NVIDIA Vera Rubin/VR200 NVL72 的主要更新
(结合我最新的供应链调查和英伟达CEO黄仁勋在2026年国际消费电子展上的主题演讲)
1. NVIDIA 已将 AI 服务器 VR200 NVL144(基于芯片)更名为 VR200 NVL72(基于封装)。
我的供应链调查显示,VR200 NVL72 将提供两种功耗配置:Max Q 和 Max P。
➢ Jensen 引用了 NVL72,而不是之前使用的 NVL144 命名。
➢ Max Q 和 Max P 采用相同的硬件设计。
➢ Max Q GPU/机架功耗(TGP/TDP):~1.8/190(kW)。
➢最大 P GPU/机架功率(TGP/TDP):~2.3/230(kW)。
➢两者均明显高于 GB300 NVL72 的 1.4/140 (kW)。
➢提供两种功率模式可提高数据中心电力限制下的部署灵活性。
这也表明 NVIDIA 越来越重视实际物理基础设施的限制,并将这些限制纳入产品规格中。
2. VR200 NVL72 升级了 MaxQ 和 MaxP 的 GPU 散热解决方案。
➢两者都使用微通道冷板(MCCP)和镀金盖。
➢市场一直期待GPU TGP达到2.3kW时会转向微通道盖(MCL);然而,MCL的大规模生产不太可能在2027年下半年之前实现。
3. 得益于 GPU、HBM 和 NVLink 交换机的升级,
VR200 NVL72 的训练/推理 AI 计算能力约为 GB300 NVL72 的 3.5 倍/5 倍,从而导致机架级电源需求急剧增加。
➢ VR200 NVL72 将电源架升级为 3*3U 110 kW(6*18.3 kW 电源),
而最常见的 GB300 NVL72 配置为 8*1U 33 kW 电源架(6*5.5 kW 电源)。
➢ VR200 NVL72 电源架采用 3+1 冗余设计。
➢ VR200 NVL72 将功率鞭状电流额定值提高到 100A(而 GB300 NVL72 为 60A),
提高了数据中心电源基础设施的要求(例如,母线槽和分接盒)。
4. VR200 NVL72 更加依赖液冷散热。
➢计算托架和 NVSwitch 托架均采用无风扇设计。
➢机架技术冷却系统流量(TCS 流量)几乎翻了一番(~+100%) ,与 GB300 NVL72 相比,
有利于 CDU、歧管、冷板和快速断开装置 (QD) 的规格和/或数量升级。
➢机架气流需求比 GB300 NVL72 下降约 80%(以 CFM 为单位)。
5. VR200 NVL72 计算托架首次采用中板,实现了真正的无电缆设计。
➢关键中平面规格包括 44 层(22+22) 、M9 CCL(EM896K3)和大约 420×60 毫米的尺寸。
➢ Jensen 指出,这种设计可以将计算托架组装时间从大约 2 小时缩短到大约 5 分钟(与 GB300 NVL72 相比)。
6. Rubin CoWoS 预计到 2026 年将生产 30 万至 35 万片晶圆,预计将于 2026 年第一季度初开始试生产,
并于 2026 年第二季度末开始量产。VR200 NVL72 机架组件预计将于 2026 年第三季度末进入量产阶段;
考虑到良率提升,预计 2026 年下半年机架出货量约为 5000 至 7000 台。
7. 当机架功率接近约 200 kW 时,54V 配电在空间(母线/电缆)和转换效率方面开始面临材料限制。
因此,VR200 NVL72 可以被视为 Oberon 机架架构的最后一代产品。
为了应对 AI 计算持续扩展带来的机架功率需求增长,
NVIDIA 预计将过渡到支持 800V 高压直流 (HVDC) 的下一代 Kyber 机架设计。
原文参考
Key Updates on NVIDIA Vera Rubin/VR200 NVL72
(Integrating my latest supply chain checks and NVIDIA CEO Jensen Huang’s CES 2026 keynote)
1. NVIDIA has renamed the AI server VR200 NVL144 (die-based) to VR200 NVL72 (package-based). My supply chain checks indicate VR200 NVL72 will be offered in two power profiles: Max Q and Max P.
➢ Jensen referenced NVL72, rather than the previously used NVL144 naming.
➢ Max Q and Max P share the same hardware design.
➢ Max Q GPU/rack power (TGP/TDP): ~1.8/190 (kW).
➢ Max P GPU/rack power (TGP/TDP): ~2.3/230 (kW).
➢ Both are meaningfully higher than GB300 NVL72 at 1.4/140 (kW).
➢ Offering two power profiles improves deployment flexibility under data-center power constraints. It also suggests NVIDIA is increasingly factoring real-world physical infrastructure limits into product specifications.
2. VR200 NVL72 upgrades the GPU thermal solution for both MaxQ and MaxP.
➢ Both use a micro-channel cold plate (MCCP) paired with a gold-plated lid.
➢ The market has been expecting a move to a micro-channel lid (MCL) once GPU TGP reaches 2.3 kW; however, MCL mass production is unlikely before 2H27.
3. Benefiting from upgrades across the GPU, HBM, and NVLink Switch, VR200 NVL72 delivers roughly 3.5×/5× the training/inference AI computing power of GB300 NVL72, driving a sharp increase in rack-level power demand.
➢ VR200 NVL72 upgrades power shelves to 3*3U 110 kW (6*18.3 kW PSUs), vs. the most common GB300 NVL72 configuration of 8*1U 33 kW shelves (6*5.5 kW PSUs).
➢ VR200 NVL72 power shelves adopt an 3+1 redundancy design.
➢ VR200 NVL72 raises the power whip rating to 100A (vs. 60A for GB300 NVL72), tightening data-center power infrastructure requirements (e.g., busway and tap-off boxes).
4. VR200 NVL72 relies even more heavily on liquid cooling.
➢ Both the compute and NVSwitch trays adopt a fanless design.
➢ Rack technology cooling system flow (TCS flow) nearly doubles (~+100%) vs. GB300 NVL72, benefiting spec and/or count upgrades for CDUs, manifolds, cold plates, and quick disconnects (QDs).
➢ Rack airflow requirements fall by roughly 80% (in CFM) vs. GB300 NVL72.
5. VR200 NVL72 compute tray adopts a midplane for the first time, enabling a truly cableless design.
➢ Key midplane specs include 44 layers (22+22), M9 CCL (EM896K3), and an approximate size of 420×60 mm.
➢ Jensen noted that this design can reduce compute-tray assembly time from roughly two hours to about five minutes (vs. GB300 NVL72).
6. Rubin CoWoS are estimated at 300–350k wafers in 2026, with pilot production expected in early 1Q26 and mass production by late 2Q26. VR200 NVL72 rack assembly is expected to enter mass production by the end of 3Q26; factoring in yield ramp, 2H26 rack shipments are estimated at ~5,000–7,000 units.
7. Once rack power approaches ~200 kW, 54V distribution begins to face material constraints in space (busbars/cabling) and conversion efficiency. As a result, VR200 NVL72 can be viewed as the final generation of the Oberon rack architecture. To address the rising rack-power requirements driven by continued AI compute scaling, NVIDIA is expected to transition to the next-generation Kyber rack design, which supports 800V HVDC.
标签: 液冷、数据中心等 点击: 评论: