<p>如果你在过去几年关注人工智能（AI）的发展，那么你很可能对 “大型语言模型”（LLM）这个术语并不陌生。</p>
<p>像 ChatGPT、Gemini 和 Claude 这样的 LLM 各自拥有众多功能和卖点，但这些模型的一个共同特点是，它们几乎可以瞬间回答关于任何主题的问题。在短短几年内，LLM 的崛起使得花费几分钟甚至几小时在互联网上寻找问题答案的过程变得过时。</p>
<p>虽然使用 LLM 可能很有趣，但你有没有想过这些模型是如何快速读取你的问题、从互联网上抓取信息并处理这些数据的？</p>
<p>也许 LLM 基础的最大支柱是一种称为图形处理单元（GPU）的硬件。GPU 存放在数据中心，并在服务器机架上以大型集群的形式存储。反过来，这些芯片能够处理大量数据，以生成对输入的详细响应。</p>
<p>目前，<strong>Nvidia</strong> 是 GPU 领域的领导者，拥有约 90% 的市场份额。Nvidia 在 GPU 领域的显著领先得益于其先发优势，这意味着自从 AI 成为科技行业的新热点以来，该公司几乎没有真正的竞争对手。</p>
<p>毫不奇怪，热切的投资者在过去几年推动了 Nvidia 股价创下新高。事实上，截至目前，Nvidia 是全球第二大市值公司。</p>
<p>然而，<strong>Advanced Micro Devices</strong>（AMD 3.70%）最近开始展示其在数据中心 GPU 市场的渗透能力。接下来，我将详细说明我认为 AMD 的 “ Nvidia 时刻” 可能终于到来的原因，并评估为什么该股票现在看起来是一个绝佳的买入机会。</p>
<h2>AMD 从 Oracle 收到好消息</h2>
<p>Nvidia 在 GPU 市场的先发优势也带来了定价权的丰厚利益。芯片需求的激增和缺乏竞争的结合使得 Nvidia 能以天文数字的价格出售其 GPU，使其成为公司高度盈利的产品。</p>
<p>能够持续花费数十亿美元购买 Nvidia 硬件的公司屈指可数，因此，Nvidia 拥有像 <strong>Microsoft</strong>、<strong>Meta Platforms</strong>、<strong>Tesla</strong> 和 <strong>Alphabet</strong> 这样的 “大七” 成员作为其最大客户也就不足为奇了。</p>
<p>争论 Nvidia 和 AMD 哪个芯片更优秀更像是一个主观的争论。话虽如此，AMD 的 MI300X 加速器的推出不仅仅是市场上的另一款 GPU。AMD 可以在价格上与 Nvidia 竞争，最近的趋势表明，一些大型科技公司可能正在寻找更具成本效益的方式来构建其 AI 基础设施。</p>
<p>例如，AMD 的 MI300X 加速器芯片在最近几个季度获得了 Microsoft 和 Meta 的客户。除了 Microsoft 和 Meta，<strong>Oracle</strong> 最近向投资者透露，它 “与 AMD 签署了一份价值数十亿美元的合同，以建立一个包含 30,000 个最新 MI355X GPU 的集群。”</p>
<p>虽然 AMD 进入 GPU 领域仍处于初期阶段，但我认为该公司在获取领先 AI 企业作为客户方面做得相当出色，尤其是那些在过去两年与 Nvidia 密切合作的企业。</p>
<p>鉴于 AMD 计划发布下一代芯片，我认为该公司在当前架构下的早期胜利是未来几年可能发生的积极迹象。</p>
<p><img src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MTEzNDkvZ2V0dHlpbWFnZXMtMTQ4MDYzMzI0MC01LmpwZyZ3PTcwMA?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://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MTEzNDkvZ2V0dHlpbWFnZXMtMTQ4MDYzMzI0MC01LmpwZyZ3PTcwMA"/></p>
<p>图片来源：Getty Images。</p>
<h2>AMD 的估值是个便宜货</h2>
<p>与其同行相比，AMD 的股价并没有受到与 Nvidia 相同程度的热情追捧。但在过去一年中，AMD 股价下跌了 47%，因此该股票变得便宜到不容忽视。</p>
<p>截至目前，AMD 的前瞻市盈率（P/E）为 22，是一年多以来的最低水平。</p>
<p>我认为投资者没有给予 AMD 的 GPU 业务足够的认可，并将该公司视为落后于 Nvidia 的企业。但在我看来，AMD 不需要在数据中心业务上超越 Nvidia，才能被视为值得投资的公司。</p>
<p>相反，我认为 AMD 当前的轨迹和令人印象深刻的客户胜利可能会导致持续加速增长的长期时期，这一切都得益于该公司在 GPU 领域对增量市场份额的不懈追求。</p>
<p>换句话说，AMD 可能不会成为芯片领域的第一玩家。但如果该公司能够展示其继续赢得大型合同的能力，并以与 Nvidia（已经是 GPU 行业的成熟玩家）相当甚至更快的速度增长，那么我可以看到增长型投资者涌向 AMD 股票，作为 Nvidia 的替代品。</p>
<p>尽管在过去一年表现平平，我认为 AMD 股票可能正朝着与 Nvidia 在 AI 热潮初期相似的路径发展。</p>

<p>AMD 在 GPU 市场上取得了进展，挑战 Nvidia 的主导地位。随着 MI300X 加速器的推出，AMD 已与微软、Meta 和 Oracle 等主要科技公司签订合同，表明其向更具成本效益的 AI 基础设施转变。尽管过去一年股价下跌了 47%，但 AMD 的股票现在被视为被低估，前瞻市盈率为 22。分析师认为，AMD 的增长轨迹和客户获取可能会在 GPU 领域带来显著的市场份额增长</p>

<p>If you've been following the artificial intelligence (AI) narrative over the last couple of years, then chances are you're familiar with the term &#34;large language model&#34; (LLM).</p>
<div class="lb-trans"><p>如果你在过去几年关注人工智能（AI）的发展，那么你很可能对 “大型语言模型”（LLM）这个术语并不陌生。</p>
