概念词典 · 产业链
Coreweave 模式
以 CoreWeave 为典型的"借钱买 GPU + 签长约出租"商业模式，本质是高杠杆赌 AI 算力需求持续，Serenity 反复警示其信贷紧缩风险。
别名： Coreweave 模式／ CoreWeave／ \$CRWV／ 新云杠杆模式／ GPU-as-a-Service
是什么
CoreWeave 模式 是 Serenity 反复使用的一个分析框架，描述一类商业模式：
- 签长约：和微软等大客户签 5 年 GPU 租赁合同
- 借大钱：以合同为抵押发行高息债（10%+ 利率）
- 买大量 GPU：用债务现金一次性大量采购
- 出租回收：合同期内租金抵债 + GPU 折旧
这种模式的财务杠杆通常是 5-10 倍。只要 AI 算力需求继续增长、GPU 残值不暴跌、利率不大幅上行，就能滚动赚钱。
为什么 Serenity 反复警示
他不全盘看空 CoreWeave 模式，但反复指出三个尾部风险：
1. 信贷紧缩风险
他举的典型案例是 $APLD（Applied Digital）：
“$APLD 发债认购不足，最终以 97 美分折价发了 23.5 亿美元高收益债，收益率约 10%。”
利率任何一次上行，整个杠杆模式利息支出激增。
2. 客户集中度
CoreWeave 多数收入来自单一客户（微软）。“如果微软 AI capex 计划修订，CoreWeave 利润表立刻塌方”。
3. GPU 残值塌方
H100 → B200 → B300 的迭代速度远超传统数据中心硬件。3 年前买的 H100 现在二手价不到 40%，但财务上按 5 年线性折旧。实际折旧远快于会计折旧，潜在亏损被滞后。
Serenity 的具体判断
| 公司 | 他的看法 |
|---|---|
| $CRWV CoreWeave | 警示信贷紧缩 + 客户集中度风险 |
| $NBIS Nebius | 偏好这家：“比 $CRWV 资产负债表更健康、客户更分散” |
| $IREN | 列入 “Fake it Until you make it” 警示名单 |
| $APLD | 警示已发生信贷紧缩信号 |
看 Coreweave 模式公司的几个细节
- 现金流量表 > 利润表：看真正回收的现金，不看会计利润
- 客户合同期限分布：3 年期 vs 5 年期对应不同折现压力
- GPU 利用率披露：粉饰过的数字往往用 “capacity booked” 代替 “capacity used”
- 建设阶段机构关注度：进入交付期前的建设阶段是最危险窗口
提到此概念的推文 (39)
按时间倒序 · 滚到底自动加载更多
-
AI产业链各环节财报强劲，需求与订单爆发式增长。
今天AI主题的财报简直疯狂，涉及$CRWV、$SMCI和$LITE。 Coreweave：荒谬的~1040亿美元+积压订单，这还不包括Q3初新增的250亿美元+客户协议。 Neoclouds（如Nebius、Iren等）的计算需求传导效应巨大。 Supermicro：荒谬的650-720亿美元营收指引，新增600亿美元+订单... AI数据中心建设的传导效应极其巨大。 Lumentum：营收从8.08亿 -> 10.1亿 -> 12.5亿美元攀升，运营利润率逐季增长。 光子学玩家一直在疯狂印钞。 简而言之：AI建设的每一层都在疯狂爆发，且体现在财报中。
英文原文
Insane earnings today for the AI theme with $CRWV, $SMCI, and $LITE. Coreweave: Absurd ~$104B+ backlog, which doesn't include $25B+ of new customer agreements added in early Q3. Compute demand read through for Neoclouds is enormous (For Nebius, Iren and others) Supermicro: Absurd $65-72B in revenue guidance, adding $60B+ in new orders... AI DC buildout read through is just enormous. Lumentum: $808M -> $1.01B -> $1.25B revenue ramp, with operating margins growing quarter after quarter. Photonics players just keeps printing. TLDR: Every layer of the AI buildout goes brrrrr, and it's showing up in earnings.
-
惊讶于 Allbirds 转向 GPU-as-a-Service 和 AI 云。
我已经不知道该说什么了，wtf 这是啥。 $BIRD 在以下消息后已经涨了 572%： “Allbirds 完成了一笔 5000 万美元可转换融资……用于转向 GPU-as-a-Service 和 AI 云基础设施。” 这是一个鞋品牌啊？ https://t.co/QViPJDXj1E
英文原文
I’m at a loss for words. wtf is this $BIRD is up 572% after: “Allbirds executed a $50M convertible financing facility… to fund a pivot into GPU-as-a-Service and AI cloud infrastructure” This is a shoe brand? https://t.co/QViPJDXj1E
-
对比三只Neocloud股票，认为$NBIS是最佳选择，$CRWV融资结构致命，$IREN长期增值但短期散户不利
我对$NBIS、$IREN、$CRWV以及当前Neocloud市场的看法。 其中一家将在5年内成为下一个AWS： 我的猜测是Nebius。 这不是赢家通吃（DigitalOcean与亚马逊并存），但显然有更优越的结构和可能的赢家。 不利因素： ->伊朗冲突导致降息的可能性很低。 ->更广泛的市场似乎不愿意资助CapEx周期。但又想从中获益。 关于$IREN： 我们理解，4.5GW = X收入。但谁来资助GPU？ 无论谁现在买入$6,000,000,000的ATM。 赢家将是稀释完成后进入的任何人。 现实是，如果没有colo模式，他们没有足够的资金来通过GPU货币化其容量。 而且他们没有找到其他融资方式，因此由于一个会买入他们出售的任何东西的邪教社区，他们选择了ATM。 然而，我同意长期来看这是增值的。只是对现在买入的散户没那么有利。 关于$CRWV： 他们做对了一切... $NVDA支持。超大规模客户... 但他们融资完全错了。现在，超过$1.5B+的年度债务利息正在吞噬Coreweave的利润并侵蚀FCF。 就像信用卡债务一样，Coreweave找到工作来偿还债务，但最终债务利息太高，工作收入不足以覆盖债务和扩张。 如果有任何公司倒闭，$CRWV是第一个因巨额债务和利息而倒下的。 关于$NBIS： 他们正在尽可能做好... $NVDA提供$2B资助capex。 可转债发行（可转债做空对冲对短期价格上涨不利）。 但这是比Coreweave更好的融资结构，利息低得多。 他们现在有来自$META和$MSFT的约$460亿+订单积压，这两家是最盈利的超大规模云服务商之一，且没有像Coreweave那样直接关联OpenAI的风险传染。 与其他公司不同的是；他们的其他公司也有增值（Clickhouse股权增值：avride机器人出租车规模化；toloka三位数增长）。 从我来看：Nebius是明显的赢家。 然而，当前宏观环境不利于整个板块的短期持有者，指数下跌7%。 尤其是如果他们在ATM活跃时买入的话。 长期来看，当他们规模化后获益时（例如$NBIS 2026年Q4，是的，甚至$IREN），收益将是巨大的。
英文原文
My thoughts on $NBIS, $IREN, $CRWV and the current Neocloud market. One of them ends up as the next AWS in 5 years: My guess it’s Nebius. It's not winner takes all (DigitalOcean is there with Amazon), but there's clearly superior structures and likely winners. The downside: -> Low chance of rate cuts from Iran conflict. ->Broader market doesn't appear to want to fund the CapEx cycle. But want to reap the benefits With $IREN: We get it, 4.5GW = X revenue. But who is funding the GPUs? Whoever is buying into the $6,000,000,000 ATM right now. The winners will be whoever enters after holders get fully diluted. The reality is, they don't have enough funding to monetize their capacity through GPUs without colo models. And they didn't find other financing methods, so they went through ATMs because of a cult community that will buy into anything they sell. However, I agree it will be accretive long term. Just not as much for the retail buying in now. With $CRWV: They did everything right... $NVDA backing. Hyperscaler clients... But they financed completely wrong. Now, $1.5B+ yearly debt interest is eating Coreweave alive and cuts into FCF. Almost like credit card debt, Coreweave gets a job to pay off that debt, but eventually, the debt interest is too high that working doesn't really cover that and expansion too. If any company goes down, $CRWV is the first to go the massive debt load and interest. With $NBIS: They're doing as much as they can right... $NVDA funding $2B to fund capex. Convertible note offerings (convertible note short hedging is annoying for short term price appreciation). But this is the best way to do financing structures with much lower interest than Coreweave. They now have ~$46B+ in backlog from $META and $MSFT, two of the most profitable hyperscalers out there, without direct OpenAI linked contagion like Coreweave. And unlike others; there’s appreciation from their other companies (Clickhouse equity appreciation: avride robotaxi scale up; toloka triple digit growth) From my take: Nebius is the clear winner. However, current macro environments does not favor short term holders across the board with indexes dropping 7%. Especially so if they're buying into active ATMs. Long term, the benefits when they scale up eg. $NBIS Q4 2026 (yes, even $IREN), will be immense.
-
CRWV大跌验证看空逻辑，警示新云债务风险差异。
这条推文经得起时间考验，$CRWV 过去一个月下跌约 28%。 我觉得最好的空头其实是多头投资者或波段交易者。 如果你看到 Coreweave 或 $CRCL 在 200 美元时的危险信号，与其回避，不如反向操作。 话虽如此，$NBIS 等其他股票正随着板块被算法性打压。 我一次又一次地注意到，并非所有新云服务商(Neoclouds)都是一样的。 - 一家新云服务商可能有数十亿美元的年度债务利息，损害自由现金流(FCF) - 另一家则可能没有……
英文原文
This aged well, $CRWV down ~28% in the past month. I feel like the best bears would actually be long investors or swing traders. If you see red flags, with Coreweave or $CRCL at $200, can always play the other direction instead of avoiding it. That being said, $NBIS and others have been being brought down algorithmically with the sector I’ve noticed this again and again, not all Neoclouds are built the same. - One Neocloud might have billions in yearly debt interest hurting FCF - The other Neocloud might not…
-
Oracle因OpenAI项目面临巨大交易对手风险及盈利不确定性。
我关注 Neoclouds 已有一段时间，因此对 $ORCL 很熟悉。情况很微妙： 主要看点是资本支出(CapEx)从 350 亿美元增至 500 亿美元，以及约 2000 多亿美元的数据中心(DC)租赁。 这主要是为了“星门(Stargate)”项目，主要客户是 OpenAI，这带来了巨大的交易对手风险(counterparty risk)。 众所周知，OpenAI 拥有超过 1 万亿美元签约资本支出，但没有足够的资金来支付。因此，你在财报中看到的用于投资回报(ROI)支出的 3000 多亿美元签约收入，可能会轻易消失。 你可能认为寻找其他客户很容易，但任何利用率滞后(utilization lag)，加上资产/租赁折旧，都会造成巨大伤害。 第二部分是： -> 将其转化为高利润率业务极其困难。 你必须结合 GPU 折旧、GPU 编排（这真的很难，目前只有 Nebius、Coreweave 等少数几家公司能做好），以及其他因素，才能将收入转化为自由现金流(FCF)。 当像 $CRWV 和 $ORCL 那样承担大量债务时，这会侵蚀任何自由现金流，且你仍需偿还债务。 《信息(The Information)》在一两个月前泄露的消息称，Oracle 在其 GPU 业务板块面临大量盈利问题（如 GPU 利用率问题等）。这造成了影响。 总体而言，主要不确定性在于与 OpenAI 的交易对手风险。其核心业务强劲，但新建设施 shaky（不稳定/令人担忧）。 因此我认为，这是针对不稳定交易对手的过度资本支出，且存在该支出能否转化为自由现金流或更多亏损的疑问。
英文原文
I've been covering Neoclouds for awhile so I'm familiar with $ORCL. It's nuanced: Main thing is CapEx -> $50B from $35B, and $200B+ or so in DC leases. This is mainly for the Stargate project, mainly for OpenAI, which presents incredible counter party risk. As you know OpenAI has $1T+ in contracted capex spend, but not enough money to pay for it. So that $300B+ in contracted revenue you see in earnings for ROI spend, could easily just disappear. You might think it's easy to just get another customer but any utilization lag, combined with asset/lease depreciation, hurts. A lot. Second part is -> It's incredibly hard to turn this into a high margin business. You have to combine GPU depreciation, GPU orchestration (which is really difficult Nebius, Coreweave are the few currently that can do this well), amid other things to turn that revenue into FCF. And when you take on a lot of debt as seen with $CRWV, and $ORCL that cuts into any FCF + you still need to pay it off. The Information leaked news a month or two ago that Oracle was facing a lot of issues with profitability on their GPU segment (GPU utilization issues, etc). So that took a toll. Overall, main uncertainty is counterparty risk with OpenAI. It's core business is strong, but new buildout is shaky. So I would argue it's excessive Capex for a shaky counterparty and questions on whether that turns into FCF or more losses.
