Why NVIDIA Will Save The Market | Futurum Equities Ep. 31

[音乐] 各位女士们，各位先生们，欢迎收听《未来股权》播客第31期。有我，股票达人Shy，还有Daniel Newman。我们今天的播出时间比平时晚了一些，但…… >> 是我的错。是我的错。 >> 没人有错。我的意思是，现在是特朗普的记者会。所以，我想市场可能会有所变动。很高兴能和您一起直播，Daniel，因为市场正在经历一些波动，所以您可以看到我们的反应。但在此之前，您这周过得怎么样？我知道，在媒体方面，我们非常忙碌，《软件已死》。《软件已死》。我们该怎么办？ >> 是的，软件的终结。数据中心的终结。内存的终结。没有更多内存了。这是安全漏洞。我们没有足够的网络安全，但没人想要网络安全。每天都在变化。但是的，您在纽约的广播节目中表现出色。我远程参与。我也接了很多媒体采访。我们很忙。这是忙碌的一周。但是的，我注意到现在SoFi跌破20美元了。我不再是Danny Dimes了。我失去了一切。我……我…… >> 必须重新赢得那个称号。不，我的意思是，听着，我还在持有SoFi的股票。我的SoFi头寸大概亏损了20%多。 >> 嗯，我不知道是什么在驱动它，但我所知道的是 >> 显然，它深受关税、数据中心建设和比特币的影响。我认为SoFi在这个周期中确实将自己与金融科技公司区分开来，因为其他所有公布业绩的金融科技公司都显示出其商业模式的周期性风险，比如Robin Hood，他们显然仍然严重依赖加密货币交易。加密货币的寒冬，您看到了Pagaya Technologies的大幅失误，因为他们不想在资产负债表上承担那些信贷风险。还有很多其他公司，SoFi却能从中解脱出来，不受所有这些周期性的影响。除此之外，Nodo似乎正在经历一场转变，他的CEO风格发生了变化，他更加直言不讳地表达他对增长的看法。他还通过一些非主流媒体来传达这一点。这对股东来说是件好事。但无论如何，我认为他说过，最近的收益增长或指引将是今年剩余时间里指引的最低情况。SoFi公布的数字简直是疯狂的，但由于他们看到的下跌，你基本上得到了过去两次财报的免费赠送。但我不知道。您呢？ >> 还有稀释？是的。 >> 我认为他们分享了一些关于稀释的细节以及它是如何被利用的，是的。 >> 有效地说，他们正在利用这次资本筹集进行套利。所以我认为这只是一个耐心的问题。我仍然认为今年会被纳入标普500指数。顺便说一句，我不是故意完全陷入SoFi的讨论。我只是在笑，因为我不再是Danny Dime了。我长大了。我长大了。但是，当您投资组合中如此大的一部分大幅下跌时，您就会觉得自己不那么像一个人了。 >> 任何零钱。是的。说得对。 >> 任何零钱。但我们知道，我正在做各种疯狂的赌注，但到目前为止，我喜欢波动性。我喜欢波动性的期权交易。所以我并不总是……我并不总是看涨，因为我从来不确定事情什么时候会改变，但我对长期价值有信心，尤其是在好公司里，如果有什么东西跌得非常快，那就会产生巨大的期权波动性，而期权波动性会产生巨大的收入机会。 >> 嗯，我喜欢它。但我想有这两个阶段。一个是波动性阶段，事物横盘整理，市场有很多不同的玩法。另一个是所有东西都大幅上涨，您喜欢看您的资产负债表，您就喜欢…… [笑声] >> 您真的不想买，因为您觉得它太贵了。您觉得…… >> 我知道，我知道。 >> 所以您想增加您的头寸。但是的，很有趣，很多AI交易都横盘整理或下跌。软件行业受到了重创。加密货币仍然大幅下跌。您仍然看到很多轮动。您仍然看到Costco和Walmart的远期市盈率达到50倍。所以仍然存在这些市场中的异常情况。6500亿美元的资本支出，其中40-50%将用于Nvidia，但它几乎没有移动。 >> 价格。所以，什么？ >> 我认为很有趣的是，2026年出现了一个新主题。为什么Costco、Walmart、John Deere、可口可乐这些没有增长的公司，它们的市盈率估值却突然溢价了？我认为这是一个很好的讨论话题，因为我认为市场对所有这些AI基础设施支出以及AI对所有公司可能产生的颠覆性影响显示出了一些令人担忧的迹象，而今年表现优异的公司是那些拥有实体资产且完全排除在AI交易之外的公司。我认为它们提供了一种避风港。我认为这就是为什么苹果在MAG 7中表现如此出色的原因，因为它们显然没有参与AI的资本支出竞赛，这一点我和您都不同意，它们绝对应该参与，但市场正在奖励它们，因为它们没有过度支出的风险。我认为现在市场在两个极端力量之间非常不理性。一种力量是它们在没有证明的投资回报率的东西上花费巨资。荒谬。另一种极端是，它将具有颠覆性，影响巨大，没有人会购买这些产品，它将吞噬社会。这两种截然不同的极端同时发生。这完全说不通。要么是其中之一。不能两者兼有。所以我相信市场最终会在这两个极端之间找到一个平衡点。然而，我不知道什么时候会揭晓。基本上，这可能是一个漫长的时期，特别是当您有6500亿美元的资本支出时，正如您提到的，今年的资本支出增长与收益增长不匹配，可能会有一个漫长的消化期，直到它们拥有足够的计算能力，这样它们就不会那么受限，您就可以真正看到AI支出的投资回报率是否有效，这可能需要几个季度才能实现。我认为这是新的行为，因为市场大幅回调时，通常是V型反弹。它短暂而迅速。它是黑天鹅事件。这不是黑天鹅事件。这更像是，我刚吃了一顿丰盛的牛排晚餐，我需要一些胃药来缓解过度饮食引起的胃灼热。这需要一夜的睡眠。我不知道。您对导致AI基础设施公司动荡的原因有什么看法？嗯，我的意思是，您今天看到了新闻，比如……Cororeweave和Blue Owl的交易。是的。不是我的意思。 >> 诸如此类的事情出现，然后它们立即成为结构性问题。这是Corewave的问题，还是整个AI数据中心建设交易的问题？您知道，我们之前也看到过类似Oracle的情况。Oracle当时正在经历什么？Oracle现在正在经历的，是否会与Google、Microsoft和Amazon经历的一样？我认为随着时间的推移，他们一直在区分精华和糟粕，他们认为Oracle的风险比其他公司更大。虽然其他公司没有像Oracle那样跑得快，但它们也没有经历过同样的下跌。现在我认为Cororeweave周围存在一些脆弱性，围绕着许多这类AI GPU相关的公司。您看看Ciphers、Irons、Hut 8、Bitfarm，我不知道他们是否认为他们拥有与其他人相同的资本稳定性，能够筹集资金，尽管他们中的许多人最近成功地筹集了资金。但我想人们在想，这是否是结构性的，担忧是否更严重？我认为Cororeweave是独特的，因为我认为Cororeweave拥有非常独特的债务结构。他们不得不筹集大量资金，尽管增长迅速，但他们处理GPU折旧的方式，他们利用资产负债表上的资产进行借款的方式，我认为他们拥有独特的结构，不仅仅是财务上的结构，这使得他们比Nebuula或Cipher更难。所以我认为我们不能因噎废食。因为我仍然没有看到，我认为OpenAI的1000亿美元将正式宣布。它即将到来。我们看到了Anthropic的融资。私人市场完全没有回落估值。我们看到了Google的百年期债务融资。所有超大规模公司的债务融资都被超额认购。>> 我认为他们的风险是独特的。我不认为它……我认为人们想把它变成什么。我认为它不是什么。>> 他们不应该。我认为我们一直非常明确地指出Coreweave是……我可能比您更激进，但我认为它们是这个AI生态系统中可能出错的坏榜样。我认为这非常具体于Coreweave，今天的行动。我认为Blue Owl正在做出一个理性的决定，也许是稍微放慢速度，看看投资回报率在哪里，当您向一家初创公司投入如此多的资金时，因为归根结底，Cororeweave仍然是一家初创公司，它们还没有盈利，而且它们正在变得非常有创意，正如您提到的，在为这些交易融资方面。您只是不知道未来会是什么样子。所以我认为Cororeweave在这笔交易中的任何痛苦都不应被解读为AI需求的问题，因为正如您所说，如果您想要证据，说明真正的需求是什么，看看那些最大的参与者的超额认购。我认为我将Cororeweave列为AI生态系统中第三梯队的参与者。我不会……任何发生在第三梯队公司身上的事情，我都会忽略，就像忽略整个AI需求图景一样。我指的是，如果您想要一个大概的了解，显然超大规模公司是第一梯队。我也认为第二梯队是像Nebuula这样的公司，它们可以自我融资。我认为他们提到了60%，他们可以自我融资，这显然是第二梯队的公司。我认为您看到Cipher正在进行去风险化，他们的房东方法。我认为Iron拥有能够真正奏效的经济效益。我只是……我们需要看到这笔交易。但像其他拥挤的公司，它们正从比特币矿工转向AI数据中心，我把它们都列为第三梯队，当然，Coreweave的风险最大。但Coreweave未能满足的需求，我认为将由四大公司来满足。在我们看到需求崩溃之前，这一切都只是噪音。我不喜欢看到Nebuula因为这个消息下跌10%。我认为这是过度反应，但它确实创造了机会。我只是希望我有更多的现金。我每周增加1%的现金头寸。希望它能累积到5%或6%。但就目前而言，还记得我们上周谈到的那个水平吗，Daniel，21周EMA水平是一切。我们再次被钉在这个水平上。即使有所有这些波动性。我们现在在606美元。我们需要收盘价高于这个水平，才能在短期到中期内有所缓解。也许要到周三Nvidia的财报才会实现，但我不知道。我认为这是个有趣的时期。我确实想谈谈……实际上，您要说什么。>> 不，不，我当时……>> 不，我……所以，下周将是那种……希望您系好安全带的财报季。我不知道。您有机会看到下周谁在报告吗？我来谈谈。>> Nvidia很容易。是的，您说得对。但>> 在Nvidia的财报周里还有其他人吗？还有其他人会在Nvidia报告时一起报告吗？>> 太对了。>> 您看我怎么做的，对吧？每个季度我都会拿出Earnings Whispers的东西，然后把所有东西都划掉，只留下Nvidia，然后说下周的报告。[笑声] 我稍后会在推特上发布。这就像我的季度笑话，因为当他们报告时，其他所有东西都……我显然，如果您是任何其他公司的散户投资者，但整个市场都在等待Nvidia的表现。如果他们……如果他们失误了，其他任何人的表现都不重要了。>> 嗯，剧透一下，财报会非常强劲。我们看到了四大公司的资本支出增加。我认为评论比实际财报更重要。但对于所有散户投资者来说，我同意下周是Nvidia的成败时刻，但这是一个疯狂的时刻。周一，Him Health将在收盘后报告，它现在正成为一个战场股票，最近有所回落。然后，周二，我们有Marcato Libra。我们也有Cipher，周二早上。我们还有Digra Ocean，周二早上。这是我们一直以来谈论的最喜欢的名字，它正在 gaining popularity。所以，Digital Ocean实际产生我们一直在谈论的20%的增长加速，以及作为低成本、持久的云计算公司在生成式AI世界中的赢家。但我想软件将是关键。我确实有一些关于ServiceNow的评论，表明无论您报告什么，它可能都不会移动针，管理层做什么也不重要，这就像Bill Bill和他的团队一样，我确实想为Bill和他的团队所做的事情鼓掌。我认为这些CEO不断谈论AI不会颠覆软件，软件会没事的，软件会没事的，诸如此类的话太容易了。软件领域的内部人士一直没有买入，尽管情况很糟糕。>> 糟糕。但Bill买入了300万美元的开放营销。整个管理团队都说：“我们将停止执行您的……销售计划。”>> 他们取消了所有的销售计划，顺便说一句，这是值得注意的，他们不是积极销售。他们有计划销售。所以，很多高管都持有SBC，这意味着他们只获得X现金，而他们的大部分上涨空间都在股票中。所以他们经常出售股票作为补充。他们通常会提前几个月甚至更长时间来计划。而这正是他们如何逐步增加现金的方式。股票很好，但它不一定能支付你的账单。>> 事实上，大多数管理层都聚在一起说我们不想卖。首先，这说明了两件事。一是我们认为我们被低估了，这是一个重要的信号。这比股票回购更好。当公司用自己的钱回购股票时，这是一回事。当人们说：“我将推迟……或者……我当时想找什么词？”你知道，推迟现金，在这种情况下，对吧？>> 是的。>> ……以期获得长期收益。>> 嗯。>> 不，这很重要。我非常印象深刻。我失望的是市场忽略了它。我当时想……就是这样。>> 所以我认为这很棒。而且我认为……我跑题了。我只是想承认Bill对ServiceNow的管理有多么不可思议。一旦我们度过了AI将摧毁软件的叙事，我相信ServiceNow将在几年后成为被误读的股票之一。但也有趣的是，我不知道您是否有机会查看任何媒体。我提出了一个“遗留层测试”，我认为每个投资者都应该问自己，如果您是软件投资者。您必须问自己这三个问题。这三个问题基本上是AI会吞噬您，还是您将在AI经济中生存并蓬勃发展。第一个是最常见的。您是基于座位的模式，生活在一个AI代理将减少知识工作者的世界里吗？这是最明显的。每个人都在说这是颠覆。我同意。这将是一个颠覆，您将看到很多公司转向其他方向。第二个问题是，您的产品是纯数字的，还是它实际上触及了物理世界？我认为，如果它是纯数字的，您将面临AI成为您主要竞争对手的风险。这就是为什么我相信Palunteers这样的公司，编排层，他们的产品触及现实世界，他们在工作场所……他们影响战争，他们影响物流，他们影响现实世界的很多方面。这就是为什么他们从中受益匪浅。安全也是如此。设备上的安全。这是现实世界。Cloudflare，他们拥有那些巨大的塔楼，这有助于他们连接所有设备到边缘，这就是为什么他们控制着互联网的20%。所以这是一个例子。这是第二个问题。您的产品是纯数字的吗？第三个，这可能是最重要的。您是否拥有AI代理无法生成的数据？这是个大问题。所以，如果您的软件公司符合这三个问题中的每一个：它是基于座位的模式，它是纯数字的，而且我们不拥有数据，那么您就有可能成为新的代理AI堆栈的遗留层。哦。>> 如果您是一个……如果您只是其他数据的零拷贝，而您只是某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……某种……**

**规则：**

- 将所有英文单词翻译成中文。

- 保留原文的语气、含义和风格。

- 添加适当的标点符号。

- 使用双换行符将文本分成逻辑结构清晰的段落。

- **请勿**混合语言。

- **请勿**将所有内容合并成一个段落。

- 仅输出纯文本。

**待翻译文本：**

[音乐] Ladies and gentlemen, welcome to the 31st episode of the future equities podcast. We got myself, Stock Savvy Shy, we got Daniel Newman. We got a later than usual time slot today, but uh I don't know. >> My fault. My fault. >> Nobody's fault. I mean, it's Trump the Trump press conferences right now. So, I think maybe the market's going to be moving. Uh love being live with you, Daniel, as uh some volatility hits the market, so you can see our reaction to it. But before we get into it, how's your week been? I know uh it's been a pretty busy one for us in the media front with software is dead. Software is dead. What are we going to do? >> Yeah, end of software. It's the end of data centers. It's the end of memory. There's no more memory. Um it's the uh security meltdown. We don't have we don't have enough cyber, but nobody wants cyber. It's the you know, each day it changes. But yeah, you know, you were in New York crushing it on the airwaves. I was remoting in. Uh had had quite a bit of media as well. Uh we were busy. It was a busy week. Um, but yeah, you know, I noticed now that the that SoFi is down under 20. I'm no longer Danny Dimes. I'm I've lost the I've lost it. I'm I'm Danny. >> Got to earn that title back. No, I mean, listen, I'm bag holding on SoFi. I'm down like 20 something% I think on my Sohi position. >> Um I I don't know what's I don't know what's driving it, but all I know is >> obviously it's very influenced by tariffs and and data center buildouts and Bitcoin. I do think SoFi really separated themselves from the fintech group this cycle because every other fintech name that reported showed the cyclicality risk in their business models like um Robin Hood clearly they're still heavily exposed to crypto transactions. Crypto winter you saw Pagaya Technologies like show like a big miss because they don't want to really have those credit risks on their balance sheet. There's a ton of other names as well where SoFi was insulated from all that cyclicality. On