2025-11-20 · 个股论点

因执行风险及融资压力降级$IREN，维持$NBIS跑赢大盘评级。

涉及标的： $CIFR $CRWV $IREN $NBIS $ORCL $WULF

中文翻译

我的措辞非常严谨。你在阅读关于“引导性签约容量”的讨论时，跳过了“预期”一词。因此，我再次重申上述观点，但提供更详细的解释： $IREN 最初的护城河是吉瓦(GW)级容量+垂直整合。$NBIS 的护城河是全栈(Full-stack)。然而，Nebius 在长期吉瓦级容量上实现了与原本被认为是 $IREN 最大护城河的平价。以下是有帮助的解释： - 短期：$IREN 0.66 GW 已连接 vs. $NBIS 22 GW 已连接。目前，$IREN 拥有更多前置容量。 - 中期：$NBIS 和 $IREN 之间的已连接电力差距正在缩小，$IREN 2026年为 1 GW，Nebius 为 1.4 GW。 - 长期：$IREN 规划 3 GW 电力扩张，$NBIS 规划 2.5 GW 签约电力扩张，Nebius 在已连接电力方面实现长期平价。 Nebius 管理层还将容量预期提高了 150%，并且有很大可能再次增加，以超越其他新云(Neocloud)公司。此外，$IREN 最终转向全栈 IAAS 而非风险较低的托管(Colo)($CIFR, $WULF)，以追求更高的收入和利润率潜力，正如他们的 ER 所述。但考虑到 $CRWV + $ORCL 在软件/编排方面的建设难度，与拥有现成且即插即用现有堆栈的 $NBIS 相比，$IREN 的新转型执行风险极高。如果 $IREN 能做好这一点，其上行潜力巨大，但由于执行风险以及为 AI 云转型的容量变现所需融资面临的信贷收紧，我已将其降级。如上所述，Nebius 拥有更大且多元化的客户群，这增强了收入可见性和防御性，因此我维持对 $NBIS 的跑赢大盘评级，因其具有非对称上行潜力。

英文原文

I'm very precise with my wording. You skipped over "expected" while reading discussions about guided contracted capacity. So reiterating what I've said above but with a longer explanation: $IREN's original moat was GW capacity + vertical integration. $NBIS's moat was full-stack. However, Nebius achieved long-term parity with GW capacity that was originally thought to be $IREN's biggest moat. Here's a helpful explanation: - Near term $IREN .66 GW connected vs. 22 GW connected with $NBIS. Today, $IREN has more frontloaded capacity. - Mid-term, The connected power argument is closing between $NBIS and $IREN with 1 GW 2026 for IREN and 1.4 GW for Nebius. - Long term, $IREN projects 3 GW power expansion, $NBIS projects 2.5 GW contracted power expansion with Nebius achieving longer term parity with connected power. Nebius management also raised capacity expectations by 150%, and there's a large chance it increases again to exceed other Neocloud companies. On top of that $IREN ended up pivoting to full-stack IAAS instead of lower-risk colo ( $CIFR, $WULF ) for higher revenue potential due to "higher revenue and margin potential" as per their ER. But looking at $CRWV + $ORCL buildout difficulty with software/orchestration, this immensely raises execution risk for a new pivot with $IREN compared to $NBIS, which has it down + plug and play with existing stack. $IREN has high upside potential if they get this right, but due to execution risk and credit tightening required for future financing for capacity monetization for their AI cloud pivot, I've downgraded it. Nebius also has a much larger diversified client base as mentioned above for both revenue visibility + defensibility, which is why I maintain outperform on $NBIS due to asymmetrical upside.

在 X 上查看原推 ↗