-
作者分享CRWV交易经历，指出其巨额债务偿付压力。
@Veganhippo21 我去年在 $34 买入 $CRWV，今年在 $70 做了波段交易。话虽如此，“你酸吗”这种论调无助于解决 Coreweave 每年 $12.5 亿+ 债务偿付的问题。 https://t.co/DrP4LIKpYm
英文原文
@Veganhippo21 I was buying $CRWV at $34 last year, and I swing traded it at $70 this year. That being said, "u mad" doesn't help Coreweave's issue with $1.25B+ yearly debt payments. https://t.co/DrP4LIKpYm
-
看空CRWV因债务与折旧风险，建议多NBIS/CIFR空CRWV对冲。
$CRWV 可能是我能想到的唯一一家市值超 500 亿美元且很可能破产的公司。至少 $ORCL 拥有盈利的核心业务，并持有大量资产，例如 15% 的 TikTok 股份。 对我来说，做空 Coreweave 的理由很明确，纯粹因为其大部分收入将用于支付债务利息。再加上 GPU 折旧，情况更是雪上加霜。 但进行对冲是个好主意。我个人会做多 $NBIS、$CIFR，同时做空 $CRWV，这对长期投资来说似乎不错。Nvidia 目前仍在维持它们的生存。
英文原文
$CRWV is probably the only $50B+ company I can think of that is likely going to go bankrupt. At least $ORCL has a profitable core business and owns tons of assets like 15% of Tiktok. Coreweave is a clear short to me purely because a large part of their revenue will be used to pay off debt interest. Then you compound that with GPU depreciation. But it's a good idea to hedge. I would personally long $NBIS, $CIFR short $CRWV, seems decent for long term. Nvidia is keeping them alive for now.
-
看好NBIS和CIFR，指出Coreweave高息债务侵蚀现金流。
我认为 $NBIS 目前极具吸引力。$CIFR（指其托管业务而非新云业务）等特定个股也是如此。相比之下，$CRWV 的说服力则稍逊一筹。 Nebius 的商业模式本质上是将五家年同比增长(Y/Y)达100%的公司合并而成。（包括与 $UBER 合作的自动驾驶公司 Avride、估值约150亿美元的 Clickhouse、AI 标注公司 Toloka 以及教育科技公司）。此外，还包括其核心业务部门，该部门已实现约70亿美元的年度经常性收入(ARR)。 因此，尽管我们需要通过实际产生的自由现金流(FCF)来仔细审视这些乐观的收入数据，但公司方面表示长期 EBIT 利润率有望达到20%~30%。 最关键的差异在于，Coreweave 超过12亿美元的债务利息正在侵蚀其自由现金流，而其他公司则获得了2~3%的低利率融资。
英文原文
저는 $NBIS 가 현재 매우 매력적이라고 생각합니다. $CIFR(네오클라우드가 아닌 코로케이션 부문) 같은 특정 종목들도 마찬가지입니다. 반면 $CRWV 는 그만큼 설득력이 있지는 않네요. 네비우스(Nebius)의 비즈니스 모델은 기본적으로 전년 대비(Y/Y) 100%씩 성장하는 5개의 회사가 하나로 합쳐진 형태입니다. ($UBER와 협력하는 자율주행 Avride, 약 150억 달러의 가치로 평가받는 Clickhouse, Toloka(AI 라벨링), 그리고 에듀테크). 여기에 약 70억 달러의 ARR(연간 반복 매출)을 기록한 주력 사업 부문이 포함됩니다. 따라서 이러한 낙관적인 매출 수치들은 실제 창출되는 FCF(잉여현금흐름)를 통해 면밀히 살펴봐야 하겠지만, 사측에서는 장기적으로 20%~30%의 EBIT 마진을 제시하고 있습니다. 가장 중요한 차이점은 코어위브(Coreweave)의 경우 12억 달러 이상의 부채 이자가 FCF를 잠식하고 있는 반면, 다른 기업들은 2~3% 수준의 저금리를 확보했다는 점입니다.
-
对比新型云融资结构，指出Coreweave债务风险高且英伟达损失可控。
其他如 $CIFR、$WULF 是数据中心托管(colocation)商，它们不购买 GPU。$CRWV、$NBIS、$IREN 则购买。前两家由 $NVDA 注资。后者 $IREN 由 $MSFT 通过预付款购买 GPU。大多数新型云(neoclouds)拥有极佳的融资结构，如 Nebius 的低息可转换债券。Coreweave 是唯一陷入困境的，背负巨额债务，因此若无垃圾债(junk bonds)支持，很难获得额外资金购买更多 GPU。英伟达给 Coreweave 泼了冷水，投入 Coreweave 的 20 亿美元最终仍回流至 $NVDA。如果 Coreweave 因巨额债务和贬值的 GPU 而倒闭，英伟达损失并不大。
英文原文
Others like $CIFR, $WULF are colo. They don't buy GPUs. $CRWV, $NBIS, $IREN do. The first two $NVDA funded. The latter one, $MSFT did with prepayment for Iren to buy GPUs. Most neoclouds have great financing structures like low interest rate Nebius convertible notes. Coreweave is the only one on fire, with massive debt, so it's harder for them to secure additional funding to buy more GPUs without junk bonds. Nvidia threw a bucket of water on Coreweave and the $2B into Coreweave goes back to $NVDA anyways. Nvidia doesn't lose much if and when Coreweave goes under with just a ton of debt and depreciating GPUs.
-
建议押注英伟达而非高负债的Coreweave
考虑因为 $NVDA 的交易而买入 $CRWV 吗？ 你可以这样思考： 赌场给你更多资金，因此你可以利用抵押品承担更多债务去赌博。 然后所有额外的钱都流回了赌场（即 $NVDA）。 这并不意味着 Coreweave 是一个好的做多标的。 但这确实让英伟达成为了一个。 押注庄家，而不是那个背负 147 亿美元+债务且年利息支付超过 12 亿美元的人。
英文原文
Thinking about buying $CRWV on the $NVDA deal? Here's the way to think about it: A casino gives you more money, so you could take on more debt with collateral to gamble. Then all the extra money goes back to the Casino (aka. $NVDA). That doesn't make Coreweave a good long. But it does make Nvidia one. Bet on the house, not the one with $14.7B+ in debt with $1.2B+ in yearly interest payments.
-
发布1月25日美股评级，强烈看好AI、内存及美国供应链瓶颈股，回避高估值与稀释风险标的。
1月25日评级。欧盟关税及$INTC财报后更新。 强烈买入： $SNAP $META 三星电子 SK海力士 $MU 欣兴电子 $TSM $CRCL $AXTI $LPTH $COPX $LIT $AEHR $FORM $AMKR $AVGO $MRVL 买入： $COIN $SMCI $GOOGL $FIG $AMZN $IBIT $RDDT $TTD $HIMS $HOOD $COHR $AMBA $IREN $POET $AAOI $LASR $VPG $OSS $INTC $UMAC $ONDS $AIRO $DPRO $AVAV $BULL $ETOR VLN NBIS GLXY CIFR HUT WULF 存疑 $VELO $SKYT 回避 $UAVS $BKKT WLMT SLNH $PLTR CRWV $ORCL $BMNR $IONQ, RGTI, QBTS _ 强烈买入 Snapchat - 底部约在$7.4，我认为在此位置非常强劲。随着内存资本支出减少及内存变现进入2027年，自由现金流(FCF)增加。只需等待重估。 Meta - 营收同比增长26%极其强劲，上季度产生超$100亿自由现金流。预计下季度财报后走势将走强，此前因光学(环比EPS光学效应超700%)导致BBB抛售，现在应已消化。 三星电子 - 半导体领域的圣杯，三星同时提供高带宽内存(HBM)和代工(Fab)敞口。 SK海力士 - 内存超级周期 美光 - 内存超级周期，但有美国背景支持。 欣兴电子 - 针对HBM、IC载板、玻璃基板、CoWoS及其他所有瓶颈环节的“邪恶”长线持仓。 台积电 - 印钞机，字面意义上不会出错。 Circle - 预计降息2-3次可能会大幅损害Circle净利润，因此已被定价。但在$160亿市值时是极佳的长线标的，他们正在印钞，且应开始看到美元稳定币(USDC)的扩张。 AXTI - LPTH: 磷化铟(InP)/锗等瓶颈。将成为2026年的巨大主题。只需等待AXT的供应链中断或Lightpath的黑钻(Black Diamond)在美国制造。我认为由于产能爬坡->收入增加，下行风险极低，但类似HBM的“登月式”涨价可能存疑。 COPX - 锂：稀土/材料如铜、锂是2026年的极佳长线标的。与上述瓶颈类似，来自中国的供应链中断将导致资金流向确保供应+建设新供应链。 AEHR - 说实话，他们处于AI和机器人两个热门垂直领域。$550万索诺玛(Sonoma)订单可能与美光和碳化硅(SiC)测试有关。看起来是市值低于$10亿的极佳“登月”标的。 FORM - 由于涉及DRAM/HBM及代工/逻辑，可能在美国供应链中变得重要。良率在HBM4中尤为重要。 AMKR - 美国本土制造供应链及台积电->美国转移的极大受益者。 博通 - 财报后近期大幅回调。鉴于超大规模客户ASIC将继续爬坡（尽管有一些延迟），我认为在此位置强烈买入。 Marvell - 与博通故事相同，Marvell因微软Maia延迟传闻而抛售。只需等待2027年营收约翻倍，当市场开始定价这一点，以及在Celestial收购后，他们在互连等其他领域做得很好。 买入 Coinbase - 加密货币近期回调使Coinbase在$570亿市值下价值再次合理。我从未喜欢其交易所部分，但为贝莱德IBIT ETF提供基础设施+与Circle的USDC收入分成，赋予Coinbase相当不错的长期价值。 SMCI - 从$60+暴跌回$30+呈现了极具吸引力的机会。市场极度担忧毛利率->SMCI向海外扩张，特别是主权AI+购买低端Nvidia GPU。且SMCI在那里的毛利率应会提升。也可能因为与客户达成的交易变得“粘性”。他们的营收增长并未停滞，仍达$360亿+。 GOOGL - 此时Gemini可能会接管ChatGPT，所以我会继续做多谷歌。 Figma - 软件板块抛售为许多被重挫的标的如Figma提供了良好机会，其拥有极高的毛利率+稳健增长。 亚马逊 - 价格基本与去年持平，他们一直在增长，AWS表现良好，涉足机器人+太空低轨卫星(LEO)，看起来是未来极佳的长线标的。 比特币 - 始终是极具吸引力的长线标的 Reddit - 估值高，但毛利率极高且不会消失，因为每个人都在用Reddit。 TTD - 2025年的抛售再次呈现了极具吸引力的估值 HIMS - 说实话，在$29时对我极具吸引力，可能会再次放入强烈买入，但当然营收减速非常令人担忧。主要Alpha在于市场未定价Zava收购，仅凭庞大的客户群，他们就能从新客户中衍生大量营收。 Robinhood - 从$140抛售回$100再次为Robinhood提供了良好机会。他们不会消失，加上银行+其他新产品营收扩张，应带来积极顺风。 Coherent - 长线美国供应链，特别是光子学、InP等。 AMBA - 针对边缘AI推理用于机器人爬坡+边缘计算的“登月”长线标的。 POET - 现在基本是1/2现金，通过Celestial间接进入Marvell+超大规模客户。鉴于承销商在$7.25买入，$6.8的股价具有吸引力。 AAOI - 与微软Maia和AWS Trainium绑定的长线标的。两者都尚未真正起飞，所以只是等待游戏。 LASR - 定向能武器非常酷。我不太喜欢基本面如20%左右的营收增长，但技术实在太酷了。 VPG - 与Optimus爬坡绑定的长线标的。我们应在2026年底看到工业用例，2027年底看到消费用例，所以Optimus生产可能现在开始或Q2影响资产负债表。 OSS - 国防板块及边缘AI+$2亿合同的长线标的。 INTC - 做多美国政策，财报并未改变任何观点，只是短期价格。 UMAC - 在此水平上是美国无人机制造的极佳长线标的。 ONDS, Airo, DPRO - 与AIRO, DPRO相同，看多无人机板块。相比几周前美国入侵委内瑞拉并威胁格陵兰时，没有太多巨大的顺风，但主题上看多。 AVAV - 关于将研发类合同->长期合同的错误信息导致抛售，呈现了相当大的上行空间 BULL - 我喜欢像Robinhood, Webull等拥有大量零售用户的券商，因为一旦拥有客户群，就有无尽的变现方式。抛售回$8呈现了极具吸引力的上行空间 ETOR - 抛售过度，净利润同比高，基本50%现金，下行风险低。只需等待财报重估。他们表现也不错，AUM同比70%+，所以不明白为何这样定价。 VLN - 不再像以前那样接近1:1资产/净值，曾有一段时间他们有$1100万+投资(下跌63%)毛利率，$9300万现金，所以会更接近1.1-1.2亿 : $1.4亿市值，这说不通。话虽如此，仍有$8000万远期营收，毛利率从63%->69%，看起来重估机会相当大。市场似乎只是不喜欢与某国相关的公司如Etoro，我想 Nebius - $150亿Clickhouse估值仅显示了分部加总(SOTP)，我不惊讶他们的子公司如Avride最终会超越主营业务。话虽如此，由于$20亿+ ATM在公开市场出售，近期有卖压。随着他们在2026年底达到$70亿ARR目标，应会极速爬坡。 GLXY, CIFR, HUT, IREN, WULF - 继续做多colo及其他新云板块标的。话虽如此，大多数自2025年以来已上涨30-40%+，所以它们不再是强烈的买入，因为已被定价。但仍有很多上行空间。 存疑 VELO - 很多人问我对此的看法，因为FinX喜欢这只股票。他们有很酷的客户如SpaceX，但基本面看起来糟糕。 ~$1180万现金 + $1750万发行 vs. ~$2300万债务。他们几乎没有剩余跑道，现在买入的人可能会被稀释。 Velo是拥有像IQE(欧洲InP供应链)这样极佳客户群但基本面糟糕的完美例子。 SKYT - 它是美国本土制造供应链的极佳标的，用于量子组件或边缘等酷东西。受益于芯片法案，但营收增长非常缓慢。它比Velo是更好的投机性长线，因为基本面更好。 24%左右的低毛利率，非常低的运营利润率，显然已计入市值，但美国纯代工应是一个溢价的好故事。底线是增长不够快。 回避 UAVS - 无尽的稀释机器，超过100%的市值已给予可将对市值25%以下的股份100%+转换为零售出售的套利投资者 BKKT - $3亿ATM稀释，而市值为$5.5亿。无尽的稀释机器 沃尔玛 - 43倍市盈率，不可能。 SLNH - 前方有大量稀释。 Palantir - 担忧估值P/E Coreweave - 担忧巨额债务，$10亿+债务利息严重损害自由现金流。然后是OpenAI的分配/建设，如果考虑到Gemini正在接管OpenAI的市场份额，对其能否履行合同义务存在极度、极度的担忧。 Oracle - 可能有技术性反弹，但说真的，他们为OpenAI(如Stargate)花费了太多资本支出，像Coreweave一样，OpenAI在能否履行合同义务方面存在极度担忧 BMNR - 无尽的稀释机器为愚蠢的项目融资，如$2亿投入Mr. Beast的公司。预期长期ETH质押ETF，做空BMNR，溢价将消失，例如$2亿现金投入Mr. Beast的公司流动性很差。 IONQ, RGTI, QBTS - 量子估值非常拉伸。 _ 总体想法： 我个人保持极度做多，这只是个人想法，非投资建议(NFI)。 许多小盘股和投机性公司自1月1日以来已经重估，我不认为许多50-100%的涨幅会持续（周五我们看到这些标的中有很多获利了结）。 话虽如此，特朗普正试图进一步降息（再降息2-3次），特别是因为中期选举即将到来。 SPY上涨 = 当选几率更大。所以我会保持极度做多直到中期选举后。 话虽如此，这有助于成长、投机性公司等。但我们已经看到这在很大程度上已被定价，如我最喜欢的长线标的之一Rocketlab，季度营收$1.55亿却达到$450亿+市值，所以我开始质疑估值->将许多头寸转向更多价值型（如软件下跌或内存超级周期）。 主题上我极度看多 - AI, 内存, 半导体 - 瓶颈 - 关键材料等。 非常看多 - 美国本土制造供应链 看多 - 国防板块 并会寻找软件到社交媒体公司等的波段交易/复苏/重估机会，鉴于近期的抛售。
英文原文