top of that, Nodo feels like he's like turning a uh a new like character change in his style of being a CEO where he's being more vocal about what he thinks the growth is going to be. And he's also taking some more non-leacy media outlets to communicate this. So, that's a great thing for shareholders. But either way, I mean, I think he said something about this recent earnings growth is or the guide is going to be like the minimum case of what the guy's going to look like the rest of the year. Like I it's just crazy numbers uh that SoFi's reporting, but like you're essentially getting like the last two earnings reports for free because of the pullback they've seen. But I don't know. Have you >> and and the delution? Yeah, >> I think they they they shared some details on the dilution and how it's being utilized and yeah, >> you know, effectively they they they're arbitrageing with that with that capital raise. And so I think it's just a patience thing. I still think S&P inclusion comes this year. And by the way, I didn't mean to totally go down the SoFi rabbit hole. I was just laughing because I'm no longer Danny Dime. So I'm growing up. I'm growing up. But uh uh you know when a when a piece of your portfolio that large pulls back that much, you know, you're just a little bit less of a person at that point. >> Annie nickels. Yeah. There you go. >> Any penny any pennies today. But uh we know I'm I'm, you know, I'm I'm making all kinds of crazy bets, but uh you know, so far my uh you know, some of the plays I love volatility. I love options plays on volatility. So, I don't always I don't always go long into it because I'm never quite as certain as to when things are going to change, but I'm generally certain about long-term value, especially with good companies where like if something falls incredibly fast, that creates so much option volatility and option volatility creates so much income opportunity. >> Um, and so I do like it, you know, but I guess there's like these two stages. is there's the volatility stage where things are moving sideways and there's lots of kind of different ways you can play the market and then there's like the everything's up a lot and you just love looking at your balance and you just love [laughter] >> you don't really want to buy because you're like it's so expensive you're like >> I know I know >> you know so you want to help your position so but yeah I mean it's funny it's like a lot of sideways a lot of AI trades moved sideways or down uh software's gotten crushed uh crypto still still down a lot you're still seeing a lot of this rotation You still got a what is it 50 times forward on Costco and Walmart. So there's still like these these kind of anomalies in in the market and you know 650 billion in capex that's going to go 40 50% to Nvidia has largely not moved it at all um >> price. So what >> I think it's interesting um that is a new theme uh in 2026 the why is Costco Walmart's John John Deere uh Coca-Cola all these companies who have no growth they have a premium valuation on their earnings multiple all of a sudden and I think it's somewhat of um good conversation to have because I think there is somewhat of a heartburn that the market's showing you of all this AI infrastructure spend the disruption that AI is supposedly going to have for every name possible when the companies are outperforming this year are the ones that have a physical assets that are ex completely excluded from the AI trade and I think there's they're offering somewhat of a safe haven. I think that's why Apple's performing so well among the Mac 7 because they are obviously not participating in the capex uh race for AI which you and I both disagree with like they should absolutely be participating but they are being rewarded by the market because they don't have that overspending risk and I think right now the market is just being so irrational between two extreme forces. One force is they're spending an arm and a leg on something that there's no proven ROI yet. Ridiculous. The other uh extreme is it's going to be so disruptive and so impactful and nobody's gonna be buying these products and it's going to eat up society. Those are two massively different extremes that are happening at the same time. That doesn't make sense at all. It's either one or the other. It can't be both. So, I do believe that the market's eventually going to fall in between these two extremes. However, I don't know when that's going to be revealed. essentially it might be an extended period of time especially when you have uh $650 billion of capex like you mentioned like there's going to be no earnings growth to match that level of capex increase this year and there might be a prolonged period of digestion until they get enough compute possible so they're not as constrained so you can really see are the ROI working for these AI expenses and it might not happen for a couple quarters and I think that is new behavior because anytime there's been somewhat of a massive pullback in the market, it's been a Vshape. It's been short-lived. It's been black swanie. This is not black swanie. This is more I just had a full steak course dinner and I need to get some Pepto-Bismol relieve my heartburn from just overeating. That's going to take an overnight sleep essentially. I don't know. What's your take on the what's causing this like ruffle among AI infrastructure names? Well, I mean, you saw the news today like uh what was it? Uh Cororeweave and the Blue Owl deal. Yeah. Not I mean >> things like that popping up and these are then immediately they become structural questions. Is this a problem for Corewave or is this a problem for the whole you know AI data center buildout trade? You know, we saw this a little bit with Oracle for a period of time. Like what Oracle was going through? Is what Oracle going through going to be the same thing that Google and Microsoft and Amazon experienced and I think over time they were cutting the wheat from the chaff but they saw the risk to Oracle as much more specific than the risk to the others. And while the others didn't run, they also didn't get nearly the same sort of pullback. And now I think with Core Wave, there's a bit more of like a fragility around a lot of these sort of AIG GPU read names. You look at a ciphers, the irons, the the hut 8, the you know bit farm like people I don't know that they feel they have the same amount of of capital stability ability to raise funds despite the fact that they have su very successfully a lot of them have raised funds recently. Um, but I think people are wondering if it's if it's structural if the concerns are are more significant like I look at corweave as unique because I think corweave has a very unique uh debt structure. they've had to raise a lot of capital uh and despite growing really fast, the way they you know handle the depreciation of of their u GPUs, the way they utilize the assets on their balance sheet for borrowing purposes, like I think they have a unique structure, not just structural structure to their to their financials that makes them more difficult than say a Nebus or say a cipher. And so I do think it's a little bit of a we can't throw the baby out with the bath water here. Um, because I'm still not seeing, you know, I think OpenAI's 100 billion is going to get announced official. It's coming. It's imminent. Um, we saw the anthropic raise. The private market is not pulling back valuations at all. We've seen the hundred-year debt raise from Google. U the over subscription of all the debt raises from all the hyperscalers. Like >> I think their risk is unique. I don't think it I think people want to make it something. I don't think it's something. >> They shouldn't. And I think that you we've been very vocal about coreweave being or I've been maybe a little more aggressive than you, but I think they're the bad agents of this AI ecosystem on what can go wrong. And I think that's very specific to Corey like today's action. I think blue owl is making a rational decision maybe of just slowing down a bit uh and seeing where the ROI is going to come when you have when you're investing that much money with a startup because end of the day corre is still a startup startup they are not profitable and there are getting very creative like you mentioned on funding these uh these deals and you just don't know what's going to look like down the line so I think whatever pain on Cororeweave on this deal. It should not be read through on what the AI demand is because like you mentioned, if you want proof on what the real demand is, look at the overs subscription on the biggest players in the game. I think I put Corweave in the tier three category of players in this AI ecosystem. I wouldn't anything that happens to the tier three names, I would never I would just like ignore for the total AI demand picture. Uh, and what I mean by if you want like top of a head like obviously the hypers skills are tier one. Uh, I also think that tier 2 are the names like the nebuses who can self-und themselves. I think they mentioned 60% they could sell they can selfund themselves that that is clearly a tier 2 name. I think you're seeing cipher have the d-risisk the bid on the their landlordo approach. I think iron has the economics to make it really work. I just we need to see the deal. Um, but like the other names that are very crowded that are pivoting from Bitcoin miner to AI data centers, I put them all in tier three with the obviously the big risk in core. But whatever core doesn't fulfill, I think it's going to get fulfilled by the big four. And until we see the demand cracking, like this is just all noise. And I don't love to see Nebas down 10% off this news. I think it's an overreaction, but it does create opportunity. I just wish I had more cash. Like I'm adding 1% cash position every week. Uh hopefully it builds up to five or six. But as of right now, remember that level we talked about last week, Daniel, the 21week EMA level is everything. We are pinned to that level again. Even all this volatility. We are at 606 right now. It's like we need to close above this level for there to be somewhat of a relief of like the short to medium term is going to be okay. Maybe we don't get there until Nvidia's earnings on Wednesday, but I don't know. I think it's interesting times. I do want to talk about uh actually you're gonna say something. >> No, no, I was >> No, I So, next week is one of those like hopefully you have a seat belt on type of weeks in the earnings report earning cycle. I don't know. Did you get a chance to see who's reporting next week? I talk it over. >> Easy one for Nvidia. Yeah, you're right. But >> is there anyone else in an Nvidia week? Is anyone else reporting when Nvidia is reporting? >> It's so valid. >> You see how I do that, right? Every quarter I take the earnings whispers thing and I black everything out but the Nvidia and I say next week's report [laughter] like I'll tweet that out later. It's like it's like a quarterly joke for me cuz like when they report like everything else is like so you know and I obviously if you're in an individual investor in any other name but the entire market sits back and waits for what Nvidia does the if they if they falter it doesn't really matter what anyone else does. >> Well spoiler alert it's going to be a really strong earnings. We saw the big four capex increase. I think the commentary is going to be really big on the conference call more than actual earnings report in my opinion. But going over for the all the individual investors out there, I agree it's Nvidia or bus next week, but it's a crazy one. Uh Monday we got Him Health reporting after the bow that is becoming a battleground stock a bit right now in their recent pullback. Then we got let's see we got Marcato Libra on Tuesday. We got Cipher Tuesday morning as well. We got Digra Ocean t Tuesday morning as well. And that's been our favorite name that we've talked about for as long as this podcast has been out that is gaining popularity. So like this the bar has never been higher for Digital Ocean to actually produce what we've been talking about which is that 20% growth reaceleration and Pix and Shovel's winner of this lowcost durable cloud computing company in the Gentic AI world. Uh but I think software is going to be at play. I I think we got service now a bit. We got some commentary on service now and show that doesn't matter what what you report it might not move the needle and doesn't matter what the management does which is like Bill Bill and team like I I do want to applaud Bill and team for what they did. I think it's so easy for these CEOs to constantly talk about AI is not disrupting software, software is going to be fine. Software is going to be fine. Blah blah blah blah blah. There's been no buying from the insiders in the software space while being brutal. >> Brutal. But Bill, he bought $3 million open marketing. All the whole management team is like, "We're just going to stop with your uh sales plan." >> They canceled all their sales, which by the way, it's it's noteworthy like they're not selling actively. They have plan sales. So, a lot of the executives are on SBC, which means they only make X in cash and a lot of their upside is all in stock. So, they regularly sell stock as an augmentation. They usually plan these months in advance if not further. Um, and that's how they actually incrementally increase cash. You know, since stock is great to have, but it doesn't necessarily pay your bills. >> The fact that the most of the management team came together and said we don't want to sell. First of all, it says two things. One is we believe we're undervalued, which is an important signal. It's a better one even than buybacks. It's like one thing when the company buys back with its own money. It's another thing when people are saying, "I'm going to defer gratitude or or um defer um uh what the heck's the word I'm looking for?" You know, defer cash in this case, right? >> Yeah. >> Um in order to maybe get something longterm. >> Um >> no, it's it's everything. I think I was very impressed. I was disappointed that the market ignored it. I was like, that's >> it is what it is. >> So, I think it was great. And I think that's I diverged at that. I just wanted to acknowledge how incredible management bill is for service now and that once we get past this narrative of AI will kill software, I do believe that service now will be one of those like misread plays looking back in a couple years. But it's also interesting because I don't know if you got a chance to look at any of the media. I came up came up with this like legacy layer test that I think every investor needs to ask themselves if you're a software investor. And you have to ask yourself these three questions. Three questions are essentially if AI will eat you up or you're going to survive and thrive in the AI economy. First one is obviously the most common one for everyone. Are you a seatbased model living in a world where AI agents will reduce knowledge workers? That's the most obvious one. Everyone in their mother's been saying that's a disruption. I agree with that. That is a disruption that you're going to see a lot of names pivot away from. The second question is, is your product purely digital or does it actually touch the physical world? And I think that if it's purely digital, you are going to be at risk of AI becoming a major competitor for you. That's why I I believe that the like the palunteers of the world, the orchestration layer like their product touches the real world uh in the war war place war uh they affect war, they affect the logistics, they affect so much of the real world. That's why they benefit substantially from it. Same with security. Security security on devices. That's the real world. Uh Cloudflare, they have all the physical um those massive massive towers which helps them have connect all the uh devices towards the edge which is why they control 20% of the internet. So that's an example. So that's the second question. Is your product purely digital? Third, this is probably the most important one. Do you actually own the data that AI agents can't generate? That is the big question. So if your if your software company is a falls for each of those three questions of oof they RC based model it's purely digital and we don't owe the data you are at risk of becoming the legacy layer of the new agentic AI stack and oh >> if you're a feed if you're just a zero copy of other data and you're just you know some sort of substrate or some sort of like u semantic layer that just sits on everyone else's stuff that that's gone. It's going away. It's >> and that's that's 90% of software by the way guys. That's a lot of software. >> Yeah. You know, it's it's again if that data can be accessed, it's a great way to think about it. Ask the core data that's used for that application, where is it and who owns it? Mhm. >> Because if it's if it's available in the open internet, you know, there's not a zero value. Like there's companies like Alpha Sense that have built these like interesting data aggregation layers, but they're APIs. They're basically their business model is going to be a call like some other software like like an Enthropic. You're just going to pay a little bit of money to be able to access what's behind their firewall if it's organized in a And by the way, over time, as the as the database and the ability for AI to read uh schema better and it gets better and better, it will increasingly remove the moat. >> So I I'm really happy you brought that up because what I mean by like if it qualifies for all three of those questions, I'm not saying it's going away. I'm saying it's going to become a commodity. We have a perfect example of an incredible business that eventually became a commodity and what the valuation the market is willing to price that commodity. PayPal very different business but payment services and is a commodity these days and PayPal's huge mind share VMO etc. the market's pricing them at six times earnings. So, what I'm saying is if a software company, they're not going to go away, but if they qualify for all three of those questions, they're probably going to be trapped between the five to eight times earnings for multiple. And I think if you're an investor trying to capitalize an AI, why are you trying to catch a falling night that's going to become a legacy valuation that's five to eight times earnings? You're not going to try to do that. So especially when um Dan and I have been like saying this as well where we think it spending is going to drastically increase in next five plus years. So connect the dots guys. If 90% of software is going to start getting devalued become the legacy layer where's all that incremental spend that's going to go towards all the AI native customers right now and also the names that the 10% that actually proved they will be winners. So that 10% is going to obviously start start having a scarce scarcity premium because a lot more flow is going towards those names rather than the legacy layer. So I think it's really important to be taking our emotion out of this because this rerating is real. I disagree that all software should be treated the same right now but we're at the phase where market sensing everyone then they're going to ask questions later and that's where the opportunity lies. But it's important to go through the exercise for each of those names. And I think >> a lot of retail has poured back in. They retail's been buying software like crazy, buying the dip. >> Um and and I think it's important that you know that people listen and take what you're saying seriously is do not make the mistake of thinking buying all software is a good idea. >> Yeah. you know like do your homework, pick the winners, answer those three questions plus Mike basically understand is the foundation in which this software sits on top of basically I I like a simple even a simpler question that you asked is it mission critical is that software mission critical to a business if it is you know there's be a lot harder to rip out like ripping out an Oracle or an SAP by the way a lot of people might want to but It's like it will completely um can can it will you decapitate a business. These businesses depend on these fun these capabilities. Um you know you go up and downstream like if you rip out the project management tool and and and build something else or you rip out a certain HR tool or something will it cause that same kind of pain? And if it if the answer is no then I would say steer to the ones that absolutely cannot function without that software. Um, and there's also like I said a lot about like I look at governance a lot like do these particular softwares like are they are transactions flowing through them dollars because that's also another big thing like people don't mind using an AI tool to to help with a content creation marketing SAS workflow or something like that but are you going to actually transact on that platform meaning >> money moving from across borders um taxes uh you know like literally corporate governance uh are are legal filings crossing that platform like things like that are not going to just get blown up quickly. >> Yeah, >> there's too much risk. >> So I mean what what you're saying essentially by this which I agree with is like we are still in somewhat the we're in the deployment phase of AI but still somewhat experimental. So like if you're not going to trust some AI company that got started yesterday to touch your ERP system because if something does go wrong, they haven't had the experience or time to test the thousands of bugs that I don't know like an Oracle has done or even a Salesforce has done. So there is somewhat of a advantage in that, but the these companies do have to make sure they make the necessary moves to fit what the new AI economy is going to be because you can't just latch on to this system of records advantage because eventually that's going to erode because AI is just becoming better and better and better. But as of right now and probably for the foreseeable future for the next couple years, you can't just like trust this AI native customer to touch your ERP because god forbid something goes wrong. You can't say oops sorry AI did it that that won't work as an response to these clients. So it is very real >> spilling data and don't forget security security and government it's like it's [snorts] not trivial. Well, I actually have a question for you because I heard this Chimat said this and I was like I disagree with this take but it's a really interesting exercise to have in your head is the reason that maybe a lot of these cloud stocks software stocks are struggling is there might be a full circle moment with going back towards onrem and I think by there is always leakage if you're on the cloud like it's almost 100% there's somewhat of leakage do you want to risk your sauce that you have an AI that your competitor might be able to extract ract from through that leakage or do you want everything on prem protected and you're willing to suck up the margins in order to do so? I disagree with that but curious your take on that. >> Well, I think there's a there's some arguments for prem. I mean first of all if you're a large enterprise and you're doing a lot of token generation like you're paying a premium so if you can build your own AI infrastructure run your own you know some of your own inference on your own infra um if you have scaled applications there's definitely a case to be made for being your own token factory. So I think some big companies will run inference on on uh prem because I think you'll get good economics when you start like think about what agents at scale like what we're going to spend on cloud it's going to get expensive. Um I I mean this has been an argument as old as time like there's been little indicators over time of work workloads coming back on prem but it just has never quite manifested or materialized the way the stories are like everyone's going to it's like I don't know if people are going to want to build out a lot more like I think large companies do this like they can do the the economics on prem versus cloud but I think these businesses is they're trying to get lean and efficient. The cloud just gives such great consumption economics in terms of like they and I just think that you really have to be a fairly scale business with very specific talents and resources to make it make sense to to to build more infrastructure on prem right now. >> I like that. That makes sense. I mean I think I was trying because the reason I'm trying to pinpoint this is like security stocks like cyber security stocks have gone beaten down to extent where it doesn't make sense on this AI narrative. So like is there something else under underplaying this? So maybe it was on prem thing but yeah I agree it's more of a hyperscaler might do this but I think a lot of the the SMBs of the world uh mediumsiz companies will not but it's it's interesting exercise to play. Um, next question I wanted I think we should talk about as well is Palo Alto because there was a very big divergence between what the street was expecting on guidance and what was actually produced. What was your takeaway from Palo Alto's earnings? Because I know you and I both talked about this during our media circuit this week. I mean, Palo Alto, first of all, this is a company that's trading at multi-year lows right now. And so, I struggle with that valuation based on its it's accreative just the accretive nature of Cyber Arc and other deals that it's done. >> It's built such scale and I think that scale is really really going to be important. Think about like OpenClaw or Claudebot, Mobot, whatever it is. Like think about all the security vulnerabilities that are being brought in as people are trying to race to agentic capabilities in their in their enterprises. Every agent is like a badgeless, nameless, faceless employee with endless credentials and access to tools that can can get into the system like >> Yeah. >> So everything from perimeter to cloud to AI sort of security feels like such a big opportunity. And so it's one of those where the market is clearly not rewarding this trend yet. It does not fully appreciate it or see it. I don't think AI is going to eat cyber security. I think AI is going to uh accelerate the need for cyber security. So then the question is is you know Nikesh Aurora the CEO's master plan of being a acquisitive super platform the route and the bet is basically can he bring synergies out of these these all these companies he's kind of strapping and bolting together and so you know the numbers the guide people didn't love it but like I think he outperformed what each of the two individual businesses were estimated at. >> Yeah. in the combined, but it just wasn't by enough. But I think it takes more time for the synergies to come out. I I still like the trend and like crowd Paulo and then like the Google's um and to some extent like on the on the on the enterprise side like the Cisco there's only a small number of players that really have these kind of full stack security offerings and I think that I think the market needs more consolidation. Security is too fragmented. >> I agree 100%. It's just >> But I think there's some patience. >> I think you have to be patient. I don't I don't see this thing running up anytime soon. >> No, I think it's it's in the penalty box a bit. And I think I was trying to think about past instances where the Inquisitive model at this scale has worked out. Like I think a lot of people are very worried that this is Salesforce 2.0. And what I mean by that is Salesforce obviously went down this >> bolt on bolt-on, bolt, >> bolt-on, bolt-on. And now they they had a couple errors, large errors, and they have 50 billion in goodwill sitting on their balance sheets that is dedicated to companies in the old world where maybe it could have gone towards what the new AI economy is going to be. There is somewhat of a fear of like you make these large purchases. I agree with you. I think Cyber Arc was a steal. I think Cyborg probably made the mistake of selling out at the when they did because the genetic AI moment was just had two months ago and machines are going to outnumber humans 100 200 times uh to one and that's going to be huge for identity like you mentioned like every AI agents a device that just you can't track unless you need identity associated with it. So I think that the worry is it wasn't just Cyber Arc they spent Chronosphere was five or six bill and they had another one. I think there's been like if you continue having make these acquisitions multiple times a year there's going to be obvious synergy risks and you can't hit 100% on all of them. Eventually there might be some duds that are going to be anchoring your margin expansion and that is a worry but the theme is super strong and has a lot of legs in the runway and if you want Palo Alto to become this network security platform they kind of need to make these acquisitions because they came from a hardware lens and trying to become more of a cloud lens. They're succeeding in that but it's going to take some time. So I do think that's Paulo Alto is one that's trading at 30 times free cash flow for example. Like it feels fair for the amount of noise that earnings report had. Like I think that pullback was very fair. You have to wait. >> It's like multi-year low. It's a multi-year low >> with all the accretive value of the acquisitions. So it's it's you're getting kind of like the Cyber Arc deal for free right now. If you liked it before, you're kind of getting it like it's included in the price. That's that's a good that's a good question to ask yourself like does is it fair to pinpoint to the pre agentic AI world valuations now because there's always going to be somewhat of a risk now if AI is just going to disrupt everything I disagree with it but there might be a philosophy on >> I seen a sign yet that AI will I mean we've been by the way cyber have been trying to use AI for a long time to you know I guess AI is moving faster and agents are moving faster but like using AI and ML tactics to identify uh cyber fraud and stuff has been being done for decades like literally like more than a decade. Um so it's it's not like just a new thing like AI and cyber you know I think you have to basically I think you made a good point about kind of thinking about consolidation and picking winners. What I do think is coming out of this AI trend is industries that maybe could have 20 or 30 names that were highly performant like in SAS might be five and in cyber it's kind of the same thing like where you know in some of these industries like we are going to get to the point where there's going to be like the leader there's going to be like the second thirds and then there's everyone else and you know in some of these industries they enjoyed having 20 30 40 names that could all perform all keep growing all keep going up and I think fragmentation is going to trade for consolidation in cyber and so it's basically are you picking crowd are you picking um Cisco are you picking Paulo are you picking like a hyperscaler to build more security in-house like a Microsoft which by the way the largest security company in the world Microsoft >> um are you picking someone there and that's where you want to put your dollars because you don't see cyber as an independent platform you see cyber as a bakedin capability that sits inside of a broader platform play and those are the questions you really have to ask. >> So, I I think that framing is so accurate what the future's going to be like because I think that the distribution is going to be insanely uneven going forward towards the clear winners and everyone else. I think you're seeing that you're going to see that in enterprise IT. You're going to see that in cyber security. You're going to see that across the board where there's going to be even more insane premiums and then everyone else is going to be trading at multi-year averages. And I think if you're caught in between, that's the risk. Like I think that's Crowdstrike. I saw one of the questions here like what's