference System​

Inspur launched the MetaBrain SD200 Ultra inference system around the same time:

Metric	Value (company figures)
Domestic AI chips	128 chips (vendor did not disclose specific models)
Model served	Kimi K3
Token throughput	2.8T tokens (system throughput figure under that methodology)
Latency	5.85 ms

Three readings:

• Hundred-chip-class system integration: 128 chips cooperating across a heterogeneous setup to run very large MoE inference is a test of interconnect topology, scheduling, and fault tolerance — OEMs are now demonstrably capable of assembling domestic chips into large systems;
• Aimed at top open-source models: targeting Kimi K3 (already live on AWS Bedrock) as the benchmark workload shows the acceptance standard for domestic inference systems is "runs today's most popular models", not a self-referential demo;
• Mind the methodology: the 2.8T token throughput and 5.85ms latency are both vendor claims; the tested model version, concurrency, and batch size were not disclosed, so procurement evaluations should re-test under real workloads.

3. EVAS Closes RMB 2 Billion B+ Round: The Cloud Compute Chip Race Still Attracts Capital​

EVAS (Yixing Intelligence) completed a B+ round of RMB 2 billion, at a post-money valuation near RMB 15 billion, with Oriza Capital following on; earlier in the first half of the year it had closed a RMB 1.5 billion Series B and brought in China Mobile as a strategic investor. The company focuses on next-generation cloud compute chips.

In 2026, when the flagship chip lineup (Ascend / Cambricon / Moore Threads / Hygon / Zhenwu) looks settled, a cloud compute chip newcomer still raising RMB 2 billion suggests:

• Primary-market conviction in the "second tier of domestic compute": with top vendors' capacity booked into 2027, overflow demand gives newcomers a window;
• The role of carrier strategic investment: China Mobile is both a buyer and an investor — the demand side of domestic compute is using capital to lock in the supply side;
• Differentiated room remains for cloud inference DSA routes (see the ecosystem positioning of routes like Tsingmicro and Houmo).

4. Assembling the Week's Signals: The Loop Is Taking Shape​

Put the three signals together with this month's flagship launches, and every segment of the domestic compute loop now has players filling the gaps:

Segment	This Month's Evidence
Flagship chips	Ascend 960 early official unveiling (9-17), Zhenwu V900 launch (9-22)
Training validation	China Telecom's Xing4.0 end-to-end trained on Ascend and open-sourced, DeepSeek betting on Ascend training
Inference systems	Inspur SD200 Ultra running Kimi K3 on 128 chips
Software ecosystem	CANN routine open-sourcing, external developers at 61%
Capital supply	EVAS B+ round of RMB 2 billion (post-money near RMB 15 billion)
Demand side	DeepSeek's alignment, Kimi K3 live on overseas cloud platforms

Independent players are investing across all four segments — "chips, systems, models, capital" — the biggest difference from the "isolated breakthroughs" of 2024-2025.

5. Takeaways​

• China Telecom Xing4.0-29B: the first open-source model end-to-end trained on Ascend outside the Huawei ecosystem;
• Inspur SD200 Ultra: a 128-chip domestic inference system at 2.8T token throughput / 5.85ms (company figures, pending re-testing);
• EVAS RMB 2 billion B+ round: the cloud compute chip second tier is still getting real money;
• What to watch: Xing4.0's training cluster scale and MFU, third-party re-tests of the SD200 Ultra, and EVAS's tape-out cadence.

---

Further Reading​

• Ascend 960 Spec Page / Ascend 910C Spec Page (this site)
• DeepSeek Confirms Bet on Huawei Chips for Large-Model Training
• Domestic Top Three in 2026 H2: Localization Rate Breaks 40% and Marches Toward 60%

References​

• The GPU Daily (2026-09-24): China Telecom open-sources Xing4.0-29B; Inspur MetaBrain SD200 Ultra
• Toutiao (2026-09-21): EVAS completes RMB 2 billion B+ round, post-money valuation near RMB 15 billion
• CSDN AI Daily (2026-09-21): Huawei CANN enters routine open-source operations, OceanStor M900 launched

This article is compiled from public reports. Xing4.0 training details and SD200 Ultra performance figures are company statements; testing conditions are subject to subsequent disclosures.

In 2026, China's AI chip market landscape has shifted from "NVIDIA unipolar dominance" to "overseas vendors leading, domestic multi-route catch-up." According to industry research, China's overall AI accelerator market was ~4M units in 2025, of which 1.65M were domestic, with share first breaking 40%; as products iterate and fabs follow up, the localization rate is expected to rise to 60%-70% by 2027. This article focuses on the latest H2 2026 progress of Huawei Ascend, Cambricon, and Moore Threads — the domestic "Big Three."

---

1. Huawei Ascend: 950 Capacity Fully Booked, 960 Roadmap Unveiled​

Ascend's core advantage is "architecture + full-stack ecosystem synergy," with ~800K units shipped in 2025, capturing 50% of the total domestic vendor share. The product iteration cadence is clear:

Time	Product	Note
2025 Q1	Ascend 910C	Main transitional model
2026 Q1	Ascend 950PR	Inference flagship
2026 Q4 (planned)	Ascend 950DT	Training flagship, drives domestic HBM iteration
2027-2028	Ascend 960 / 970	Roadmap products

950 series capacity has entered a "fully booked" state: 950PR entered mass production in April 2026; June monthly capacity jumped to 500K-600K units (nearly 10x MoM), with a full-year target of 1.2M units at 100% certainty; ByteDance locked in 350K units for $5.6B, while Tencent / Alibaba / Baidu combined locked in 400K units.

Ascend 960 roadmap specs (per roadmap disclosure):

Metric	Ascend 960
Architecture	Ascend 6th gen (Da Vinci v6)
FP8 compute	~4 PFLOPS
Memory	288GB
Memory bandwidth	9.6 TB/s
Super-node	Atlas 960 SuperPoD, 15,488 cards, Lingqu optical-electrical converged bus
Debut	2027 Q4 (roadmap)

The previous-gen Ascend 384 super-node has cumulatively shipped over 750 sets, deployed across 20+ industries including internet, operators, finance, education, and healthcare — Huawei calls it "the only domestic super-node that has trained a SOTA model."

---

2. Cambricon MLU690: H2 Mass Production, Entering ByteDance Bidding Window​

Cambricon is the core domestic compute leader in the absence of an Ascend IPO, with the technology gap continuously narrowing:

• Siyuan 590 (7nm): Performance equivalent to 80% of A100, already supports DeepSeek, continuously adapting to mainstream large models like Qwen 3 and GLM
• Siyuan 690 series: Will enter mass production in H2 2026, expected to achieve order scale-up during ByteDance's H2 bidding window
• Revenue certainty: Equity incentive targets show >100% revenue growth for the next 3 years: 2026 revenue target 13.5B RMB, 2027 27B RMB, 2028 60B RMB

Cambricon fully benefits from the industry dividend of "domestic CSP capex + full adaptation of domestic large models and domestic chips," making it the most direct elasticity play on rising localization rate.

---

3. Moore Threads MTT S5000: Full-Function GPU + Ecosystem Breakthrough​

Moore Threads takes a differentiated "full-function GPU" route, with the flagship MTT S5000 based on the 4th-gen "Pinghu" MUSA architecture:

Metric	MTT S5000
Dense AI compute	1000 TFLOPS
Memory	80GB
Memory bandwidth	1.6 TB/s
Inter-card interconnect	784 GB/s
Precision	FP8 to FP64 full precision (training + inference)
Security	First batch to pass national "Safe and Reliable Evaluation" (Level I)

Its engineering capability is verified: the Kuae (KUAE) intelligent computing cluster based on S5000 achieves 95% training linear scaling efficiency, with compute efficiency loss within 5% at ten-thousand-card scale; supports checkpoint-resume training with effective training time ratio >90%; and has trained a MoE-236B base model with >25 trillion tokens of corpus from scratch.