Jan 25th Ratings. Post EU Tariffs and $INTC ER. Strong Buy: $SNAP $META Samsung Electronics SK Hynix $MU Unimicron $TSM $CRCL $AXTI $LPTH $COPX $LIT $AEHR $FORM $AMKR $AVGO $MRVL Buy: $COIN $SMCI $GOOGL $FIG $AMZN $IBIT $RDDT $TTD $HIMS $HOOD $COHR $AMBA $IREN $POET $AAOI $LASR $VPG $OSS $INTC $UMAC $ONDS $AIRO $DPRO $AVAV $BULL $ETOR VLN NBIS GLXY CIFR HUT WULF Questionable $VELO $SKYT Avoid $UAVS $BKKT WLMT SLNH $PLTR CRWV $ORCL $BMNR $IONQ , RGTI, QBTS _ Strong Buy Snapchat - Bottomed around $7.4, imo very strong at this level. Increased FCF from memory opex reduction and memory monetization into 2027. Just a waiting game for re-rating. Meta - 26% Y/Y revenue growth is extremely strong, produced $10B+ FCF last quarter. Expect it to pick up after next quarter earnings due to optics (700%+ Q/Q EPS optics) that caused selloff last time from BBB. Samsung Electronics - Holy grail for semis, samsung provides exposure to both hbm and foundry. SK Hynix - memory supercycle Micron - memory supercycle, but with US backing. Unimicron - unholy long for hbm, ic substrates, glass core, cowos, and all other bottlenecks. TSM - money printer, literally can't go wrong with this. Circle - 2-3x projected rate cuts would likely hurt circle net income a lot, hence why it's being priced in. But amazing long at $16B as they print money and should start seeing expansion of USDC. AXTI - LPTH: Bottlenecks for InP / Germanium, etc. Will be a huge theme going into 2026. It's just a waiting game for both supply chain disruption (in AXT) or made in America w/ black diamond in Lightpath. Low downside risk imo due to capacity ramp -> revenue increase, but moonshot HBM type price increases might be questionable. COPX - LI: Rare Earths/Materials like Copper, Lithium are great longs for 2026. Similar with bottlenecks above, supply chain disruptions from China will cause money to flow into securing supply + buildout out new supply chains. AEHR - Honestly, they sit in two different hot verticals in AI and Robotics. $5.5m Sonoma order might be linked with Micron and SiC Testing. Seems like an extremely good moonshot sub $1B MC. FORM - Likely to be important in US supply chains since they do DRAM/HBM, and Foundry/Logic. & Yield is especailyl important w/ hbm4. AMKR - extreme beneficiary of made in america us supply chains and tsm -> US AVGO - Large correction recently post earnings. Strong buy IMO at these levels given hyperscaler ASICs will continue to ramp (even though there's been some delays). MRVL - Same story with Broadcom, marvell selloff after rumors of Microsoft maia delays. It's just a waiting game for ~2x revenue in 2027 and when markets start pricing that in, and after celestial acqusition, they're doing great stuff in other segments like interconnects. Buy Coinbase - Recent correction to Crypto makes Coinbase value decent again at $57B. Was never a fan of their exchange portion, but providing infra for Blackrock IBIT etfs + USDC revenue sharing with Circle, gives Coinbase pretty good long term value. SMCI - Extreme selloff from the $60's+ back to $30's presents attractive opportunity here. Markets are extremely concerned about gross margins -> SMCI expanding overseas, especially with soverign AI + buying lower end nvda gpus. and SMCI's margins should increase over there. Also likely due to deals to become sticky w/ customers. It's not like they're dying revenue growth to $36B+. GOOGL - Gemini at this point would likely take over chatgpt, so i'd remain long google. Figma - Software selloff provides good opportunity into a lot of the hammered names like Figma which extremely high gross margins + sturdy growth Amazon - Basically same price as last year, they've been growing, AWS is doing fine, they're in robotics + space LEOs, and just seems like a great long going forward Bitcoin - Always an attractive Long Reddit - High valuations, but extremely high gross margins and not going anywhere since everyone uses reddit. TTD - Selloff from 2025 presents attractive valuations again HIMS - Honestly extremely attractive for me at $29, might be put into strong buy again, but of course revenue deceleration is very worrysome. Main alpha is that markets arent pricing in Zava acqusition and just from sheer customer base, they can derive a lot of revenue from new customers. Robinhood - Selloff from $140 back to $100 presents a good opportunity for Robinhood again. They're not going anywere, plus new product revenue expansion from banking + others, should present positive tailwinds. Coherent - Long US supply chains, esp. for photonics, inp, etc. AMBA - Moonshot long for edge AI inference for robotic ramps + edge compute. POET - Basically 1/2 cash now, backdoor into marvell + hyperscalers through celestial. Attractive upside at $6.8 given underwriters bought at $7.25 AAOI - long play tethered to msft maia and aws trainium. both of them haven't really taken off yet so it's just a waiting agme LASR - energy directed weapons are super cool. i dont quite like the fundamentals like low 20% revenue growth, but the technology is just way too cool. VPG - Long play tethered to optimus ramp. we should see industrial use cases EOY 2026 and consumer EOY 2027, so maybe optimus productions starts hitting balance sheet now or q2. OSS - Long play on defense sector and edge AI + $200m contract. INTC - long on us policy, earnings didn't really change any perspective, just short term price. UMAC - Great long play at these levels on drone manufacturing in US. ONDS, Airo, DPRO - Same with AIRO, DPRO, bullish on drone sector. There's not much of a massive tailwind compared to a few weeks ago when US was invading venezuela and threatening greenland, but thematically bullish. AVAV - selloff from misinformation about converting r&D type contracts -> long term contract presents considerable upside BULL - I do like brokerages like robinhood, webull, etc. that have a ton of retail users since there's endless ways to monetize once you own the customer base. selloff back to $8 presents attractive upsdie ETOR - selloff way overblown, high net income y/y, basically 50% cash, low downside risk. just waiting for re-rating per earnings. they're doing well too, 70%+ Y/Y AUM, so not sure why they're being priced in like this. VLN - not quite the same anymore as close to 1:1 assets/nav, at one point they had $11m+ inv (off 63%) gross margins, $93M cash, so would have been closer to 110-120m : $140m MC, which made no sense. That being said still $80m fwd revenue off 63% -> 69% gross margins, seems like considerable opportunity for re-rating. Markets just don't seem to like companies eg. Etoro related to a certain country, I guess Nebius - $15B clickhouse valuation just goes to show Sum of Parts, where I wouldn't be surpirsed if their subsidaries like Avride ended up overtaking the main business. That being said, near term selling pressure due to $2B+ ATM being sold on open market. Should ramp up extremely fast as they meet their $7B ARR target EOY 2026. GLXY, CIFR, HUT, IREN, WULF - Remain long on the colo, and other neocloud sector plays. That being said most are up 30-40%+ since 2025, so they're not exactly a strong buy anymore as they've been priced in. But lot of upside remains. Questionable VELO - Lot of people asked my opinion on this since FinX loves this stock. They have really cool customers like SpaceX, but fundamentals look terrible. ~$11.8M cash + $17.5M offering vs. ~$23M. debt. They barely have any runway left and people buying now are likely to be diluted. Velo is the perfect example of amazing customer base like IQE (EU for inP supply chain), but terrible fundamentals. SKYT - It's a great made in america supply chain company for a lot of cool stuff like quantum components or edge. Benefits from CHIP act, but very slow revenue growth. It's a lot better speculative long than Velo since it has better fundamenatls. Lower gross margins like 24%, very low operating margins, is obviously priced into MC but U.S. pure-play foundry should be a good story for premium. Bottom line are not really growing too fast though. Avoid UAVS - Endless dilution machine with over 100%+ of marketcap given over to arbitrage investors that can convert 100%+ of the shares under 25% market value to sell on retail BKKT - $300m ATM dilution right now while MC is $550m. Endless dilution machine Walmart - 43 p/e, there's no way. SLNH - Lot of dilution ahead. Palantir - Concern over valuation P/E Coreweave - Concerns over large debt, $1B+ in debt interest hurts FCF a ton. Then there's allocation/buildout for OpenAI, which has extreme, extreme concerns if they can fulfill contract obligations, especially since gemini is taking over market share of openai. Oracle - There might be technical rebound, but seriously, they've spent so much capex just for openai (eg. stargate), and like coreweave, OpenAI, which has extreme concerns over if they can fufill contracts obligations BMNR - endless dilution machine to fund silly projects like $200m into mr. beast's company. Expect long eth staking etfs, short bmnr plays, and premium to go under as $200m cash into mr. beast's company for example is not very liquid. IONQ, RGTI, QBTS - Quantum valuations are very stretched. _ Overall Thoughts: I'm personally staying extremely long, this is just personal thoughts NFI. A lot of small caps and speculative companies have already been re-rated since Jan 1st and I don't expect many of the 50-100% moves to continue (we've seen a lot of profit taking Friday on some of these names). That being said, Trump is trying to cut rates even more (another 2-3x projected), esp. since Midterms is coming up. SPY Up = better chance of getting elected. So I'm staying very long until after Midterms. That being said a lot of this helps growth, speculative companies etc. But we're already seeing this largely priced in like Rocketlab, one of my favorite longs, reaching $45B+ MC off $155m quarterly revenue, so I'm questioning valuations a bit -> pivoting a lot of positions into more value (eg. software drop or memory supercycle). Thematically I'm extremely bullish on - AI, Memory, Semis - Bottlenecks - Critical Materials, etc. Very bullish on - Made in America supply chains Bullish on - Defense Sector And would look for swing trades/recovery/re-rating for stuff like software to social media companies around now given the recent selloff.