your uh which cyber stock is your number one pick? Uh for me it's been crowdstrike because they've organically proven to be able to create their own additional modules past couple years of the latest and greatest through the Falcon platform. Yeah, they had a huge slip up on shutting the world down for a couple hours uh years ago, but George did an incredible job getting himself out of that hole. And I think now they're about to finally get out of that fundamental hole because they have still have to pay for the credits uh from that mess. I think they're positioned really well. And I think that next one for me, we haven't talked about this too much, rubric. I really do believe that AI agents it's there's going to be millions of always on AI agents are going to be continuously operating autonomously. I think it's really important to have that undo button essentially for these AI agents if something goes wrong for these attacks like you need a data vault. I'm I'm curious on your input uh u thoughts about Zcaler because Zcaler is one where I don't know if it could be a full platform. I love how they fit in the cyber security stack, but it is hyper like if you want to talk about fragmented niches in the ecosystem. What Zcaler operates in is highly competitive and I think they have a they're not all public and that's the tricky thing about Zcaler's competition but super competitive but they've gone destroyed but their fundamentals are great. So what's your take on Zcaler? Um, you know, I I don't know if they're going to be they become too niche risk of being too niche. I'm not sure that the platform scale, you know, it's why I kind of like I said when I was coming even I forgot about them even when I was naming names just a minute ago. I feel like they're going to get either get acquired or kind of kind of sucked into something because what I think is you know right now and by the way this isn't in terms of businesses you know business surviv survival I'm just talking about in terms of investability like I really want broad platform I want agents I want AI I want cloud I want um you know scale I want to be thinking about like are you the type of security offering that could basically be the core of an enterprise's security strategy >> and if you're just serving one purpose like like you know endpoint security, you know like I get that there's a business for you out there I just don't know how investable you are. So, I think that's kind of what I'm seeing with some of these names that don't have sort of what I would call a either a current or a um at least a a pathway to being sort of full stack. >> I I I agree with that. No, I agree with that. I I always thought that the perfect marriage is bless you. Perfect marriage is Zcaler and Crowd Strike is the perfect marriage. I just don't think that Zcaler will always sell out. But I don't know like also another point is like Microsoft right now I'm looking at the chart they're trading 22 times forward earnings. That's their 7-year low is 21 times earnings. So we talk about pure play cyber security names all the time. Microsoft has the best cyber security product in the world. It's the largest as well. So like it's difficult when you can buy that business for seven times the like near the seven years lows. And that goes back to my kind of point though is like when you like a theme, you got to pick how you get exposure to the theme. >> Yeah. >> It's like a lot of people want exposure to robotics, but it's like I can get exposure with a, you know, with Elon Musk by getting into Tesla as opposed to picking something that's maybe much more specifically like a true robotics company. And what I'm saying is a lot of the MAG 7 type names can give you the exposure. Google's got some serious security exposure. Microsoft has serious uh security exposure. Um so I think right now especially when some of these names have pulled back to like you said to multi-year lows. Um it's hard to weigh against that. It's it's hard to get narrow and go niche when you can actually get the broad theme, especially when there's not real clear uh upside. meaning like you're not really super if you're not super confident that there's going to be that winner. >> Um so in moments like this I I I like that idea. I mean you know like a Microsoft right here with after it's just been it's had the crap beaten out of it. >> It's it's the Sam Alman effect for sure that just shows it to you. When 45% of your backlog is open AI I think the market's kind of penalizing for them. >> Your backlog's twice the size by the way. I know it's they essentially their backlog is showing everyone that they have like five plus years of revenue already recognized for Azure. It's kind of nuts to think about it that way. But um I I do want to bring up Nvidia earnings because I want us to give us our commentary ahead of that because that is the Super Bowl of every earning cycle. And I think this one specifically, Nvidia stock has gone nowhere for 6 months. And I think there's a lot of spreading of FUD uh that's probably impacting them. Maybe Google's TPU success is being used as a zero sum framing on why Nvidia shouldn't be catching a bit, which is ridiculous. But for me, I think everyone's going to be watching this name. I think everyone's be for me specifically, I'm going to see how Blackwell is ramping. I want to see what if Jensen provides more commentary when he was like preparing the chips the world has never seen before. It feels like he might be adding some more commentary on >> little Grock in there. Little Grock. >> Yeah, the Grock. >> That Grock would be big. I would love to see Reuben timing as well on a 2026 road map. Like I think there's a lot of gems that could be found this. >> Yeah, I think Evercore came out with a comment saying they think it might be three to six months forwards forward. I saw that and there there's just so much misinformation though out there that I don't know what's real. That's why I want to hear Jetson say it. Um that would be a game changer to be honest like that's a big change 3 to 6 months like that time window is what what's the time window from black to reuben now if that happens it's like alltime low of like >> yeah to be like a nine month cycle it's wild >> that's crazy that's early production >> that's it's just so nuts to me and that again like I do believe that it just shows how they're so different than everyone else Nvidia like they have so they can offer so many system solutions to everyone else that nobody else can offer that it's like we're all living in their world. But what's your take what's your pre-earnings thoughts that you want to share to everyone? >> Yeah. Um, [clears throat] so couple things. I mean, you hit it on the head. I mean, the ramp and and Ruben visibility is going to be really important, but I think he's proven that he can keep the train on the tracks. He's bought the memory capacity. He's bought the uh you know the the the the wafer capacity packaging. He's he's ahead of the game on all that stuff. Um the meta deal, by the way, what a win. I mean, >> oh my god. Yeah, it's >> not zero sum with AMD. And so people please stop thinking that just cuz they won something means like AMD's out. It's not. But it was a big win. And Jensen is absolutely uh you know brutal about like going in and just taking all the market he can every deal. I mean he's great. I mean God you should want him >> on your payroll if you can get them but you can for about $190 a share. Um the the margins will be important. Shai I mean everyone's watching. Can they creep the margins back up? People keep saying the margins are going to get hit. They haven't really. I mean I think they're going to expand. I mean the the consensus is like around 73 half%. Um the data center revenue up

66% 是预期。我认为他们会超越这个数字。我认为它会比预期的还要大。所以，你知道，我仍然看好那里。嗯，你知道，他们还可以暗示一些其他我认为会很有趣的事情，比如中国。我的意思是，我们从 AMD 的数据中看到了一些中国的流动。他们会从中国获得一点意外的顺风吗？嗯，当然，嗯，所以，主权人工智能趋势，他们每个季度都谈论过，但没有真正的数字，比如主权，我们看到世界各地都在建设数据中心，比如大型有意义的数据中心出货量，超出了超大规模用户，因为我们知道超大规模用户的数字非常可预测，但我们是否看到任何趋势线，中东发生了什么，欧洲发生了什么，发生了什么，比如建设是否发生在...

