<p>如果您是一位了解情况的投资者，您很可能会阅读您投资的公司的季度报告。如果您想要更深入的了解，您可能会听管理层在季度收益电话会议上的评论。但如果您仍有额外时间，听取投资者会议上的演讲可能会提供更多见解。</p>
<p>9 月 12 日，数据公司 <strong>Snowflake</strong>(SNOW -3.29%) 在 <strong>高盛</strong>Communacopia + Technology Conference 上做了一个演讲。在他的讲话中，首席财务官 Mike Scarpelli 说了 16 个对投资者非常重要的词。这甚至让我重新考虑这支股票作为一个潜在的投资。</p>
<h2>Snowflake 正在进行重大改变</h2>
<p>人工智能（AI）趋势不可否认，Snowflake 也被卷入其中。作为证据，只需看看它的新闻稿。其 2022 财年第二季度的报告开始时简单地称自己为&#34;数据云公司&#34;。但在其 2025 财年第二季度的报告中，它称自己为&#34;AI 数据云公司&#34;。</p>
<p>在我看来，几乎所有科技公司都在努力说服投资者他们多年来一直是 AI 领域的领导者。但实际上，大多数公司都在积极投资于 AI，拼命地试图跟上。我认为 Snowflake 也属于这一类。事实上，该公司一直在购买图形处理单元（GPU）以满足 AI 工作负载的需求。</p>
<p>问题在于 GPU 价格昂贵，这损害了 Snowflake 的利润率——稍后会详细讨论。管理层一直表示对 AI 的投资正在进行中。但在最近的会议演讲中，Scarpelli 的语气听起来大不相同。</p>
<p>在谈到 GPU 昂贵的价格标签时，Scarpelli 出人意料地说:&#34;在看到支持它的收入之前，我不会再购买任何 GPU。&#34; 这 16 个词既令人震惊，对我来说也是一股清新的空气。</p>
<h2>这为什么重要</h2>
<p>在我的投资生涯中，我见过股票的估值超过 100 倍销售额，也见过市值超过 1000 亿美元的公司。但很少见到两者兼具，就像 Snowflake 股票在上市后不久那样。</p>
<p>投资者愿意让 Snowflake 股票成为我见过的最昂贵的股票之一，因为它有惊人的增长率和令人羡慕的利润率。普遍共识是，该业务的增长速度如此惊人，以至于它很快就会价值远远超过 1000 亿美元。</p>
<p>AI 也被认为是 Snowflake 的催化剂。毕竟，它是一家数据公司，而没有数据，AI 是无法运行的。但 AI 并没有像许多人预期的那样成为催化剂。公司的增长率仍然强劲，但却稳步下降。对 AI 基础设施的投资也促使管理层下调了利润预期。</p>
<p><img src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/c2a6aeb5b53a2d7b713649d8b911c42e.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/c2a6aeb5b53a2d7b713649d8b911c42e.png"/></p>
<p>YCharts 提供的 SNOW 营收（季度同比增长）数据。</p>
<p>具体来说，Snowflake 在 2025 财年开始时预计其产品收入增长 22%，运营利润率为 6%，调整后的自由现金流利润率为 29%。但截至第二季度，预计增长率为 26%，运营利润率为 3%，自由现金流利润率为 26%。</p>
<p>Snowflake 的营收前景是向上的，是的。但不要急于将其归因于 AI。正如 Scarpelli 关于 AI 所说:&#34;事实是今天很少有人大规模使用它。&#34; 因此，似乎 AI 并没有推动收入增长。但 AI 正在推动支出，因为公司正在购买 GPU——这可能是为什么沃伦·巴菲特的公司最近出售了 Snowflake 股票。</p>
<p>一般来说，投资于支出上升且没有收入增长的公司可能是危险的。这就是为什么我觉得 Scarpelli 的评论令人耳目一新。他不会再购买任何 GPU，除非它实际开始为业务增长做出贡献。</p>
<h2>是时候重新考虑 Snowflake 股票了吗？</h2>
<p>当公司不加区分地在最新的&#34;大事&#34;上花钱时，这可能是永远浪费的钱。但当企业审慎审视支出并要求看到财务效益时，股东可能会受益。<strong>Meta Platforms</strong> 是一个很好的最近例子。在将 2023 年定为&#34;效率年&#34;之后，该公司的利润飙升，而不牺牲增长。</p>
<p><img src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/e9398d87725d525958ee6bf88306f400.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/e9398d87725d525958ee6bf88306f400.png"/></p>
<p>YCharts 提供的 META 营收（TTM）数据。</p>
<p>就 Meta Platforms 股票而言，自 2023 年初以来已经翻了四倍。简言之，当它专注于证明其支出合理时，其利润飙升，将股票推至新高。</p>
<p>Snowflake 股票可能会发生类似的情况吗？现在还为时过早。但该公司的首席财务官要求在 AI 方面看到投资回报。这是正确的心态，可能会为 Snowflake 带来更好的效率。</p>
<p>以 11 倍销售额计算，Snowflake 股票从未如此便宜。如果它继续以目前的速度增长并控制支出，我相信更好的日子可能就在前方。</p>

SNOW

<p>雪花计算（Snowflake）的首席财务官 Mike Scarpelli 最近表示，在收入支持之前，公司不会再购买更多的 GPU，这表明在人工智能相关支出不断上升的背景下，公司战略发生了转变。尽管雪花计算的增长强劲，但利润率受到 GPU 成本的影响，导致了谨慎的展望。公司预计到 2025 财年产品收入增长 26％，但人工智能尚未显著提升收入。这种谨慎的支出方式可能提高效率，潜在地提升股东价值，类似于 Meta Platforms 在专注于合理支出后取得的最近成功</p>