-
分析AI算力股与云厂商的算法关联及市场定价不确定性。
$GOOGL / Gemini 可能正在公开为这一结果欢呼。像 $CIFR 这样的公司主要通过 Fluidstack 与 Google 紧密绑定。然而，正如我们在上一轮下跌中所见，$NBIS 到 $IREN 在算法上被关联到 Oracle 和 Coreweave，无论 OpenAI 的关系影响如何。不确定市场这次会如何对此进行定价。
英文原文
$GOOGL / Gemini is probably openly cheering for this outcome. Companies like $CIFR are largely tethered to Google through Fluidstack. However, as we've seen last drop, $NBIS to $IREN were algorithmically associated with Oracle and Coreweave regardless of OpenAi relational impact. Not sure how markets price this in this time.
-
马斯克起诉OpenAI或引发AI产业链连锁反应，市场已部分定价该风险。
突发：埃隆·马斯克起诉 OpenAI，索赔 1340 亿美元。 预测市场今日给出的马斯克胜诉概率为 59%。 这可能令人担忧。 如果 OpenAI 败诉： 市场可能会引发与 OpenAI 相关的 AI 生态系统连锁反应（Contagion）。 $AMD 或 $CRWV 以及像 $ORCL（Stargate 项目）到 $MSFT 这样的公司将是巨大的输家，因为它们严重依赖 OpenAI。 此外，还有涉及 Coreweave 的 $APLD 和 $CORZ 等公司的三阶效应（Third-order effects）。 $AMD 的合作伙伴如 $RIOT 也是如此。 例如： 1. 微软/甲骨文 -> 数十亿美元的投资减值，建设 + GPU 利用率滞后将造成数十亿美元损失 2. CoreWeave -> 积压订单收入面临巨大风险，利用率滞后 + 已因巨额债务利息而处于风险之中 3. APLD / CORZ -> 依赖 Coreweave 的收入，如果租户倒闭，其短期收入也会随之下降。 如果 OpenAI 被迫支付巨额赔偿金，回归严格的非营利地位，或因流动性危机破产（无法履行合同义务）： 这可能会触发许多与 OpenAI 相关的 AI 板块股票的连锁反应事件。 当然，这是最坏的情况。 我的预期是中间地带： -> 马斯克赢得审判（或在判决前迫使达成和解）。 -> OpenAI 被迫支付巨额罚款（例如 100 亿 -200 亿美元）并可能开源其旧模型 -> $MSFT 可能会介入支付罚款或重组债务，因为利害关系太大。 或者在马斯克 41% 的败诉概率下：无事发生，与 OpenAI 相关的公司将安然无恙。 然而，市场似乎已经通过甲骨文（$330 -> $195）和 Coreweave（$150 -> $101）的价格大幅调整，定价了 OpenAI 连锁反应风险带来的大量不确定性。 这起诉讼绝对是市场关注的最大事件之一。
英文原文
Breaking: Elon Musk has sued OpenAI for $134 Billion. Prediction markets have priced the odds of Elon winning at 59% as of today. This may be alarming. If OpenAI loses this: Markets may trigger a contagion across the AI ecosystem tied to OpenAI. $AMD or $CRWV and companies like $ORCL (Stargate) to $MSFT are massive losers as they're heavily levered to OpenAI. Then there's third order effects on companies as $APLD and $CORZ involved to Coreweave. Same goes with $AMD partners, such as $RIOT. For example: 1. Microsoft / Oracle -> Write down billions in investment, buildout + GPU utilization lag would cost billions 2. CoreWeave -> Massive risk to backlog revenue, utilization lag + already at risk due to massive debt interest 3. APLD / CORZ -> Depend on Coreweave for revenue, if the tenant goes down, so does their near term revenue. If OpenAI are forced to pay massive damages, return to a strict non-profit status, or declare bankruptcy due to liquidity crises (and cannot fulfill its contractual obligations): This might trigger a contagion event across many AI sector names tied to OpenAI. Of course this is the worst case scenario. My expectation is middle ground: -> Musk wins the trial (or forces a settlement just before the verdict). -> OpenAI is forced to pay a massive fine (e.g., $10B–$20B) and possibly open-source its older models -> $MSFT likely steps in to pay the fine or restructure the debt since there's too much at stake. Or in the 41% chance Musk loses: nothing happens, and companies tied to OpenAI will be fine. However it looks markets already priced in a lot of that uncertainty with OpenAI contagion risk with Oracle ($330 -> $195) and Coreweave ($150 -> $101). This lawsuit is definitely one of the biggest events markets are watching.
-
OpenAI 高估值融资改善 AI 数据中心与供应链风险偏好
最新消息：OpenAI 在完成 5000 亿美元估值、100 亿美元以上的 Amazon 轮融资后，正在以 7500 亿美元估值继续融资。 $CRWV：+15.85% $NBIS：+10.28% 有了这笔交易，像 CoreWeave 和 Oracle 这样的公司，在为 OpenAI 的资本开支需求建设容量时，对手方风险在结构上降低了，因为 OpenAI 的资产负债表更强了。 像 Nebius 这种算法上和行业龙头绑定的公司，也因此上涨。 从 Rocket Lab 到 Bitcoin，高 beta 资产普遍上涨。 最近因为日元套息交易解除，加上今天大量未平仓合约到期，市场出现了极端波动。 不过，AI 交易的基本面（尤其是 Micron 神级财报显示了极强内存需求）以及 Neocloud 的基本面，仍然比以往更好。
英文原文
Latest news: OpenAI is raising funds at a $750B valuation after their $10B+ Amazon round at $500B. $CRWV: +15.85% $NBIS: +10.28% With this deal, companies like Coreweave and Oracle structurally have less counterparty risk with OpenAI's stronger balance sheet to fund capex requirements. Companies like Nebius that are algorithmically tied to the sector leaders are up as a result. High-Beta Assets from Rocketlab to Bitcoin are up across the board. There was extreme volatility recently with the Yen carry trade unwinding + large open interest expiring today. However, the fundamentals of the AI trade (especially with Micron's godlike ER showing extreme memory demand), and Neoclouds remain better than ever.
-
Amazon 投资 OpenAI 对 AI 数据中心和 ASIC 供应链的连锁影响
最新：Amazon 的 100 亿美元 OpenAI 融资，以及 AI 供应链的涟漪效应。 $AMZN 将以 5000 亿美元以上估值向 OpenAI 投资 100 亿美元以上。 为什么这是 AI 股票的重大结构性变化： 1. AI 数据中心交易去风险：（$ORCL、$CRWV、$APLD、$CORZ） 结合前面提到的 SPEED 法案，影响 Neocloud 的主要问题是： 1. 数据中心延迟和递延收入 2. 不可持续的 CapEx -> 没有 FCF 3. OpenAI 传染/积压订单风险 SPEED 法案直接处理了第 1 和第 2 点，但没有处理 OpenAI 相关的第 3 点。 影响 Oracle、CoreWeave 等最大 Neocloud/数据中心提供商的主要担忧，是它们为 OpenAI 这个对手方投入了巨额 capex，而 OpenAI 本身没有足够资金承诺这些资本开支。 但现在，随着 Amazon 新融资，第 3 点开始被解决。 - 新增 100 亿美元资金，加上 Amazon 的资产负债表支持 ChatGPT 的创造者，OpenAI 对 Oracle 和 CoreWeave 的早期承诺开始获得兜底。 - 下游影响：这直接降低了 $CRWV 和 $ORCL 的风险，因为它们在为 OpenAI 建设容量。再往下两层，依赖 CoreWeave 作为租户的 $APLD 或 $CORZ 也会被去风险。 由于 CoreWeave 和 Oracle 被视为“行业龙头”，这会立刻改变整个 Neocloud 板块的情绪，包括 $NBIS、$IREN、$CIFR、$WULF 等。 2. Hyperscaler AI capex 浪潮（$AMZN、$MRVL、Alchip）： 我们之前看到，$AVGO ASIC backlog 之后，市场开始担心 hyperscaler 支出减弱，很多相关玩家因此大跌。 但这笔交易的关键条件之一，是 OpenAI 采用 Amazon 自研 Trainium ASIC。这代表非 Nvidia 集群会激进扩张。 - 设计与 IP：直接利好 Marvell（$MRVL）和 Alchip 等 ASIC 设计伙伴。 - 定制硅集群需要大量光互连和 HBM。这会为光子（$AAOI、$LITE、$COHR）和内存（$MU、SK 海力士）创造新的 capex 超级周期。 - 代工厂，比如 $TSM。 以及更多参与 hyperscaler ASIC 建设的相关公司。 唯一输家？Nvidia（$NVDA）。 Amazon 正在成功用它庞大的资产负债表，迫使领先 LLM 摆脱对 H100/Blackwell 的依赖，同时推动整条 AI 供应链。 核心结论是，AI 交易由世界上最富有的公司资助，比如 $AMZN；OpenAI 也正在证明，它可以通过出让股权来扩大资产负债表，以满足需求。 做多 AI 板块。
英文原文
Just In: Amazon's $10B OpenAI Funding and The AI Supply Chain Ripple Effect. $AMZN is set to invest $10B+ in OpenAI at a $500B+ valuation Why this is a MASSIVE structural shift for AI stocks: 1. De-Risking the AI DC trade: ( $ORCL, $CRWV, $APLD, $CORZ ) With the SPEED Bill mentioned earlier, the main issues affecting Neoclouds were: 1. DC Delays & Deferred Revenue 2. Unsustainable CapEx → No FCF 3. OpenAI Contagion/Backlog. the Speed bill directly addresses #1 and #2. But not #3 with OpenAI. The main fears affecting the biggest Neocloud/Datanceter providers like Oracle, Coreweave was their immense capex spend for a counterparty (OpenAI) that doesn't have the funding to commit to it's capex spend. But now, #3 is starting to be addressed with the new Amazon funding. - With a fresh $10 Billion and Amazon’s balance sheet backing the creator of ChatGPT, OpenAI's early commitments to Oracle and CoreWeave are now starting to be backstopped. - Downstream Impact: This directly derisks companies like $CRWV and $ORCL, who are building capacity for OpenAI. And by two hops, companies like $APLD or $CORZ that rely on Coreweave as a tenant. As Coreweave and Oracle are seen as the "sector leaders" this immediately changes sentiment across the whole Neocloud sector from $NBIS, $IREN, $CIFR, $WULF and others as well. 2. The HyperScaler AI capex wave ( $AMZN, $MRVL, Alchip): We've seen fears after $AVGO ASIC backlog about hyperscaler spending waning. And many related players tanked on the news. However, a key condition of this deal is OpenAI’s adoption of Amazon’s proprietary Trainium ASICs. This signals an aggressive scaling of non-Nvidia clusters. - Design & IP: Direct benefit to ASIC design partners like Marvell ( $MRVL ) and Alchip. - Custom silicon clusters require massive optical interconnects and HBM. This creates a new capex supercycle for photonics ( $AAOI, $LITE, $COHR ) and memory ( $MU, SK Hynix). - Foundries such as $TSM. and many more related companies involved in the buildout of hyperscaler ASICs. The only loser? Nvidia ( $NVDA ). Amazon is successfully using its massive balance sheet to force the leading LLM to diversify away from H100/Blackwell dependence and boosting the whole AI supply chain alongside it. The main takeaway is that the AI trade is funded by the richest companies in the world, such as $AMZN, and OpenAI is showing it can scale up its balance sheet to meet requirements by trading off equity. Go long on the AI sector.
-
解释 SPEED Act 为什么比三次降息更利好数据中心链条
谢谢！之前也有一些帖子提到 SPEED Act 通过，以及“利好 $NBIS 和 $CIFR”，这很有帮助。 但我只是想拆解“为什么”，因为数据中心板块极其复杂，而且互相连接很深。 SPEED Act 通过比三次降息还好，因为在看利用率拖累、时间敏感的折旧成本，以及利润率如何影响 $ORCL 和 $CRWV 这类公司时，延迟可能是盈利能力面临的最大单一问题之一。 对于 CoreWeave 这样的直接受益者来说，这项法案会降低关于延迟和“不可持续 capex”的空头 thesis 风险。$APLD 和 $CORZ 这类下游公司，也会因为对手方风险降低而改善。
英文原文
Thanks! There were some other posts about the Speed act passing and “bullish for $NBIS and $CIFR”, which is helpful. But I just wanted to break down “why” since the DC sector is EXCEPTIONALLY nuanced and interconnected. Speed act passing is better than 3x rate cut, since delays are probably one of single biggest issue for probability whe we look at utilization drag, time sensitive depreciation costs, and margins affecting companies like $ORCL and $CRWV. Bear case thesis with delays and “unsustainable capex” gets derisked with this act on direct beneficiaries like Coreweave. And companies downstream from $APLD and $CORZ improve as well from lower counterparty risk.
-
长文拆解 SPEED Act 对 Neocloud 和 AI 数据中心的去风险作用
刚刚，SPEED Act 在众议院取得进展。 这是今年 Neocloud 板块（$NBIS、$CRWV、$IREN）最大的单一去风险法案/事件。 下面是原因和梳理： 美国政府正准备以美国对中国的国家安全为理由，支持从 Oracle 到 Nebius 的 AI 数据中心建设。 Oracle 和 CoreWeave 最近跌了 40%+（也把 $NBIS 从 140 美元打到 79 美元、$IREN 从 80 美元打到 35 美元、$CIFR 从 24 美元打到 14 美元），核心恐惧有三点： 1. 数据中心延迟和递延收入 2. 不可持续的 CapEx -> 没有 FCF 3. OpenAI 传染/积压订单风险 SPEED Act 和美国政府干预，修复了数据中心建设延迟的空头论点，也解决了利用率滞后带来的盈利问题（利润率）。 #1 数据中心延迟和递延收入 空头 thesis：多年许可延迟（NEPA、输电）把高价值合同变成了递延收入风险。$CRWV 明确把供应商延迟作为下调指引的原因，并在财报后大跌，因为相当一部分收入被推迟到 2026 年 Q1/Q2。 如果 SPEED Act 通过，alpha 在这里： 强制速度和诉讼保护。 - SPEED Act 要求联邦环境和监管审查遵守严格且不可协商的截止日期，通常是 1-2 年。 - 诉讼保护：该法案大幅缩短针对已批准许可提起诉讼的时效，比如缩到 150 天，并指示法院即使许可被临时挑战，也要允许数据中心建设继续推进。 结果：从签约到“GPU 上架 -> 收入流入”的时间线，被压缩了，并由联邦政府在政治上去风险。 递延收入被提前确认，修复了 $CRWV、$APLD 和 Neocloud 板块此前面临的延迟、递延利润/收入问题。 #2 不可持续 CapEx -> 无法从资产变现出 FCF 空头 thesis：公司在 GPU 和建设上花费数十亿美元（$ORCL 的 capex 很巨大），但从购买 GPU 到变现之间的利用率拖累严重影响盈利能力和 FCF。 这也大幅影响 AI Cloud 供应商，因为它们缺乏足够电力来把 GPU/capex 变现。 公司因此面临巨大的减记风险，也就是利用率拖累：GPU 闲置时，折旧和通电成本的时钟仍在跑。 这个拖累对数据中心部门盈利能力影响巨大，The Information 关于 $ORCL AI 利润率极薄的报道也提到过这一点。 SPEED Act 和美国政府干预会直接降低 CapEx 风险，因为速度上的立法要求（修复点 #1）实际上保证电力基础设施会在一个确定且较短的时间线内到位。 这种确定性让 $NBIS、$CRWV 和 $IREN 可以更有信心安排数十亿美元 GPU 的采购和部署，知道资产到货后就能立即开始变现，同时也通过降低利用率拖累来加快 FCF。 这种结构性变化会流向整个行业。它会立即降低主要 AI Cloud 供应商（$AMZN、$MSFT、$ORCL）的风险，因为它们现在能更确定地保证产能；同时也会保证 Colo/Infra/Energy 提供商（$CIFR、$WULF 等）的需求，因为它们的核心业务就是提供电力容量。 关于 capex -> FCF 以及建设延迟时间线的空头论点，现在已经被 SPEED Act 直接处理。 现在美国政府准备加速 $NBIS、$CRWV 和 $IREN 这类公司，因为 AI 数据中心已经被放到美国和中国 AI 国家安全竞争的前线。 它能否在众议院通过，是每个投资者都该关注的事。但如果通过，这会是 Neocloud / AI 数据中心建设里最大的、尚未被充分讨论的顺风之一。
英文原文
Just now, the SPEED Act ADVANCES in the House. This is the single biggest de-risking bill/event for the Neocloud sector ( $NBIS, $CRWV, $IREN) this year. Here's why and a rundown: The U.S. GOVERNMENT is set to support the AI data center buildout from Oracle down to Nebius on national security grounds for US vs China. Oracle and CoreWeave recently dropped 40%+ (tanking $NBIS $140 -> $79, $IREN $80 -> 35, $CIFR, $24 -> $14 as well) on three core fears: 1. DC Delays & Deferred Revenue 2. Unsustainable CapEx → No FCF 3. OpenAI Contagion/Backlog. The Speed Act and US Government intervention fixes bear-case points for data center buildout delays and addresses utilization lag profitability issues (margins). #1 DC Delays & Deferred Revenue Bear Thesis: Multi-year permitting delays (NEPA, transmission) turned high-value contracts into deferred revenue risk. $CRWV explicitly cited vendor delays for lowering guidance and tanked on earnings shifting a large portion of revenue to from Q1 Q2 2026. The alpha if the Speed Act passes: Mandatory Speed and Litigation Shields. - The Speed Act mandates strict, non-negotiable deadlines (often 1-2 years) for federal environmental and regulatory reviews. - The Litigation Shield: The bill drastically shortens the statute of limitations for filing lawsuits against approved permits (e.g., to 150 days) and instructs courts to allow DC buildout to continue even if a permit is temporarily challenged). The Result: The timeline from contract signing to "GPUs on racks -> revenue flowing" is now compressed and politically de-risked by the Federal Government. Deferred revenue is pulled forward and fixes delays and deferred profitability/revenue that plagued $CRWV, $APLD, and the Neocloud sector. #2: Unsustainable CapEx -> No FCF from monetizing the assets Bear Thesis: Companies were spending billions on GPUs and construction ( $ORCL's capex is massive) with utilization drag (from the point of purchasing the GPUs to monetization) largely affecting profitability and FCF. This also largely affects AI Cloud vendors (lacking power to turn monetize the GPUs/capex). Again This forced companies to take a massive write-down risk due to Utilization Drag (the time the GPU sits idle while the clock runs on depreciation/power-up). This drag is HUGE for profitability on DC segments, as cited in The Information reports on $ORCL's razor-thin AI margins. The SPEED Act and the US Government intervention directly de-risks CapEx as the legislative mandate for speed (Fix #1) effectively guarantees that power infrastructure will arrive within a defined, short timeline. This certainty allows $NBIS, $CRWV, and $IREN to time the purchase and deployment of billions in GPUs with high confidence that the assets will begin monetizing immediately upon arrival as well as accelerates FCF from reducing utilization drags. This structural change flows down the entire industry. It instantly de-risks the major AI Cloud vendors ($AMZN, $MSFT, $ORCL) who can now guarantee their capacity, and it guarantees demand for the Colo/Infra/Energy providers ( $CIFR, $WULF, and others) whose core business is supplying that power capacity. The bear case on capex -> FCF + buildout delay timeline has now directly addressed with the Speed ACT. Now the US government is set accelerate companies like $NBIS, $CRWV, and $IREN as AI Datacenters is now placed on the forefront of the AI national security battle between the United States and China. Whether it passes legislation in the House is what every investor should be watching, but if does, this is one of the largest (not-talked about) tailwinds for the Neocloud /AI Decenter buildout.
-
博通财报后AI板块抛售是误解造成的买入机会，新云板块中OpenAI依赖股除外
博通[$AVGO]业绩及其对AI板块的影响，如$LIITE和$NBIS： 博通的业绩"双重超预期"，营收$180.2亿(+28% Y/Y)，EPS $1.95，超出共识预期。 但AVGO下跌-11.64%，并拖累了整个AI板块。 这是买入机会吗？ 是的。 博通被视为超大规模云厂商ASIC代理增长的代表，因为亚马逊$AMZN Trainium、微软$MSFT Maia、尤其是谷歌$GOOGL TPU V7 Ironwood都通过它进行规模化部署。 而像$ALAB(-13.2%)、$CRDO(-5.11%)、$LITE(-12.23%)、$TSM(-3.71%)、$COHR(-9.25%)等公司都是TPU/ASIC建设以及博通作为公司的直接受益者。 博通下跌有三个原因，市场下跌有一个原因： 就博通而言，有一些小问题，如税率变化影响EPS模型，或因更多定制AI芯片而非更高利润率软件导致的"利润率压缩"，但这只是会计处理框架问题。（类似于$META在一次性税收后最初的下跌） 对博通和整体市场而言，是积压订单预期问题。以上所有引用的内容与ASIC市场预期增长相比都是小问题。 博通披露未来18个月$730亿的AI积压订单。而有关Anthropic和META购买价值数十亿$GOOGL TPU的传言，人们隐含预期是$800亿+。 然而，这次抛售是由算法和短期AI泡沫情绪驱动的价格错位，而非基本面破裂。 这条积压订单引用是确认订单的最低合同底线。谷歌$GOOGL、亚马逊$AMZN等公司可能会继续增加ASIC订单，而市场未能辨别这一细微差别。 分析师预期营收转化会更加前置，Q4之后积压订单应该会减少，这给出了2026年更高的可能范围$550-600亿+，而非$730亿预期中的$500亿。 TLDR：关于超大规模云厂商ASIC加速以与$NVDA依赖竞争这一论点没有改变。$AVGO和其他如$COHR、Sk Hynix、$MU、$VRT和$LITE都将受益。 这不是关于营收积压订单的最佳消息，但由于交付周期/订单周期和最低底线而被误解。 如果非要说什么，较低的超级云厂商ASIC需求对$NVDA及其生态系统是有益的，但我们也看到$CRWV、$SMCI、$NBIS和$NVDA GPU/DC计算生态系统今天都从盲目抛售中下跌超过5%，尽管存在负相关性。 这又是典型的"AI泡沫"周期因误解而再次来袭。AI股票普遍下跌10-12%的恐慌是一个绝佳的买入机会。 （该推文引用了@aleabitoreddit的推文，引用内容仅供理解语境）: 甲骨文[$ORCL]业绩及其对新云板块如$NBIS和$IREN的影响： 甲骨文报告EPS超预期且积压订单创纪录，但盘后下跌12%。 甲骨文较9月11日高点下跌39.8%，并拖累整个板块。 原因如下： 这次抛售不仅仅是对边际营收miss的反应，而是算法做空和投资者对AI资本支出周期可持续性以及该板块主要租户信用资质的不看好： OpenAI。 甲骨文宣布2026年资本支出增加$150亿至近$500亿，这与报道的与OpenAI $3000亿合作伙伴关系密不可分。 最初，OpenAI是前沿LLM，对甲骨文、Coreweave等公司有前景良好的资本支出承诺，促成了该板块的初始重新定价。 然而，随着超过$1万亿的义务以及Anthropic、Gemini、XAI等竞争对手的增加，市场严重质疑甲骨文、Coreweave等公司是否为无法从运营现金流履行其义务的租户在建设。 我们看到市场有效地发出信号：甲骨文正在为OpenAI创造一种不可持续的债务驱动型"供应商融资"，而OpenAI无法履行其承诺。 因此下跌是理性的：这次抛售是由信用风险和资本密集度的理性重新定价所驱动。 OpenAI融资担忧是合理的：OpenAI缺乏资金履行合同的假设得到其营收($130亿)与义务($600亿/年)之间明显不匹配的支撑。 信用担忧是真实的：甲骨文CDS利差扩大显示"信用事件"降级或违约的概率上升。 此外，我们看到这在新云板块引发传染效应，$NBIS从$140跌至$90s，$IREN从$80跌至$40s，$CIFR从$24s跌至$17s。 但这对$WULF、$NBIS、$IREN等新云公司来说是买入机会吗？ 是的。 这对$ORCL来说是好的买入机会吗？ 不是。 前瞻展望： $ORCL（很大一部分）、$CRWV（25%积压订单）是两个主要依赖OpenAI的公司，这一叙事可能因OpenAI的融资活动而瞬间翻转（+30%+变化）。 如果OpenAI在2026年以高估值超额认购IPO，且其新GPT模型击败Gemini/Claude，我们可以看到这种改变。 然而，许多其他公司与OpenAI无关。新云板块的原始论点是Mag7资本支出从其现金牛业务（Azure、AWS、GCP）向下流向：$NBIS、$IREN、$CIFR、$WULF等。 但随着最大玩家($ORCL、$CRWV)下跌，这些公司算法性地拖累了整个板块。 如果你看各家公司，$CIFR和$WULF由$GOOGL兜底，$IREN/$NBIS由$MSFT资助。 这些是与超大规模云厂商/Mag7的锁定合同积压订单，而非OpenAI。 这种因误解风险而导致的不理性抛售为新云板块提供了绝佳的买入机会，但不是与OpenAI相关的公司如$ORCL和$CRWV。
英文原文
