# 产业日报 2026-09-18

来源范围：`/workspace/intel/archive/2026-09-18/`（industry/ + hpe-morpheus/nutanix/smartx/redhat 等 vendor 旁证）。**主轴**为 **Gartner Magic Quadrant for Server Virtualization Platforms（报告日 14 Sep 2026）** 象限簇扩容——本日新增 **HPE = Challenger**（Connect Converge 伙伴/媒体，16 Sep 2026；Morpheus/VME 叙事）；此前已知厂商博客自称：**Leaders** Nutanix + RH OpenShift Virtualization；**Visionaries** SUSE + Platform9。**本地仍无完整 MQ PDF**，不编造 Broadcom/VMware 或其他未声明厂商象限/轴分。并行 NEW：Nutanix 产品博客（NDB on NC2 / RDS Custom Oracle EOS、Neocloud AI 风险、统一运维）、**SmartX 榫卯 AI 1.2**、**RH Alquimia × OpenShift AI** agent 评测栈。industry/ 另大量再抓 Register **Feb–Aug 2026 旧稿**（Cisco hypervisor 传闻、半数 VMware 用户拟减量、AWS nested virt、Western Union→Nutanix 等）——一律标 **再抓旧稿 / 续档指针**，不作 09-18 战略新发。Astera/onsemi/Gartner HV×存储/Google agentic = 续档。数字未独立核实标「厂商自称」「研究推算」或「媒体转述」。

## 摘要

今日产业信号可以收成六句话：

1. **MQ 簇扩容：HPE 被点名为 Challenger（Connect Converge，16 Sep 2026）。** 文面：HPE「named a Challenger in Ability to Execute and Completeness of Vision」于 2026 Gartner MQ for Server Virtualization Platforms；叙事绑「不到两年推出虚拟化替代」与 **HPE Morpheus Software** / 客户信任（**伙伴/媒体转述 + HPE 视角，非 PDF 核验**）。叠加已知：**Leaders** Nutanix、RH OpenShift Virt；**Visionaries** SUSE、Platform9（均为**厂商博客自称**）。**仍无完整 MQ PDF；Broadcom/VMware 及其他未声明象限一律不写。**
2. **Nutanix 产品博客三连（非 MQ 新象限）：** (a) **NDB on NC2** 对标 AWS **RDS Custom for Oracle EOS 2027-03-31**，主张保留 OS 级访问 + 自动化生命周期（**厂商产品主张**）；(b) **Neocloud AI 七维风险**（基础设施/网络/数据/安全/合规/运维/环境功耗）推 hybrid 治理（**厂商白皮书式博客**）；(c) **统一混合云与 AI 运维**——HCI 从硬件一体升格为「运营哲学」、外置存储/GPU/云/边缘一致策略（**厂商叙事**）。
3. **SmartX 榫卯 AI 平台 1.2 产品发布（中国本地 HCI AI 栈信号）：** 自研 **Flex Engine**（MinerU/PaddleOCR 等非 LLM）；昇腾 **910B/910C** + **vNPU** 切分与可观测；模型仓库可视化 + **自动估算 KV-Cache/显存**；APIKey 按 Project 分组。附匿名船舶制造/医疗案例——**厂商产品与案例自称，非独立 VoC。**
4. **RH Alquimia × OpenShift AI：企业 agent 舰队评测栈（Gaussia / EvalHub / MLflow）**——主权边界内行为基准与安全探针；属 **agent-hosting / openshift** 旁证，**非新 hypervisor SKU。**
5. **再抓旧稿（非今日 NEW）：** Register 等 Feb–Aug 2026 条目（Cisco 自研 hypervisor 传闻 02-16、半数 VMware 用户拟 2028 前减量 03-24、AWS nested virt 02-17、Western Union→Nutanix 04-08、NodeWeaver/Xen 等）及 Astera Leo、Gartner HV×存储、Google agentic、onsemi EPP——作续档指针，**勿抬成 09-18 战略发布。**
6. **对规划：** 早报主抬 **「MQ 短名单语汇从双 Leader + 双 Visionary 扩到含 HPE Challenger」**——换平台买家可引用的分析师地图更完整，但仍须标「厂商/伙伴自称、无 PDF」。并列中国 **HCI 上 AI 平台版本化（SmartX 1.2：异构 NPU + KV 预估）** 与 **OpenShift AI 上 agent 评测工业化**；Nutanix NDB/Neocloud 作数据库迁云与 AI 基础设施风险旁证。**勿发明 Broadcom 象限；勿把旧 Register 再抓当新发。**

## 原文摘录

### 1. Gartner MQ · Server Virtualization Platforms · HPE = Challenger（新增）

- 归档：`hpe-morpheus/hpe-named-a-challenger-in-2026-gartner-mq-svp-12549764.*`
- URL：https://connect-converge.com/hpe-named-a-challenger-in-2026-gartner-magic-quadrant-for-server-virtualization-platforms/
- 来源：Connect Converge（HPE 用户社区/伙伴媒体）；作者 Brandi Boyd；**文面日期 September 16, 2026**（meta `article:published_time` 2026-09-16）
- 口径：**伙伴/媒体转述 + HPE 视角**；本地 **无** MQ PDF；不编造轴分坐标或其他厂商位置

> HPE has been named a Challenger in Ability to Execute and Completeness of Vision in the 2026 Gartner® Magic Quadrant™ for Server Virtualization Platforms. This recognition comes less than two years after announcing HPE’s intent to launch a virtualization alternative. In HPE’s view, this recognition reflects not only product innovation, but also the trust customers and partners have placed in HPE Morpheus Software to run business-critical applications and modernize their virtualization environments.

#### 1b. 已知象限声明（续；非本日新发 PDF）

| 自称定位 | 厂商 | 本地主要归档（本日或 09-17 再抓） |
|---|---|---|
| Leader | Nutanix | `nutanix/nutanix-named-leader-in-gartner-mq-for-server-virtualization-nutanix-aae2f686.*`；PR `nutanix/nutanix-pr-leader-gartner-mq-svp-31f163a6.*` |
| Leader | Red Hat（OpenShift Virtualization） | `redhat/red-hat-named-a-leader-in-the-2026-gartner-magic-quadrant-for-server-v-fda46b90.*` 等 |
| Visionary | SUSE Virtualization | 见 09-17 日报 / `suse-virtualization/` |
| Visionary | Platform9 | 见 09-17 日报 / `platform9/` |
| Challenger | **HPE（Morpheus）** | 本日 NEW，上节 |

报告引用日：**14 September 2026**（Nutanix PR 等文面：Tony Harvey et al.）。**Broadcom/VMware 象限：本地源未声明 → 不写。**

### 2. Nutanix 产品博客 · NDB / Neocloud / 统一运维

#### 2a. NDB on NC2 · RDS Custom for Oracle EOS

- 归档：`nutanix/beyond-self-managed-ec2-ndb-on-nc2-for-rds-custom-for-oracle-databases-a8dcb329.*`
- URL：https://www.nutanix.com/blog/ndb-on-nc2-for-rds-custom-for-oracle-databases
- 口径：**厂商产品博客**

> AWS has announced that Amazon RDS Custom for Oracle will reach end of support on **March 31, 2027**. Customers can continue using the service as-is until then; after that date, RDS Custom for Oracle database instances, snapshots, and custom engine versions will no longer be accessible. AWS's own prescriptive migration guidance recommends transitioning affected workloads to self-managed Amazon EC2.  
> … the Nutanix Database Service (NDB) solution running on Nutanix Cloud Clusters (NC2) in AWS is a solid alternative… It preserves the OS-level access that made RDS Custom for Oracle valuable… while keeping the database lifecycle automation…

#### 2b. Neocloud AI 风险评测框架

- 归档：`nutanix/how-to-evaluate-neocloud-risks-for-enterprise-ai-nutanix-08ba55de.*`
- URL：https://www.nutanix.com/blog/7-enterprise-ai-infrastructure-risks-evaluate-before-choosing-neoclouds
- 口径：**厂商博客**（推 hybrid + NCP/Enterprise AI）

> Enterprise requirements often outpace the operational maturity of specialized compute providers. While neoclouds accelerate initial testing, artificial intelligence (AI) workloads create technical pressure where architectural gaps surface late in production. Assessing risk across six core dimensions can help maintain workload performance, cost efficiency, and compliance.  
> 1. Infrastructure Risk: GPU Access is Not the Full Architecture  
> … Storage performance, orchestration maturity, hardware refresh cadences, and multi-tenancy isolation dictate operational stability.

（文内展开网络/数据/安全/合规/运维/环境功耗等维；末段推 Nutanix Sizer / Carbon and Power Estimator / NCP——**厂商主张**。）

#### 2c. Unify infrastructure operations

- 归档：`nutanix/unify-infrastructure-operations-for-hybrid-cloud-and-ai-874cb4fd.*`
- URL：https://www.nutanix.com/blog/unify-infrastructure-and-operations-to-minimize-silos
- 口径：**厂商叙事**（引用第三方 Gartner 云趋势 PR 数字时标研究推算/转述）

> In the past, hyperconverged infrastructure (HCI) simplified operations by tightly integrating compute, storage, and networking into a unified platform. That principle remains valuable, but its application has changed.  
> Instead of viewing hyperconvergence purely as a hardware architecture, think of it as an **operational philosophy**. The underlying infrastructure may now include external storage, third-party networking, GPUs, cloud resources, and edge systems.