嗯，你知道，我们也会对指引非常感兴趣，因为指引完全关乎这里的持久性。我们知道我们很确定这个看起来不错。资本支出表明了接下来的几个，但你知道，他接下来会怎么指引，你知道，因为这是他们第四季度的结束，对吧？>> 是的。>> 所以，所以他会，你知道，他会谈论他明年全年的看法，这可能会非常令人兴奋。但我认为这可能是一个我们在这里稍微转向的时刻。就像再次，如果我所想的，如果 OpenAI 宣布他们的一千亿美元，而且它实现了，而 Nvidia 绝对碾压了它，它只会让整个泡沫熊市的故事少一些空气。整个泡沫熊市，你知道，一切都会破裂。就像，不。你知道，似乎资金一直在流入。购买是普遍的。数据中心正在建设。是的，系统中存在一些小裂痕。仅仅因为一个人被取消抵押品赎回权，并不意味着每个人都会被取消抵押品赎回权。>> 所以这里的类比也是一样的。会有一些坏账。会有一些糟糕的交易。但我想总体来说会很好。>> 是的，我认为这也会是一个强劲的。希望这将改变这个人工智能基础设施交易的叙事，这个交易显然已经暂停了。嗯，让我们看看还有什么在报告。我很好奇，另一个对我来说很重要的事情，因为你们很多人都知道我是 Snowflake 的多头。他们在周三报告，与 Nvidia 同一个下午。这就像他们每个季度都这样做，我就是不明白为什么。在我看来，一家人工智能公司在 Nvidia 的同一个下午报告，这简直是荒谬的。我知道这可能不是他们的错，但在我看来，这确实是他们的错。当你提前一两天看到的时候，你显然可以调整它。但对我来说，这是一个大问题，因为我认为我在这只股票上是看多的。但如果你回到传统的测试，三个问题之一是你的产品是纯数字的吗？Snowflake 是纯数字的。我认为数据平台是新人工智能经济的一个令人难以置信的切入点。但数据平台存在一个巨大的围墙花园方法。我认为你可以从 AWS 的产品中获得很多。你也可以从类似的超大规模用户产品中获得很多，而且我认为 Snowflake 长期以来一直存在一个数据砖问题，而且这个问题不会很快消失，因为它们无法与数据砖在私有市场上的利润率竞争，当它们定价过低时，而数据砖是在人工智能的视角下推出的，而 Snowflake 最近才成功地从数据仓库转向了人工智能数据平台视角，那么它就没有同样的优势了。所以对我来说，具体来说，我关注 Snowflake 的不是为了超越预期，因为我认为它们会超越。我只是认为我们需要看到关键指标的重新加速。就像 Cloudflare 能够在周期中获得关注，而几乎所有其他软件名称都无法做到这一点，是因为它们连续三个季度证明了 RPO 的重新加速。就像它们从两个季度前的 35% 的增长到上个季度的 39%，到这个周期它们达到了 49%。所以我认为这显示了加速的清晰迹象。我认为对于 Snowflake 来说，它们的净收入留存率一直固定在 125%。它们需要显示出加速，因为上个周期它们并没有真正达到那个水平的加速，但 RPO 从 33% 增长到 37%。这意味着它们签了很多大交易，但由于某种原因，部署或工作流程没有以应有的规模进行部署。所以对我来说，这是一个有点脱节。我认为它们需要更一致才能获得关注。我对此感到兴奋，因为我认为这是那些可能获得我所说的稀缺性溢价的股票之一，即成为人工智能的赢家，而现在它们的价格是，让我查一下，它们仍然不是最便宜的，这有点疯狂，即使在这次回调之后，它们的自由现金流市盈率也达到了 43 倍，而且它们有一个 SBC 问题，所以它们实际上并不盈利，所以它们实际上必须显示增长数字，而且它们需要增长，我不知道，30% 才能让市场满意。你对 Snowflake 有什么具体的看法吗？>> 我必须花些时间与他们的 CEO 交谈。我的意思是，他们确实有一个 SBC 的挑战。我确实认为他们有一个很好的人工智能转型故事。我确实认为数据基础层对于企业能够利用人工智能仍然非常重要。嗯，你知道，我没有像那样仔细地研究它，看看它的交易水平等等，来对它在这里的投资价值有一个强烈的看法，但我仍然认为它们是什么，基本上，这是一个几匹马的比赛。我听不清你。>> 不，我在这里。>> 对不起，我的屏幕变黑了。它回来了。它回来了。嗯，所以，你知道，我认为你已经很好地涵盖了那个。我正在查看我们下周的整个列表。而且，这将是，这将是激动人心的一周。我的意思是，开玩笑不说，你知道，我正在看我要看什么。周二市场开盘前，你有 Digital Ocean 和 Cipher。我知道你喜欢 Marcato Libre。嗯，>> 你有 Axon。你有 Zeta。Zeta 是一个有趣的公司。我知道我们在评论中有很多关于 Zeta 的问题。我喜欢 Zeta，因为它是一个真实的 AI ROI 案例。就像一家公司真正地在利用人工智能，如果你能利用人工智能并创造营销规模，它是可衡量的。嗯，Workday，人们会关注它。另一家 SAS 公司受到了打击。Tempest。Tempest 是人工智能医疗保健领域的颠覆者。嗯，你和南希·佩洛西都喜欢它。嗯，HP，我的天，他们买了足够的内存吗？他们能跟上 PC 的需求吗？他们的 CEO 离开了。他们公司内部有很多大的变动。那会很有趣。你有 HUD 8，所以周三开盘前又有一个矿工。嗯，我们有 Circle，你知道发生了什么。稳定币。还记得曾经的狂潮吗？>> 是的。>> 它回到了现实。然后周三收盘后，我的意思是，顺便说一句，是 Nvidia、Tradeesk、Salesforce、Snowflake、Ion Q、Synopsis、Pure Storage 和 Nutanix。我的意思是，多么棒的科技公司阵容啊。我的意思是，当然，就像 Nvidia 一样，对吧？嗯，周四早上你有更多的量子计算，有 DWave。你有 EOS Energy。嗯，周四收盘后，你有 Corewave。所以，那个可能很有趣。还有，伙计，周四你有 Rocket Lab。>> 我知道。>> Rocket Lab。哦，你问了 Zcaler。他们下周。Dell，伙计。你下周会得到一个 Dell。嗯，它们一直是服务器需求的一个有趣的晴雨表。嗯，Dolingo，我不太认真对待它，但它一直是一个表情包名称。所以，我的意思是，但说真的，多么棒的一周啊。我的意思是，>> 不，这很紧张。这将是紧张的一周。我认为我之前评论过。嗯，这个周末好好休息一下，因为这将是你会看到很多变动的一周。而且我非常沮丧，我现在只有 3% 的现金头寸，因为我认为下周会有机会，有这么多强大的公司在报告。但每一天都会有事情发生。这将非常有趣。如果你是收益周期迷，这就是你的超级碗。所以享受它。做好准备。但让我们来回答一些问题，因为我认为我们有很多，我喜欢互动，伙计们。我们有很多问题。所以，我想尽可能多地回答，因为我们有 >> 我试图在进行过程中巧妙地回答一些问题，因为我确实觉得，你知道，有时候需要很长时间，然后人们等待，你知道，然后他们会伤心，因为我们没有实时回答他们。>> 我们看到了你。我们看到了你。>> 我们看到了你。>> 我们看到了你。>> 所以，这是一个很好的问题。Sales um Service Now、Adobe、Salesforce。