Broadcom [ $AVGO ] earnings results and its effect on the AI sector like $LITE and $NBIS: Broadcom's ER was "double beat" with $18.02B revenue (+28% Y/Y) and $1.95 EPS, beating consensus. But AVGO dropped -11.64% and brought down the AI sector. Is this a buying opportunity? Yes. Broadcom is seen as a hyperscaler ASIC proxy growth as companies like $AMZN Trainium, $MSFT Maia, and most importantly $GOOGL TPU V7 Ironwood are scaled through it. And by proxy companies like $ALAB (-13.2%), $CRDO (-5.11%), $LITE (-12.23%), $TSM (-3.71%), $COHR (-9.25%), and are direct beneficiaries of the TPU/Asic buildout and Broadcom as a company. There's three reasons why Broadcom fell and one why the market fell: For Broadcom, there's minor things such as tax rate changing EPS models or "margin compression" from accounting from just more custom AI chips than higher-margin software, but this is just accounting framing. (Similar to how $META dropped initially on one-time tax post-ER) For both Broadcom general market, it was backlog expectations. Everything cited above is all minor compared to expected growth of ASIC markets. Broadcom cited $73B in AI backlog for the next 18 months. And rumors of Antrophic and META buying billions of $GOOGL TPUs, people were implicitly expecting $80B+. However, the selloff represents a dislocation in price driven by algorithms and short-term AI Bubble sentiment rather than a fundamental breakage. This backlog quote was the MINIMUM CONTRACTUAL FLOOR of confirmed orders. Companies like $GOOGL, $AMZN, will likely continue ramping up ASIC orders and the market failed to discern this nuance. Analysts are expecting revenue conversion to be more front loaded, and that there should be less backlog beyond Q4 given the cycles, which gives a higher likely range of $55-60B+ for 2026 rather than $50B expected of the $73B. TLDR: The thesis regarding hyperscaler ASIC ramp to compete vs $NVDA dependency has not changed. $AVGO and other players like $COHR, Sk Hynix, $MU, $VRT, and $LITE all stand to benefit. It's not the best news regarding the revenue backlog, but it's misunderstood due to lead-time/order cycles and minimum floors. If anything, lower hyperscaler ASIC demand is beneficial to $NVDA and their ecosystem, but we've also seen $CRWV, $SMCI, $NBIS and $NVDA GPU/DC compute ecosystem drop over 5%+ today from an indiscriminate sell-off despite inverse correlation. This is just the typical "AI Bubble" cycle hitting again from misunderstanding. The widespread panic of AI stocks dropping 10-12% is a great buying opportunity.
-
甲骨文因OpenAI信用风险大跌，新云板块错杀，建议买入非OpenAI依赖标的。
甲骨文($ORCL)的财报结果及其对$NBIS和$IREN等“新云(Neocloud)”板块的影响： 甲骨文财报EPS超预期且积压订单创纪录，但盘后仍下跌12%。 甲骨文自9月11日高点以来已下跌39.8%，并拖累了整个板块。 原因如下： 抛售不仅是对营收轻微不及预期的反应，更是算法做空和投资者对AI资本支出(Capex)周期可持续性，以及该板块主要租户OpenAI偿债能力的担忧。 甲骨文宣布2026年资本支出增加150亿美元至近500亿美元，这与据报道的与OpenAI的3000亿美元合作伙伴关系密不可分。 最初，OpenAI作为前沿大语言模型(LLM)，向甲骨文、Coreweave等承诺了诱人的资本支出，推动了板块的初步重估。 然而，随着超过1万亿美元的债务义务以及来自Anthropic、Gemini、XAI等的竞争加剧，市场严重怀疑甲骨文、Coreweave等是否为一家无法通过经营性现金流履行义务的租户建设基础设施。 市场有效发出了信号：甲骨文正在为无法履行承诺的OpenAI提供不可持续的债务融资“供应商融资(Vendor Financing)”。 因此，下跌是理性的：抛售由信用风险和资本密集度的理性重估驱动。 OpenAI资金担忧是合理的：OpenAI缺乏资金履行合同的假设，由其收入(130亿美元)与义务(600亿美元/年)之间的巨大错配所支持。 信用担忧是真实的：甲骨文信用违约互换(CDS)利差的扩大表明“信用事件”降级或违约的概率上升。 此外，我们看到这在“新云”板块中引发了传染效应：$NBIS从140美元跌至90美元区间，$IREN从80美元跌至40美元区间，$CIFR从20多美元跌至17美元区间。 但这对于$WULF、$NBIS、$IREN等新云公司是否是买入机会？ 是的。 这对$ORCL是否是好的买入机会？ 不是。 前瞻展望： $ORCL(大部分)和$CRWV(25%积压订单)是两大主要依赖OpenAI的玩家，这一叙事可能因OpenAI的融资活动而瞬间反转(+30%以上波动)。 如果OpenAI在2026年以高估值进行超额认购的IPO，且其新GPT模型击败Gemini/Claude，情况可能改变。 然而，许多其他玩家与OpenAI隔离。新云板块的原始逻辑是Mag7从其现金牛业务(Azure, AWS, GCP)向下漏斗资金至：$NBIS, $IREN, $CIFR, $WULF等。 但随着最大玩家($ORCL, $CRWV)下跌，算法交易拖累了整个板块。 如果单独看公司，$CIFR和$WULF由$GOOGL背书，$IREN/$NBIS由$MSFT资助。 这些是来自超大规模云厂商/Mag7的锁定合同积压，而非OpenAI。 这种因误解风险而导致的非理性抛售，为新云板块提供了绝佳的买入机会，但不包括与OpenAI绑定的$ORCL和$CRWV。
英文原文
Oracle [ $ORCL ] earning results and its effect on the neocloud sector like $NBIS & $IREN: Oracle reported earnings with a beat on EPS and a record backlog but, dropped 12% after hours. Oracle is down 39.8% from September 11th highs and brought down the sector with it. Here's why: The sell-off was not merely a reaction to marginal revenue miss, but both an algorithmic short and investor selloff on the sustainability of the AI capex cycle and the creditworthiness of the sector's primary tenant: OpenAI. Oracle's announcement of a $15 billion increase in 2026 capital spending to nearly $50 billion was inextricably linked to a reported $300 billion partnership with OpenAI. Originally, OpenAI was the frontier LLM, with promising capex promises to Oracle, Coreweave and others, contributing to the initial repricing of the sector. However, with over $1t+ in obligations and increasing competition from Anthropic, Gemini, XAI, and others, the markets have serious doubts on whether Oracle, Coreweave, and others are building for a tenant that cannot currently fund its obligations from operating cash flow. WE're seeing the market effectively signaling that the market Oracle is creating an unsustainable debt-funded "vendor financing" for OpenAI, which cannot fulfill its promises. So, the drop was rational: The sell-off was driven by a rational repricing of credit risk and capital intensity. OpenAI Funding Fear is Valid: The hypothesis that OpenAI lacks the funds to honor its contracts is supported by a glaring mismatch between its revenue ($13B) and its obligations ($60B/year). Credit Fears are Real: The widening of Oracle's CDS spreads sees a rising probability of a "credit event" downgrade or default. Furthermore, we're seeing this trigger a contagion effect across the "Neocloud" sector from $NBIS dropping from $140 to $90s, $IREN dropping $80 to $40s, $CIFR dropping from $24s to $17s. But is this a buying opportunity for Neoclouds like $WULF, $NBIS, $IREN, and others? Yes. Is this a good buying opportunity for $ORCL? No. Forward Outlook: $ORCL (large portion) , $CRWV (25% backlog) are the two players largely dependent on OpenAI and this narrative can flip in an instant (+30%+ change) depending on capital raising activity from OpenAI. If OpenAI files for an oversubscribed IPO in 2026 at high valuations and it's new GPT models beats out Gemini/Claude, we can see this change. However, many other players are isolated from OpenAI. The original thesis of the Neocloud sector was Mag7 capex funndel from their cash cows segments (Azure, AWS, GCP) down into: $NBIS, $IREN, $CIFR, $WULF, and others. But as the largest players ( $ORCL, $CRWV) fall, these algorithmically bring down the whole sector. If you look at the companies individually, companies like $CIFR and $WULF are being backstopped by $GOOGL, and $IREN / $NBIS are funded by $MSFT. These are locked in contract backlogs from Hyperscalers/Mag7, not OpenAI. This irrational selloff due to misunderstanding of risks presents an amazing buying opportunity for the Necoloud sector, but not companies tied to OpenAI like $ORCL and $CRWV.
-
反驳Hedgeye做空逻辑，对比Coreweave指出Nebius财务更健康且业务多元。
做空机构 @Hedgeye 正在做空 Nebius $NBIS。股价现为 $97.8。 他们的观点？Nebius 是 Coreweave 2.0。 Coreweave [ $CRWV ]： - 客户多元化程度低，主要是超大规模云服务商，约四分之一的积压订单来自 OpenAI。 - 剩余现金 18 亿美元。 - 由于 8-10% 利率的有毒融资，每年债务利息超过 13 亿美元。 Nebius 也是如此吗？ Nebius [ $NBIS ]： - 在 Neocloud 市场中拥有最高的客户多元化，包括 $META、$MSFT、政府、$SHOP、Mistral、$NOW 等。 - 剩余现金超过 48 亿美元。 - 票据结构约 40%+ 价外(OTM)，年利率约 2%，每年利息支出约 7660 万美元。 - 拥有对 $UBER 至关重要的 FSD 4 级 Robotaxi 自动驾驶汽车业务板块（最后一轮融资 $UBER 投入 3.75 亿美元），以对抗 Waymo。 - AI 训练板块由贝索斯资助，并被 $AMZN、Anthropic 等使用。 - 持有子公司 28% 的股份，该子公司被 $TSLA、$NET、Tiktok、$META 等大多数上市公司使用。 - 4 家子公司同比增长 100%，伴随核心运营业务增长，Tripleten/Toloka 每年增加净收入。 这是同一家公司吗。
英文原文
A short seller firm @Hedgeye is now shorting Nebius $NBIS. The stock price is now $97.8. Their claim? Nebius is Coreweave 2.0. Coreweave [ $CRWV ]: - Small diversification in customers, mainly hyperscalers, ~quarter of their backlog is OpenAI. - $1.8B in cash left - $1.3B+/yearly in debt interest from toxic financing at 8-10% interest. Same thing as Nebius right? Nebius [ $NBIS ]: - Highest diversification in the Neocloud market of customers from $META, $MSFT, governments, $SHOP, Mistral, $NOW, and many others. - $4.8B+ in cash left - ~$76.6M/yearly from note structure 40%+ OTM at ~2% interest. - FSD level 4 Robotaxi self driving car segment that is of critical importance to $UBER (last round $375M with $UBER) to compete vs Waymo. - AI training segments funded by Bezos and used by $AMZN, Anthropic, and others - 28% in subsidiary that most of the public companies from $TSLA, $NET, Tiktok, $META and so on use - 4 subsidiaries growing 100% Y/Y alongside their core operational business, with Tripleten/Toloka adding to net income every year. Same company.