### 3. SmartX · 榫卯 AI 平台 1.2

- 归档：`smartx/榫卯ai平台1-2-更丰富的模型与ai芯片支持-更简单的资源规划与管理-0b34eed0.*`
- URL：https://www.smartx.com/blog/2026/09/sunmao-ai-release-1-2/
- 口径：**厂商产品发布**；案例为匿名船舶制造/医疗——**非具名独立 VoC**

> 近日，SmartX 正式发布榫卯 AI 平台 1.2 版本…  
> - 模型支持：引入自研推理引擎 **Flex Engine**，支持 MinerU、PaddleOCR 等非大语言模型…  
> - 算力支持：扩展对昇腾 NPU 的适配范围…新增对 **910B、910C** 等型号的支持。同时，支持…对昇腾 NPU 进行切分（**vNPU**）…  
> - 运维管理：通过模型仓库可视化管理、**自动计算 KV-Cache 与显存需求**以及 APIKey 分组管理…

> 某船舶制造用户…某医疗用户…（**厂商案例，客户匿名**）

### 4. Red Hat · Alquimia × OpenShift AI（agent 评测）

- 归档：`redhat/scaling-enterprise-ai-fleets-with-alquimia-and-red-hat-openshift-ai-0cba7962.*`
- URL：https://www.redhat.com/en/blog/scaling-enterprise-ai-fleets-alquimia-and-red-hat-openshift-ai
- 口径：**厂商联合方案博客**；**非新 hypervisor SKU**；映射 agent-hosting / openshift

> Alquimia's agentic platform running on Red Hat OpenShift AI and Red Hat OpenShift provides this functionality.  
> By combining Alquimia's **Gaussia** evaluation framework, **EvalHub** job orchestration, and **MLflow** run-history logging, this solution turns unmonitored agent conversations into repeatable, parallelized benchmark jobs… all within a protected, sovereign infrastructure boundary.

### 5. 再抓旧稿 / 续档指针（非 09-18 NEW 战略发布）

本日 `industry/` 对下列 **原发日期早于本周** 的 Register 等页有再抓取；**只作指针，不升格为今日战略新稿**（部分 txt 与同批页共享相同 sha256，正文可用性存疑，以 URL 原日期为准）：

| 主题 | 原 URL 日期（路径） | 归档 |
|---|---|---|
| Cisco 自研 hypervisor 传闻 | 2026-02-16 | `industry/cisco-set-to-release-home-brew-hypervisor-as-a-vmware-alternative-395bfd9a.*` |
| Half of VMware users plan to reduce usage by 2028 | 2026-03-24 | `industry/half-of-vmware-users-plan-to-reduce-usage-by-2028-cb33bd28.*` |
| AWS nested virtualization（部分 EC2） | 2026-02-17 | `industry/aws-adds-nested-virtualization-option-for-handful-of-ec2-instances-aa62ba13.*` |
| Western Union → Nutanix | 2026-04-08 | `industry/western-union-zaps-vmware-and-moves-to-nutanix-1735855a.*` |
| NodeWeaver perpetual licensing | 2026-04-16 | `industry/nodeweaver-says-its-perpetual-licensing-beats-vmware-s-perpetual-price-8ffef69d.*` |
| Xen 5 年支持 / safety 分区 | 2026-03-10 / 2026-08-18 | `industry/xen-project-*` |

续档（非推翻稿）：Astera Leo 2/X、Gartner HV×存储、Google agentic、onsemi EPP、Open Secure AI→LF、Nutanix agentic/DHI MQ 再抓、Proxmox 24/7、VDDK——交叉 `daily/2026-09-14.md`–`2026-09-17.md`。

### 6. 跳过项

- 0 字节/壳：`industry/virtualization-831bfd76.*` → **跳过**
- Dell Pro 5 笔记本、Infineon 汽车 RISC-V、至顶网/奇客/科技行者频道壳 → **不抬**
- Nutanix vs VMware 营销对比页、sitemap、社区壳 → **不抬为主轴**

## 映射提示

- MQ 簇 + HPE Challenger → `industry/emerging.md` / `index.md`（轻）；`topics/migration.md`、`topics/cost.md`、`topics/cloud-native.md`；**勿**写 Broadcom 象限
- Nutanix NDB / Neocloud / unify ops → `topics/ai-ops.md`、`topics/cost.md`、`topics/storage.md`（轻）；vendor 轨另详
- SmartX 榫卯 AI 1.2（Flex Engine / 昇腾 vNPU / KV-Cache）→ `topics/inference.md`、`topics/asic.md`、`industry/hardware.md`（昇腾段轻刷新）、`topics/agent-hosting.md`（轻）
- RH Alquimia × OpenShift AI → `topics/agent-hosting.md`、`topics/openshift.md`；**非**新 hypervisor
- Register Feb–Aug 再抓 → 仅作历史指针；network / security / scale-ops / edge-inference：本日无战略级独立新轴