The ecosystem is Moore Threads' deepest moat: MUSA has achieved 100% core math library compatibility, 3000+ PyTorch operator compatibility, covers 55 categories of core AI operators, has official vLLM and SGLang support, Day-0 adaptation of mainstream models, and 800K+ developers. Its PD heterogeneous-disaggregation solution achieves equivalent replacement of international high-end GPUs at a 2:1 ratio with S5000, significantly reducing inference cost.

The 5th-gen "Huagang" architecture (released 2025-12) supports FP4 to FP64 full precision, with 50% higher compute density and 10x better energy efficiency than the previous gen, supporting 100K+ card clusters; cumulative R&D investment in the "Huashan" (train-infer integrated) and "Lushan" (graphics rendering) new chips based on this architecture exceeds 900M RMB.

---

4. Software Ecosystem Decides: Day-0 Adaptation Becomes Routine​

Beyond hardware, software ecosystem realization is the watershed for domestic compute in 2026:

• Huawei's CANN heterogeneous computing architecture and MindSeries suite are fully open-sourced, with the community incubating 67 projects, 12.44M+ lines of code, and 3,500+ monthly active developers
• The "release-and-adapt" closed loop between domestic large models and domestic chips has basically formed: Tencent Hunyuan T3 (295B), DeepSeek-V4, and GLM-5.2 all completed Day-0 adaptation
• 2026 is regarded as the "first year of domestic super-nodes"; Huatai Securities estimates China's super-node architecture market will reach 341.4B RMB by 2028, with a 2026-2028 CAGR of 194%

---

5. Industry Judgment: From "Can It Be Built" to "Can It Be Used Well"​

The domestic Big Three are converging along three paths:

1. Huawei: Locks government/enterprise and internet big customers with super-node system-level capability + full-stack software
2. Cambricon: Impacts the revenue inflection point by narrowing the training-side gap + scaling up via big-customer bidding
3. Moore Threads: Covers cloud-edge-end full scenarios with full-function GPU generality + mature CUDA-compatible ecosystem

The common shortcoming of all three remains advanced process and HBM supply — precisely the core link of overseas controls. But as domestic HBM iterates and fabs follow up, a realistic path to 60%-70% localization by 2027 exists.

Related Links​

• Huawei Ascend 960 spec page
• Huawei Ascend 950DT spec page
• Cambricon MLU690 spec page
• Moore Threads MTT S5000 spec page

References​

• Domestic AI Chip Industry View Update - Robo Datayes
• Moore Threads, Two Positives in One Day - Eastmoney
• Moore Threads × SiliconFlow: PD Disaggregation Heterogeneous Compute Co-location Whitepaper - EE Focus
• After Approaching Break-even, Moore Threads Faces Hong Kong "Big Exam" Again - Top News

---

This article is compiled from public industry research, broker views, and corporate announcements as of August 2026. Some shipment and market-share figures are third-party estimates, not officially confirmed data.

The Ascend 950 series (950PR inference / 950DT training) has become the core supply of domestic AI compute. Per multiple brokerages and industry research, 950 series capacity is 100% booked with scarce spot supply; the full-year 1.2M-unit target is "100% certain," with expectations of an upward revision to 1.5M. This article summarizes capacity and order data as of July 2026.

1. Capacity pace: ~10× MoM jump in June​

Time	950PR monthly capacity	Notes
May 2026	50k-60k units	Near full production
June 2026	500k-600k units	~10× MoM; SMIC, Hua Hong tier-1 suppliers on overtime
Q3 2026 (est.)	700k-800k units	Per month
Full-year 2026 target	1.2M units	Upward revision to 1.5M expected

Supply chain delivery is tight: high-speed backplanes and liquid-cooling connectors' lead time stretched from 2 weeks to 6-8 weeks; orders are booked into 2027.

2. Order structure: top cloud providers + operators + overseas​

Customer	Locked volume	Amount / Notes
ByteDance	350k 950PR	$5.6B, concentrated delivery from Q3 2026
Tencent / Alibaba / Baidu	~250k 950PR + 150k 950DT	Combined ~400k units
Three major operators	200k+ units	Centralized procurement, for intelligent compute centers and AI private networks
Overseas	South Korea 2,000 units, Malaysia 3,000 servers, Russia ten-thousand-card cluster	From pilot to commercial

3. Shipment forecast: firmly #1 domestic​

Per CCA (Kezhi) Consulting estimates:

Metric	2025	2026 (forecast)
Huawei Ascend total shipments	812k cards	1.026M cards
Of which 950PR	—	~800k units
Of which 950DT	—	~100k-200k units

Huawei has completed the product transition from the 910 series to the 950 series. The internet industry has become Ascend's largest application market; competitive advantage is extending from single-hardware performance to software ecosystem and system capabilities.

4. Going overseas: formal South Korea entry in Q4​

Per Korean media ETNews, Huawei plans Q4 2026 to formally enter the South Korean market with the Ascend series and Atlas 950 SuperPod:

• Local distributor agreements signed; two channel partners including SK Shieldus selected
• Main products: 950PR (mass-produced and delivered since April) and 950DT (launched Q4)
• Official line: 950PR inference performance is 2.87× that of H20, priced at about 1/4 of it

5. WAIC 2026: 1024-card live debut confirmed​

At WAIC 2026 (July 17-20, Shanghai), Huawei's Atlas 950 SuperPoD live hardware made its first public appearance — a 16 compute-cabinet, 1,024 Ascend-card scale — and won the conference's top honor, the SAIL Award:

• Core metrics: total compute 1 EFLOPS FP8 / 2 EFLOPS FP4, 256 TB globally unified memory addressing, Lingqu 2.0 interconnect, 3 μs ultra-low RTT latency
• Full configuration: 128 compute cabinets + 32 interconnect cabinets = 160 cabinets, ~1000㎡ footprint, carrying 8,192 Ascend 950DT, planned Q4 2026 launch
• Commercial foundation: previous-gen 384 SuperNode has cumulatively shipped 750+ units, deployed in 20+ industries
• Software ecosystem: CANN fully open-sourced end of 2025; community incubated 67 projects, 12.44M+ lines of code, 3,500+ monthly active developers

WAIC's debut confirmed the 950 series' "SuperNode-first" product logic: beyond single-card compute, system-level effective compute (interconnect bandwidth + unified memory + low latency) is the key dimension for domestic compute to benchmark against international flagships.

Ascend roadmap recap​

Product	Positioning	Key metrics (official roadmap)
950PR	Inference	1 PFLOPS (FP8) / 2 PFLOPS (FP4), 2 TB/s interconnect
950DT	Training	SuperNode core, launched Q4
960	Train/inference	2 PFLOPS (FP8) / 4 PFLOPS
970	Next-gen	In planning

Industry interpretation​

1. Domestic substitution moves from inference to training: 950PR (inference) ramps first, 950DT (training) follows in Q4, combined with the Atlas 950 SuperPoD ten-thousand-card interconnect — domestic compute now has the complete "training substitution" puzzle for the first time.
2. Capacity is the biggest variable: order certainty is extremely high, but SMIC/Hua Hong advanced-process capacity, HBM supply, and advanced packaging remain ramp bottlenecks — the root of "scarce spot supply."
3. Going overseas opens a second growth curve: bulk procurement from South Korea, Malaysia, Russia, and Latin America marks domestic compute's shift from "internal circulation" to "external circulation."

Related links​

• Huawei Ascend 950PR specs
• Huawei Ascend 950DT specs
• Huawei Ascend 950 specs

References​

• China intelligent-compute chip competitive landscape (CCA data) - IDCquan
• Ascend 950 series capacity and orders - Eastmoney
• WAIC catalyst: Huawei Ascend 950 SuperNode live debut - 10jqka
• Huawei Ascend 950 enters South Korea in Q4: 2.87× H20 inference - Toutiao

---

Data in this article is based on official and major brokerage research; capacity/orders are dynamic figures and will be continuously updated.

Escalating U.S. export controls are forcing China's AI chip industry to accelerate self-reliance. By 2025, the discussion around domestic Chinese AI chips has shifted from "are they usable?" to "which one should I choose?"

This article systematically reviews the major players, core products, and actual deployment status of domestic AI chips, helping developers and procurement decision-makers understand the competitive landscape.