-
反驳做空 $NBIS 观点，对比其与 $CRWV 在财务及业务上的巨大差异。
做空 $NBIS 是个好主意，因为你认为它和 $CRWV 一样，不错的论点！ $CRWV -> 每年 13 亿+美元的债务利息，拥有来自 OpenAI 的大量积压订单，剩余现金 18 亿美元。 $NBIS -> 可转换票据结构，利息不会严重侵蚀盈利能力；客户基础多元化以实现高利用率（Shopify、Cursor、政府机构、Meta/MSFT 等超大规模云服务商）；拥有 48 亿+美元现金用于建设；Robotaxi 自动驾驶汽车板块被 $UBER 使用并正在扩展以与 Waymo 竞争；AI 训练板块由 $AMZN、Anthropic 等资助并使用；持有 28% 的子公司股份，该子公司被大多数上市公司使用，所有 4 家子公司同比增长 100%——市值 240 亿美元。 当然和 Coreweave 是同一家公司。
英文原文
Great idea to short $NBIS because you think it's the same $CRWV, nice thesis! $CRWV -> $1.3B+ yearly debt interest, large backlog from OpenAI, $1.8B in cash left. $NBIS -> Convertible notes structure where interest doesn't heavily cut into profitability, diversified customer base for high utilization (shopify, cursor, governments, hyperscalers like meta/msft), $4.8B+ in cash for buildout, Robotaxi self driving car segments used by $UBER and scaling up to compete vs Waymo, AI training segments funded by and used by $AMZN, Anthropic, and others, 28% in subsidiary that most of the public companies use, all 4 subsidiaries growing 100% Y/Y - $24B MC. Same company as Coreweave for sure.
-
CoreWeave CEO称其自研软件栈是核心竞争优势。
对于 $CRWV，其CEO表示，正是其软件编排(software orchestration)能力，使他们能够将裸机(bare metal)转变为功能完备的超级计算机。CNBC上有一段5分钟的采访涉及此话题。 另一段采访： “CEO强调了CoreWeave的定制软件栈(custom-built software stack)，他认为这将在长期使公司脱颖而出。他表示：‘从我们的角度来看，我们真的相信，随着时间的推移，将我们与他人区分开来的……是驱动我们在基础设施上实现性能的……软件栈。’ Intrator补充说，CoreWeave的软件栈是从零开发的，以优化为指导原则。”
英文原文
For $CRWV its software orchestration according that allows them to turn bare metal into fully functioning supercomputers to their CEO. There’s a 5 min CNBC interview on this Another interview: “The CEO highlighted CoreWeave’s custom-built software stack, which he believes will set the company apart in the long run. He stated, “From our position, we really believe that over time, what’s going to differentiate us from anyone else is the … software stack that drives the performance that we’re able to achieve on our infrastructure.” Intrator added that CoreWeave’s software stack was developed from scratch, with optimization as the guiding principle. “
-
AI云股被算法归篮联动，CRWV财务堪忧，NBIS/IREN长期价值或超CRWV。
是的，完全同意。我认为算法/市场目前将 $NBIS、$IREN 归入 $CRWV、$NVDA 的篮子中。因此，任何关于 Coreweave 的负面报道都会对其他公司产生负面影响。我们终于看到 $WULF、$CIFR 等被归入数据中心(colocation)篮子（相对不受 GPU 贬值论点影响），并表现优异。但坦率地说，$CRWV 是一个财务噩梦，只要它被视为新云(neocloud)行业领导者，就会影响其他公司。至于 $NVDA，$GOOGL 的 TPU 论据是我目前看到的针对 $NVDA GPU 云业务最强的看空理由，但这些公司已经从 $META、$MSFT 锁定了 5 年的超大规模云服务商(hyperscaler)合约。我认为市场最终会正确定价，我相信 $NBIS 和 $IREN 有一天会比 $CRWV 更有价值，但其中只有一家拥有自动驾驶 Robotaxi lol。
英文原文
Yep absolutely. I think algos/market put $NBIS, $IREN in the $CRWV, $NVDA basket right now. So any negative hit piece about Coreweave does negatively affects the others. We've finally seen $WULF, $CIFR and others get put into the colo basket (which are relatively unaffected to GPU depreciation arguments), and outperform. But $CRWV is a financial nightmare to put it bluntly, so it does affect the others as long as it's treated as the neocloud sector leader. As for $NVDA, TPU arguments from $GOOGL is the strongest bear case I've seen to date though on $NVDA GPU clouds, but these companies already have have 5 year hyperscaler deals locked in from $META, $MSFT. imo markets will price things in correctly in due time, I do think both $NBIS and $IREN will be worth more than $CRWV one day. but only one of those has self-driving robotaxis lol
-
分析GLXY评级高于APLD的原因及Coreweave风险
是的，如果仅如你所说比较高性能计算(HPC)板块，我可能会将 $GLXY 降级至与 $APLD 相同的 C 级。但由于其其他业务板块，其资产负债表要稳健得多，因此需要动用较少的债务/信贷额度，这就是它评级更高的原因。他们还从事大量其他业务，如与 $HOOD 竞争的经纪业务、加密货币流动性提供商等，这改善了自由现金流(FCF)，但这样比较有点不公平。他们确实通过与 $CRWV 在 Helios 园区的合作获得了长期的超大规模客户可见性，当时的利率约为 9.10%。但这方面的 downside 是 Coreweave，且其破产的可能性不为零。
英文原文
Yeah I'd might move $GLXY down to a C tier alongside $APLD if we just had to compare HPC sector as you mentioned. But because of their other segments, their balance sheet is vastly better so they had to tap into less debt/crdit facilitaties, which is why it's a grade higher. They do a ton of other stuff like $HOOD competitor brokerage, crypto liquidity providers, etc. and it improves fcf but a bit of an unfair comparison. They do have long term hyperscaler visibility with $CRWV on their Helios campus with the the interest rate was ≈9.10%. Downside of that is Coreweave, and there's a nonzero chance it goes under.
-
Coreweave全栈虽强但现金流紧张且依赖OpenAI，风险巨大。
同意。但关于 @_melonBird 引用的 SemiAnalysis 对 $CRWV 的排名，我认同该表格。Coreweave 和 $NBIS 是整个 Neocloud 领域中唯二在“全栈”方面做得扎实的。Coreweave 花了多年时间和数十亿美元进行收购，以完善软件。然而，当你查看公司财务报表时，该表格毫无意义。Coreweave 仅剩 19 亿美元现金等价物，且每年面临 12 亿美元以上的债务利息。在信贷收紧期间，他们还需要为其他超大规模云厂商合同的建设筹集更多资金（而像 $NBIS 这样规模小得多的玩家却有 49 亿美元以上用于资本支出）。此外，其 1/3（220 亿美元以上）的积压订单来自 OpenAI，而 OpenAI 尚未获得履行该 1 万亿美元以上承诺的资金，$CRWV / $ORCL 正在为他们建设产能，这构成了巨大风险。
英文原文
Agreed. But to @_melonBird with the SemiAnalysis ranking on $CRWV, I agree with that table. Coreweave and $NBIS are the two with full-stack nailed down in the whole Neocloud sector. Coreweave spent many years and billions of USD on acquisitions to get software right. However, that table doesn't mean anything when you look at company financials. Coreweave only has $1.9B cash equivalent left and faces $1.2B+ a year from debt interest. And they'd need to raise more for buildout for their other hyperscaler contracts during a period of credit tightening (when much smaller players like $NBIS has $4.9B+ to spend on capex) On top, 1/3 ($22B+) of their backlog comes from OpenAi which has no funding for that $1T+ promised yet, and $CRWV / $ORCL is building out capacity for them, which presents a massive risk.
-
Coreweave因高债务和OpenAI依赖被评F级，NBIS更优。
如果你花30秒读一下帖子，就会发现我并没有给 $BITF 评级。 关于 $CRWV，短期有上涨空间，因为抛售至60美元区间确实过度了。但中长期的债务利息负担过重。 相比之下，$NBIS 和 $IREN 通过可转换债券融资或低利息，每季度对利润率的侵蚀很小；而 Coreweave 每年支付的债务利息超过12亿美元。 此外，OpenAI 占其总积压订单的三分之一。虽然他们拥有一些有防御性的大型超大规模客户（$META, $MSFT）以及 $NVDA 的背书，但像 $ORCL 那样为 OpenAI 建设基础设施，鉴于其缺乏收入以及 Gemini ($GOOGL) 的竞争，存在相当大的风险。 因此，尽管 Coreweave 是备受好评的全栈式服务供应商，但由于荒谬的债务利息侵蚀利润率，加上积压订单高度集中于 OpenAI（合同不确定性），其评级为 F 级，而这并不影响像 $NBIS 这样多元化的全栈式服务供应商。
英文原文
If you spent just 30 seconds reading the post, you would see I didn't rate $BITF. With $CRWV there's short term upside because sell-off was way too much to the $60s. But medium-long term they have way too much debt interest. While $NBIS and $IREN funds through convertibles or little interest cutting into margins every quarter, Coreweave pays over $1.2B+ in annual debt expense interest. OpenAI also makes up 1/3th of their total backlog, so while they have some defensible hyperscalers ( $META, $MSFT) + $NVDA backstop, building out for OpenAI like $ORCL, presents considerable risk given their lack of revenue + Gemini $GOOGL competition. So they are F tier despite being a highly rated full-stack offering due to absurd debt interest cutting into margins + large concentration of backlog from OpenAI (contract uncertainty) that don't affect diversified full-stack offerings like $NBIS.
-
IREN转全栈致风险激增，NBIS/CIFR基本面稳健却被板块性做空。
谢谢，是的，我之前做分级列表时将 $IREN 列为 S 级，但他们因试图通过购买数十亿 GPU 与 $CRWV 在全栈领域竞争而被大幅降级。许多 $IREN 持有者争论称托管(Colo)最盈利，但在 $IREN 的财报电话会上，他们明确表示因“更高的收入潜力和更强的回报”而购买 GPU 并从事全栈业务。构建软件层比人们想象的要难得多，$IREN 的执行风险急剧上升。与此同时，$NBIS 和 $CRWV 已经掌握了这一环节（根据其白皮书，$NBIS 胜过 $CRWV），但 Coreweave 在利息债务和 OpenAI 风波中搞砸了。不幸的是，我们看到整个板块被做空，尽管存在许多细微差别，例如 $CIFR 不受 GPU 贬值论点的影响。或者 $NBIS 拥有超过 5 年的 Mag7 及企业客户、强劲的资产负债表，且不受 $ORCL 及信贷收紧引发的 AI 泡沫 OpenAI 风波影响。
英文原文
Thanks, yeah I had $IREN as S tier when I did my tierlist earlier but they got downgraded a lot because they ended up trying to compete vs $CRWV in the full-stack by buying billions in GPUs. A lot of IREN holders were arguing colo was most profitable, but in $IREN's earnings call, they literally said they were buying GPUs and doing the full-stack due to "higher revenue potential and stronger returns". It's much, much harder building out the software layer than people think and the execution risk for IREN skyrocketed immensely. Meanwhile $NBIS and $CRWV already has that nailed down ( $NBIS wins vs. $CRWV per their whitepaper) but Coreweave messed up with interest debt + OpenAI drama. Unfortunately we're seeing a whole sector short despite many of the nuances eg. $CIFR not being affected by GPU depreciation arguments. Or $NBIS having Mag7 + enterprise customers over 5Y periods + strong balance sheet and not affected by AI bubble OpenAI fiasco with $ORCL + credit tightening.