<p>If you're an informed investor, you likely read the quarterly reports from the companies you've invested in. If you're looking for even deeper insight, you might listen to management's commentary on quarterly earnings calls. But if you still have extra time, listening to presentations at investor conferences can provide even more insight.</p>
<div class="lb-trans"><p>如果您是一位了解情况的投资者，您很可能会阅读您投资的公司的季度报告。如果您想要更深入的了解，您可能会听管理层在季度收益电话会议上的评论。但如果您仍有额外时间，听取投资者会议上的演讲可能会提供更多见解。</p>
</div><p>On Sept. 12, data company <strong>Snowflake</strong>(SNOW -3.29%) made a presentation at the <strong>Goldman Sachs</strong> Communacopia + Technology Conference. In his remarks, CFO Mike Scarpelli said 16 words that are extremely important for investors. They even have me rethinking the stock as a potential investment.</p>
<div class="lb-trans"><p>9 月 12 日，数据公司 <strong>Snowflake</strong>(SNOW -3.29%) 在 <strong>高盛</strong>Communacopia + Technology Conference 上做了一个演讲。在他的讲话中，首席财务官 Mike Scarpelli 说了 16 个对投资者非常重要的词。这甚至让我重新考虑这支股票作为一个潜在的投资。</p>
</div><h2>Snowflake is making a big change</h2>
<div class="lb-trans"><h2>Snowflake 正在进行重大改变</h2>
</div><p>The artificial intelligence (AI) trend is undeniable, and Snowflake is swept up in it. For evidence, just look at its press releases. Its report for its fiscal Q2 2022 started with simply calling itself &#34;The Data Cloud company.&#34; But in its report for its fiscal Q2 2025, it called itself &#34;The AI Data Cloud company.&#34;</p>
<div class="lb-trans"><p>人工智能（AI）趋势不可否认，Snowflake 也被卷入其中。作为证据，只需看看它的新闻稿。其 2022 财年第二季度的报告开始时简单地称自己为&#34;数据云公司&#34;。但在其 2025 财年第二季度的报告中，它称自己为&#34;AI 数据云公司&#34;。</p>
</div><p>From my perspective, almost all tech companies are trying to convince investors that they've been leaders in the AI space for years. But in reality, most are investing aggressively in AI, desperately trying to just keep up. I'd lump Snowflake in this bunch. Indeed, the company has been buying graphics processing units (GPUs) to have what it needs to power AI workloads.</p>
<div class="lb-trans"><p>在我看来，几乎所有科技公司都在努力说服投资者他们多年来一直是 AI 领域的领导者。但实际上，大多数公司都在积极投资于 AI，拼命地试图跟上。我认为 Snowflake 也属于这一类。事实上，该公司一直在购买图形处理单元（GPU）以满足 AI 工作负载的需求。</p>
</div><p>The problem is that GPUs are expensive, and it's hurt Snowflake's profit margins -- more on that in a moment. Management has been saying that the investment in AI is ongoing. But Scarpelli's tone sounded much different during his recent conference presentation.</p>
<div class="lb-trans"><p>问题在于 GPU 价格昂贵，这损害了 Snowflake 的利润率——稍后会详细讨论。管理层一直表示对 AI 的投资正在进行中。但在最近的会议演讲中，Scarpelli 的语气听起来大不相同。</p>
</div><p>While talking about the expensive price tags of GPUs, Scarpelli surprisingly said: &#34;I'm not going to buy any more GPUs until I see the revenue to support it.&#34; Those 16 words were both shocking, and, for me, a rare breath of fresh air.</p>
<div class="lb-trans"><p>在谈到 GPU 昂贵的价格标签时，Scarpelli 出人意料地说:&#34;在看到支持它的收入之前，我不会再购买任何 GPU。&#34; 这 16 个词既令人震惊，对我来说也是一股清新的空气。</p>
</div><h2>Why it matters</h2>
<div class="lb-trans"><h2>这为什么重要</h2>
</div><p>During my investing career, I've seen stocks richly valued at more than 100 times sales, and I've seen companies with enormous market caps above $100 billion. But I've rarely seen both, as was the case with Snowflake stock shortly after it went public.</p>
<div class="lb-trans"><p>在我的投资生涯中，我见过股票的估值超过 100 倍销售额，也见过市值超过 1000 亿美元的公司。但很少见到两者兼具，就像 Snowflake 股票在上市后不久那样。</p>
</div><p>Investors were willing to make Snowflake stock one of the most expensive I've ever seen because it had a phenomenal growth rate and enviable profit margins. The general consensus was that the business would grow at such an impressive rate that it would be worth far more than $100 billion in short order.</p>
<div class="lb-trans"><p>投资者愿意让 Snowflake 股票成为我见过的最昂贵的股票之一，因为它有惊人的增长率和令人羡慕的利润率。普遍共识是，该业务的增长速度如此惊人，以至于它很快就会价值远远超过 1000 亿美元。</p>
</div><p>AI was also believed to be a catalyst for Snowflake. After all, it's a data company, and AI doesn't work without data. But AI hasn't been the catalyst that many expected. The company's growth rate is still strong, but it's plunged steadily nonetheless. Investments in AI infrastructure also motivated management to lower its outlook for profits.</p>
<div class="lb-trans"><p>AI 也被认为是 Snowflake 的催化剂。毕竟，它是一家数据公司，而没有数据，AI 是无法运行的。但 AI 并没有像许多人预期的那样成为催化剂。公司的增长率仍然强劲，但却稳步下降。对 AI 基础设施的投资也促使管理层下调了利润预期。</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/c2a6aeb5b53a2d7b713649d8b911c42e.png" alt="" original-src="https://media.ycharts.com/charts/c2a6aeb5b53a2d7b713649d8b911c42e.png"/></p>
<p>SNOW Revenue (Quarterly YoY Growth) data by YCharts.</p>
<div class="lb-trans"><p>YCharts 提供的 SNOW 营收（季度同比增长）数据。</p>
</div><p>To put specific numbers here, Snowflake started fiscal 2025 expecting 22% growth for its product revenue, as well as a 6% operating margin and a 29% adjusted free cash flow margin. But as of Q2, it expects 26% growth, a 3% operating margin, and a 26% margin for free cash flow.</p>
<div class="lb-trans"><p>具体来说，Snowflake 在 2025 财年开始时预计其产品收入增长 22%，运营利润率为 6%，调整后的自由现金流利润率为 29%。但截至第二季度，预计增长率为 26%，运营利润率为 3%，自由现金流利润率为 26%。</p>
</div><p>Snowflake's outlook for revenue is up, yes. But don't be quick to attribute this to AI. As Scarpelli said regarding AI: &#34;The reality is that very few are using it en masse today.&#34; Therefore, it seems AI is <em>not</em> boosting revenue. But AI is boosting expenses because the company is buying GPUs -- this might be why Warren Buffett's company recently sold Snowflake stock.</p>
<div class="lb-trans"><p>Snowflake 的营收前景是向上的，是的。但不要急于将其归因于 AI。正如 Scarpelli 关于 AI 所说:&#34;事实是今天很少有人大规模使用它。&#34; 因此，似乎 AI 并没有推动收入增长。但 AI 正在推动支出，因为公司正在购买 GPU——这可能是为什么沃伦·巴菲特的公司最近出售了 Snowflake 股票。</p>
</div><p>Generally speaking, it can be dangerous to invest in a company with rising expenses and without a boost for revenue. That's why I find Scarpelli's comments refreshing. He's not buying any more GPUs unless it actually starts contributing to the growth of the business.</p>
<div class="lb-trans"><p>一般来说，投资于支出上升且没有收入增长的公司可能是危险的。这就是为什么我觉得 Scarpelli 的评论令人耳目一新。他不会再购买任何 GPU，除非它实际开始为业务增长做出贡献。</p>
</div><h2>Is it time to rethink Snowflake stock?</h2>
<div class="lb-trans"><h2>是时候重新考虑 Snowflake 股票了吗？</h2>
</div><p>When companies indiscriminately spend money on the latest &#34;big thing,&#34; it can be money forever wasted. But when businesses scrutinize spending and demand to see a financial benefit, good things can happen for shareholders. <strong>Meta Platforms</strong> is a good recent example. After dubbing 2023 its &#34;year of efficiency,&#34; the company's profits soared without sacrificing growth.</p>
<div class="lb-trans"><p>当公司不加区分地在最新的&#34;大事&#34;上花钱时，这可能是永远浪费的钱。但当企业审慎审视支出并要求看到财务效益时，股东可能会受益。<strong>Meta Platforms</strong> 是一个很好的最近例子。在将 2023 年定为&#34;效率年&#34;之后，该公司的利润飙升，而不牺牲增长。</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/e9398d87725d525958ee6bf88306f400.png" alt="" original-src="https://media.ycharts.com/charts/e9398d87725d525958ee6bf88306f400.png"/></p>
<p>META Revenue (TTM) data by YCharts.</p>
<div class="lb-trans"><p>YCharts 提供的 META 营收（TTM）数据。</p>
</div><p>For perspective, Meta Platforms stock has quadrupled since the start of 2023. In short, as it focused on justifying its spending, its profits skyrocketed, which lifted the stock to new highs.</p>
<div class="lb-trans"><p>就 Meta Platforms 股票而言，自 2023 年初以来已经翻了四倍。简言之，当它专注于证明其支出合理时，其利润飙升，将股票推至新高。</p>
</div><p>Could something similar be in the works for Snowflake stock? It's too early to say. But the company's CFO demands to see a return on investment when it comes to AI. That's the right mindset to have, and it could lead to better efficiency for Snowflake.</p>
<div class="lb-trans"><p>Snowflake 股票可能会发生类似的情况吗？现在还为时过早。但该公司的首席财务官要求在 AI 方面看到投资回报。这是正确的心态，可能会为 Snowflake 带来更好的效率。</p>
</div><p>At 11 times sales, Snowflake stock has never been cheaper. And if it keeps growing at its current pace while keeping expenses under control, I believe better days could be ahead.</p>
<div class="lb-trans"><p>以 11 倍销售额计算，Snowflake 股票从未如此便宜。如果它继续以目前的速度增长并控制支出，我相信更好的日子可能就在前方。</p>
</div>

Snowflake's CFO Just Said 16 Words That Should Be Music to Every Shareholder's Ears

- Snowflake CFO 表示将暂停购买 GPU，直到有收入支持。  
- 公司在 AI 投资上面临高成本，利润率受到影响。  
- 该策略可能改善 Snowflake 的财务效率，股价有望回升。

Snowflake 的首席财务官刚刚说了 16 个字，这对每个股东来说应该是好消息