这三只股票中，你的主题是什么？我可以给出我的解决方案，我的答案。嗯，我认为 Service Now，如果你想通过传统的测试，我认为 Service Now 有最多的工作流程最终会产生实际的物理世界结果。就像他们有笔记本电脑发货、员工入职、技术支持已解决，就像所有这些都比 Adobe 和 Salesforce 更能触及物理世界。所以，我选择 Service Now。你呢？>> 嗯，如果我现在下单，可悲的是对于 Adobe，它将是 Salesforce，或者抱歉，Service Now，Salesforce，Adobe。我担心那个新的 Google 模型。>> 就像 Adobe 总是倒霉。尽管如此，Adobe 的数字还没有显示出任何麻烦的迹象。所以，这是我感觉他们很脆弱的一个。嗯，但我也还没有在现实世界中看到任何证据表明这种情况确实发生了。>> 是的，我同意。这是那些被定价为合法层面的股票，但如果不是爆炸性的重新定价。嗯，下一个。你说得对。我还没见过。有很多 Zeta 的评论，Zeta Zeta 的评论。>> Zelda，伙计。每个人都对 Zeta 感兴趣。>> 所以，对我来说，这是一个非常有趣的公司，因为我认为它们被困在两个现实中。我认为它们的护城河显然是它们确实拥有专有的消费者数据。它不仅仅是软件工作流程会面临风险，纯粹是数字化的。它们的人工智能确实能提高定位的准确性。嗯，然而，这些数据是否足够专有，以至于其产出能够消除所有产出都是数字化的风险？我的意思是，显然，我们看到迹象表明人工智能可以生成广告文案、细分客户，它们可以优化所有这些广告系列，如果人工智能使营销自动化层商品化，我认为 Zeta 的软件可能会面临压力，但 Zeta 拥有实际的，它们是远程的，数据就是它们。所以我认为它们正在将人工智能嵌入平台以增强结果，而且显然有效。所以价值是数据集，而不是工作流程软件，但我把它归类为 Service Now，就像你必须等待并观察它们在未来 6 到 12 个月内会拿出什么，看看它们的层面是否会成为那个传统层面。>> 是的。我的意思是，将人工智能应用于实现营销成果并非易事。好的。我的意思是，就像你和我输入一些想法，然后说给 Claude 五条推文很容易。嗯，让它实际管理营销层，拥有所有专有数据，并将其暴露给某种广告定位信息。嗯，还有很多工作要做。就像我说的，就像 Claude 会告诉你如何做，但真正做到这一点的东西并不完全存在，而且它总是会与 DIY 和那种可以为你完成所有事情的平台进行权衡。我认为这很有价值。企业不想从事开发游戏。>> 是的。所以，如果这些软件能够轻巧高效，并且能够快速搭建，并且具有代理性，我认为人们会为此付费。我只是认为人们不想使用旧的营销软件，上面加了一个代理，当他们真正地，你知道，但我认为我喜欢他们真正地试图将 ROI 应用于人工智能。>> 我认为这很重要。>> 是的。而且它们有一个很好的故事，我认为故事在这些天也很重要，尤其是在软件领域。下一个问题。嗯，我认为对我来说，我可能必须回答这个问题。Rock collab。嗯，太空主题的想法。伙计们，2026 年仍然是太空年。就像价格走势的回调表明的那样。这更多的是一种风险规避的原因，为什么所有这些太空股票都在回调。但 SpaceX 以看起来像 1.5 万亿美元的估值上市，这对整个行业来说是超级利好，因为 SpaceX 生态系统的每个组成部分实际上都会被重新估值，因为 SpaceX 以这个估值被重新估值。是的，Starlink 是它们业务的大部分，但我们终于得到了整个生态系统的锚定，以及它的估值应该是多少。当你拥有像 Rock collapse 这样价值 400 亿美元的股票时，我甚至不知道它们现在的估值是多少。嗯，我认为这将成为太空经济的第二个来源，而且只估值 400 亿美元的市值，而它们预计在未来三年内以 40% 的复合增长率增长，并在明年实现盈利。这对我来说是不可理解的。所以，Rockab 的收益有一个锚定，那就是 Neutron 会延迟，而且我们真的不知道确切的日期。Neutron 是它们下一个催化剂，以获得下一次股票价格的重新估值。这是一个中型发射，最终真正与 SpaceX 竞争。这就是今年 Rocklabs 的游戏名称，它不仅需要一个最终日期，还需要成功交付。但就目前而言，我的意思是，它正在消化一点。从技术上讲，它看起来会在短期内进一步回调，但它在某种程度上保持得很好。它没有保持得很好。今天下跌了 5%。它保持在 21 周 EMA 水平之上。所以，那个水平是 66 美元。所以，如果它跌破那个水平，可能会有更多的痛苦。但是的，嗯，我知道你对此没有太多看法。最后一个话题，让我们谈谈内存。我认为内存可能是 Nvidia 下周财报的最大话题之一。嗯，你对那有什么期望？>> 我认为他们已经比其他人更了解这一点了。我认为他们已经购买了更多。他们的规模将是 Nvidia 的巨大优势。就像 Nvidia 的巨大优势。嗯，我们知道所有内存制造商都将把内存容量转移到更高价值的产品上。更高价值的是 HBM。是数据中心。你将看到这种转变是多年的。嗯，所以这将非常有利。我长期以来一直在说关于美光的好话。嗯，我现在感觉非常 vindicated。嗯，任何听我建议的人在这只股票上都做得很好。嗯，但我认为这不是一个短期趋势。我认为这是一个非常非常长期的趋势，超过三年。嗯，然后我认为你不会看到像历史上那样的内存低迷。我们可能会在某个时候看到内存正常化，但我认为它会保持在高位更长时间。它将是一个非常强劲的类别。那些买得好的人，那些发现了定价并认为他们可以通过最后一刻购买来省钱的人，将会非常非常糟糕。而且我认为消费者将面临长期的痛苦，因为你会看到很多用于手机和 PC 的内存容量将被重新分配到数据中心。嗯，它将推高价格。它将推高价格，并将降低可用性。而且我确实担心一些 PC 业务。我有点担心 HP。联想已经发布了关于其 PC 业务的警告。>> 高通也是。上个财报周期，在他们的，你知道，他们正在转向数据中心和机器人，这很好，因为>> 手机业务将会受到打击。苹果也发出了警告。所以，苹果甚至不得不发出警告，而苹果从来不会在供应链方面出错。所以，>> 我认为它让很多人措手不及，但我怀疑它让黄仁勋措手不及。>> 是的，没有什么能。他掌控着整个经济，在我看来。但是，[笑声] 他就只有这个了。>> 他和达尔杜·特朗普。顺便说一句，我们没有真正谈论关税，但我会在这里结束。>> 是的。>> 我只是，我必须在这里结束，因为我还有另一个电话。这将是一个多年的机械练习，在行政和司法部门，因为他们将使用另一种机制，然后最高法院将不得不回去，并决定那种实施机制是否...因为他们所说的只是，他使用的紧急法案在这种情况下不是他权力的延伸。他们没有说关税是合法的。他们说他实施它们的方式不符合规定，或者不属于他的权利或他的权力范围，但现在他可以找到一种新的方式来实施它们。所以，我认为它会拖延。所以，好了，我得走了。>> 好了，这就是我们。就是这样了，各位。嗯，下周再见。忙碌的收益周期。>> 我们将讨论这个超级周期的结果。我们下周见。非常感谢。>> 再见。热。热。[音乐]