-
AI股因信贷收紧恐慌抛售，建议转向有Mag7合同且低债的新云龙头。
我们正目睹 AI + 数据中心(DC) 股票实时崩盘，$IREN 本周下跌 37.88%，$NBIS 本月下跌 35.27%。 恐慌源于信贷收紧、高息债务、OpenAI 传染效应、宏观因素，以及最重要的：非理性恐惧。 以下是实时发生的情况： 1. 信贷收紧 - 像 $APLD 这样的公司未能成功发行债券（认购不足），最终发行了 23.5 亿美元的高收益债务融资——彭博社 11 月 14 日报道。债券以 97 美分的价格折价出售，收益率约为 10%。 该债券交易的风险主要与 Applied Digital 对 $CRWV 的依赖有关，后者占其合同收入的大部分。随着建设阶段机构关注度增加（例如 Burry 关于 GPU 折旧的做空报告），数据中心更难筹集资金。 融资条件收紧影响了整个通过发行债务为自身建设融资的生态系统。 2. 高息债务 - 由于信贷收紧，$CRWV 为扩张而承担的劣质债务导致每年超过 10 亿美元的损失。这侵蚀了利润率和自由现金流(FCF)。 同样，$APLD 和其他以不良收益率筹集资金的公司也面临同样的未来问题，我们看到相关公司因风险管理而出售。 3. OpenAI 承诺的 1 万亿美元+ 资金没有余额支撑 - OpenAI 承诺超过 1 万亿美元的资本支出(capex)，其中 224 亿美元+ 流向 $CRWV，3000 亿美元+ 流向 $ORCL，他们正在为此建设产能。由于 OpenAI 手头没有余额并采取避险策略，市场现在对这些承诺持怀疑态度。 现在，我们看到这种建设从 Coreweave 等公司向下传染到 $CORZ 和 $APLD，这些公司在建设期间可能依赖 $CRWV 的未来收入。 4. $CRWV 是 AI 基础设施（以及部分 $ORCL）的行业领导者。当行业领导者抛售时，通常其他公司会跟随。 5. 12 月降息概率下降： 除了高息债务外，我们看到 12 月降息概率降至约 53% 的抛硬币水平。较高的利率使得像 $CRWV 这样背负利息债务的公司再融资以增强 FCF 变得困难。 _ 细微差别与重新定位： 我们看到全面抛售。然而，Mag7（$META, $GOOGL, $AMZN, $MSFT）的资本支出和合同已确认，且盈利能力极强。 有一种简单的方法可以安全重新定位，以避免未来收入增长 + 债务周期的问题。 1. 配置与 Mag7 有合同可见性的新云(Neoclouds)： - $NBIS - 190 亿美元 $MSFT 交易，30 亿美元来自 $META，广泛的企业支持 - $CIFR - 55 亿美元与 $AMZN 的 Colo 交易，30 亿美元与 $GOOGL 的交易 - $WULF - 37 亿美元+ 与 $GOOGL 的交易等。 - $IREN - 90 亿美元与 $MSFT 的交易 我们可能会看到更多像 $SLNH, $CLSK 等投机性容量公司下跌，并整合到具有可见企业合同（未来收入下行风险低）的新云中。 2. 配置无/极低息债务且资产负债表强劲的新云： 像 $NBIS 这样的公司因拥有 47.9 亿美元的可转换票据现金及现金等价物，而与当前信贷市场波动隔离。或者像 $CIFR 这样拥有 12 亿美元现金的公司。 3. 避免与 OpenAI 有直接合同而非 Mag7 的公司。 _ 新云是 Mag7 AI 算力的骨干。这一论点没有改变，我们可能会看到一些规模达到 1000 亿美元+ 的公司成为下一个 AWS。 然而，我们看到由 $CRWV, $ORCL 和 OpenAI + 信贷收紧（部分由于 Burry 关于 GPU 折旧的错误机构做空报告）引起的广泛、不分青红皂白的抛售。 对于像 $NBIS 这样拥有 70-90 亿美元远期年度经常性收入(ARR) 和极强资产负债表的公司，这种抛售完全不合理。但对于其他投机性公司，这是一次必要的修正。 一句好话是“散户最后卖出，最后买回”。 如果你认为你可以卖出并在两周后回来，错过抛售后的一个或两个巨大反弹日可能会让你失去大部分恢复收益。 机构可能正在通过暗池/大宗交易购买更好的名字，例如 $NBIS，在这些清洗和保证金清算之后。我们还看到机构所有权从 38%-> 43-46%，尽管股价下跌。 利用这次机会购买好名字，并持有度过这些波动时期。 （该推文引用了 @aleabitoreddit 的推文，引用内容仅供理解语境）: 13F 新云申报的机构流向评估： · Nebius ( $NBIS ): 🟢 强烈正面 (7/10) · WULF ( $WULF ): 🟢 高度正面 (8.5/10) · IREN ( $IREN ): 🔴 非常负面 (3/10) · CIFR ( $CIFR ): 🟢 高度正面 (8.0/10) · Coreweave ( $CRWV ): 🟡 中性 (5.5/10) · Cleanspark ( $CLSK ): 🟢 高度正面 (9.0/10) _ TLDR 摘要更新： $NBIS · 机构所有权整体定量增长强劲，由坚实的长期机构买家以及量化基金和对冲基金的高活跃度驱动。 $WULF · 机构基础结构稳定且安全，以 Vanguard 和 BlackRock 等大型纯多头资产管理人为特征，他们对股票持有高信念。 $IREN · 所有权由高频交易员、做市商和量化基金（例如 Jane Street, Citadel）主导。这种结构被视为负面，因为它表明缺乏稳定的短期持有压力和长期机构缺乏信念。 $CIFR · 优秀的机构设置，以纳入主要指数（被动需求）和主动基金的强烈信念购买为标志，Alyeska 的大幅增持尤为突出。 $CRWV · 高交易量表明市场兴趣浓厚，但所有权目前高度集中，做市商存在显著，导致中性评分，暗示流向的风险/回报平衡。 $CLSK · 被认为是该组中最好的机构结构，结合了广泛支持的被动指数纳入和激进的、高信念的主动积累，表明定性风险最低。 _ 评论： 例如，“看多 - Jane Street + Citadel 拥有 x 的 $IREN ”（提示：这不是正面的）。尤其是当长期持有者如 FMR (Fidelity)：212 万股，大幅削减持仓时。 基金类型很重要，例如： 被动（最正面） Vanguard, BlackRock 指数, State Street, Geode 等 多头（正面） Fidelity, T. Rowe, Wellington, BIT 等 对冲基金（中性到正面） Alyeska, Coatue, Millennium 等 量化/做市商（负面） Jane Street, Citadel, Susquehanna, SIG, Two Sigma 等 这是一个主观框架，用于仅从本季度评分流向有多正面（而非整体所有权）。
英文原文
We're seeing a live crash in AI + DC stocks right now with $IREN down 37.88% this week to $NBIS down 35.27% this month. Panic stems from Credit Tightening, High Interest Debt, OpenAI contagion, Macro, and most importantly: Irrational Fear. Here's what happening real time: 1. Credit Tightening - Companies like $APLD failed to sell bonds (under subscribed) and ended up issuing $2.35B in high yield debt offerings - Bloomberg Nov 14th. The bonds were sold at 97 cents on the dollar, a significant discount, with a yield of approximately 10%. The bond deal's risk is primarily tied to Applied Digital's reliance on $CRWV, which accounts for a substantial portion of its contracted earnings. More institutional cation (eg. Burry short report on GPU depreciation) around the buildout makes it harder for DC's to raise funds. Tightening funding conditions affect the entire ecosystem that issue debt to fund their own buildout. 2. High Interest Debt - Due to credit tightening, bad debt, which $CRWV has to fund expansion, causes $1B+ in losses a year. This cuts into margins and FCF. Likewise, $APLD and others that raise funds with bad yields face the same future issues and we're seeing a selloff in related companies for risk-management. 3. $1T+ in OpenAI promised funding without balances - Over $1 trillion+ in capex was promised by OpenAI going to $22.4B+ $CRWV and $300B+ to $ORCL, which they are building capacity for. The market is now skeptical of these promises from OpenAI due to not having the balances on hand and taking a risk-off approach. And now we're seeing contagion on that buildout from companies like Coreweave spread down to $CORZ and $APLD during this buildout that might rely on $CRWV for future revenue. 4. $CRWV is a sector leaders of AI infrastructure (and partially $ORCL). When the sector leader sell-off, usually others follow suit. 5. Rate Cut odds in December Fall: On top of the high interest debt, we're seeing rate cut odds in December drop to a coinflip ~53%. Having higher interest rates make it trouble for interest-debt burdened companies like $CRWV to refinance for stronger FCF. _ THE NUANCE and Repositioning: We're seeing a broad sell-off across the board. However, Capex and contracts from Mag7 ( $META, $GOOGL, $AMZN, $MSFT ) are confirmed with their extreme profitability. There's an easy way to safely reposition to avoid issues with future revenue growth + debt cycles. 1. Position into Neoclouds with contract visibility with Mag7: - $NBIS - $19B $MSFT deal, $3B from $META, broad enterprise support - $CIFR - $5.5B Colo deal with $AMZN, $3B deal with $GOOGL - $WULF - $3.7B+ deals with $GOOGL and more. - $IREN - $9B deal with $MSFT We will likely see more capacity speculative companies like $SLNH, $CLSK, and others fall off and consolidation into Neoclouds with visible enterprise contract (low downside risk of future revenue). 2. Position into Neoclouds with no/extremely low interest debt and high balance sheets: Companies like $NBIS are isolated from current credit market volatility as they have $4.79 billion in cash and cash equivalents from convertible notes. Or companies like $CIFR with $1.2B in cash. 3. Avoid companies with direct contracts with OpenAI that's not Mag7. _ Neoclouds are the backbone of Mag7 AI compute. This thesis has not changed and we'll likely see some scale to $100B+ companies in the future as the next AWS. However, we're seeing a widespread, indiscriminate sell-off caused by $CRWV, $ORCL, and OpenAI + credit tightening (partially due to incorrect institutional short reports from Burry on GPU depreciation). These types of sell-offs are completely not warranted for companies like $NBIS with $7-9B forward ARR and extremely strong balance sheets. But for other speculative companies, it's a well-needed correction. A good quote is “retail is last to sell, last to buy back”. If you think you can sell and come back two weeks later, missing one or two huge rebound days after a selloff can cost you most of the recovery gains. Institutions are probably buying the better names right now eg. $NBIS through dark pools/block trades now after these types of washouts and margin liquidations. We've also seen institutional ownership go from 38%-> 43-46% despite share prices dropping. Use this as a buying opportunity for good names, and hold through these periods of volatility.
-
分析CRWV收购失败及高息债务劣势，对比NBIS优势并指出政府合同潜力。
对于 $CRWV 而言，问题在于其收购 $CORZ 未能获批。Core Scientific 本应比那 90 亿美元的收购价带来更大价值。若能在其现有合同基础上，通过收购为全栈模型（Full Stack Model）增加更多 GW 级算力，将是极度看涨（omega bullish）的利好。 Coreweave 每年还有超过 10 亿美元的利息债务，这对自由现金流（FCF）和利润率造成了极大拖累（相比之下，$NBIS、$CIFR、$IREN 使用的是可转换票据）。 $NBIS 已经实现了 $CRWV 的全栈垂直整合，但拥有更高的利用率、更好的编排/性能（根据其白皮书）、无巨额债务利息以及高积压订单，因此市场看到了其他风险更低、上行空间更大的替代方案。 不过，如果 $CRWV 获得美国政府合同批准，那将是另一个巨大的看涨催化剂。但我们已经看到像与 $META 签订的 140 亿美元这样的大型合同积压订单，然而其债务利息确实荒谬。
英文原文
For $CRWV it was $CORZ acquisition failing to get approved. Core scientific would have delivered so much more value than that 9B price tag. Having more GW capacity added to their full stack model (on top of their existing contract with them), would have been omega bullish with the acquisition. Coreweave also has $1B+ in interest debt a year, which is an extremely big drag on FCF + margins (compared to $NBIS, $CIFR, $IREN with did convertible notes). $NBIS already achieved what $CRWV did with full-stack vertical integration, but higher utilization + better orchestration/performance (as per their whitepaper) + no massive interest on debt + high backlogs, so markets see other lower risk, higher upside, alternatives emerging. If $CRWV get approved for US Gov contracts though, that would be another huge catalyst upside. But we already have visibility into large contracted backlogs like $14B with $META but the debt interest is just ridiculous.
-
补充 $CORZ 分析：Coreweave 长期合同改善现金流并资助资本支出。
你说得对，我漏掉了 $CORZ。感谢指出，也感谢 @cazenove_uk 等人的提醒。 鉴于 $CRWV 的合并前几天已终止，现在应将其视为独立实体看待，不知怎么竟疏忽了。我会在下一篇 2/2 帖子中补充。 $CORZ 与 $CRWV 签署的 12 年期 87 亿美元 数据中心托管(Colo) 合同带来了更可预测的现金流，且约 80% 的资本支出(Capex)（约 1.96 亿美元/第三季度 2.445 亿美元）由 Coreweave 资助，这种模式对股东远比可转换债券或侵蚀利润的有毒债务利息友好。 此外，拥有约 13-15 GW 容量，Coreweave 仅占用 500 MW，且收入能抵消资本支出，这相当有趣。
英文原文
You’re right, I missed $CORZ. Thanks for pointing it out, and appreciate others like @cazenove_uk for that too. Should just treat it as standalone now given $CRWV merger got terminated the other day, don’t know how it slipped my mind. I’ll add that in the next 2/2 post. $8.7 billion colo with $CRWV over 12 years gives $CORZ more predictable cashflows, and majority of capex ~80% ($~196m/ $244.5 million Q3) or so was funded by Coreweave, which is a lot nicer of a model for shareholders than convertible debt or toxic debt interest cutting into margins. Also, with ~1.3-1.5 GW capacity, Coreweave only taking up 500 MW, and revenue offsetting capex, it’s quite interesting.