</div><p>LLMs such as ChatGPT, Gemini, and Claude each boast a multitude of capabilities and selling points, but one of the overlapping features of these models is that they can answer questions on just about any topic almost instantly. In just a couple of years, the rise of LLMs has turned the process of spending minutes or even hours on the internet to find the answer to a question into an antiquated function.</p>
<div class="lb-trans"><p>像 ChatGPT、Gemini 和 Claude 这样的 LLM 各自拥有众多功能和卖点，但这些模型的一个共同特点是，它们几乎可以瞬间回答关于任何主题的问题。在短短几年内，LLM 的崛起使得花费几分钟甚至几小时在互联网上寻找问题答案的过程变得过时。</p>
</div><p>While LLMs can be fun to use, have you ever wondered how these models are able to read your question, scrape the internet for information, and process this data so quickly?</p>
<div class="lb-trans"><p>虽然使用 LLM 可能很有趣，但你有没有想过这些模型是如何快速读取你的问题、从互联网上抓取信息并处理这些数据的？</p>
</div><p>Perhaps the biggest supporting pillar of an LLM's foundation is a piece of hardware called a graphics processing unit (GPU). GPUs are housed inside of data centers and stored together in large clusters on server racks. In turn, these chipsets are able to process enormous volumes of data to generate detailed responses to inputs fed through the LLM.</p>
<div class="lb-trans"><p>也许 LLM 基础的最大支柱是一种称为图形处理单元（GPU）的硬件。GPU 存放在数据中心，并在服务器机架上以大型集群的形式存储。反过来，这些芯片能够处理大量数据，以生成对输入的详细响应。</p>
</div><p>Right now, <strong>Nvidia</strong> is the leader of the GPU landscape -- having acquired an estimated 90% market share. Nvidia built its substantial lead in the GPU realm thanks to its first mover-advantage, meaning that the company had very little in the way of legitimate competition since AI emerged as the hot new ticket in the technology sector.</p>
<div class="lb-trans"><p>目前，<strong>Nvidia</strong> 是 GPU 领域的领导者，拥有约 90% 的市场份额。Nvidia 在 GPU 领域的显著领先得益于其先发优势，这意味着自从 AI 成为科技行业的新热点以来，该公司几乎没有真正的竞争对手。</p>
</div><p>Unsurprisingly, eager investors have fueled Nvidia's share price to new highs over the last couple of years. In fact, as of this writing, Nvidia is the second most valuable company in the world as measured by market cap.</p>
<div class="lb-trans"><p>毫不奇怪，热切的投资者在过去几年推动了 Nvidia 股价创下新高。事实上，截至目前，Nvidia 是全球第二大市值公司。</p>
</div><p>Nevertheless, <strong>Advanced Micro Devices</strong>(AMD 3.70%) recently started demonstrating its own ability to penetrate the data center GPU market. Below, I'm going to detail why I think AMD's &#34;Nvidia moment&#34; may have finally arrived and assess why the stock looks like a screaming buy right now.</p>
<div class="lb-trans"><p>然而，<strong>Advanced Micro Devices</strong>（AMD 3.70%）最近开始展示其在数据中心 GPU 市场的渗透能力。接下来，我将详细说明我认为 AMD 的 “ Nvidia 时刻” 可能终于到来的原因，并评估为什么该股票现在看起来是一个绝佳的买入机会。</p>
</div><h2>AMD gets great news from Oracle</h2>
<div class="lb-trans"><h2>AMD 从 Oracle 收到好消息</h2>
</div><p>Nvidia's first-mover advantage in the GPU market also came with a lucrative benefit in the form of pricing power. A combination of soaring demand for chipware and lack of competition allowed Nvidia to sell its GPUs at astronomical prices -- making it a highly profitable product for the company.</p>
<div class="lb-trans"><p>Nvidia 在 GPU 市场的先发优势也带来了定价权的丰厚利益。芯片需求的激增和缺乏竞争的结合使得 Nvidia 能以天文数字的价格出售其 GPU，使其成为公司高度盈利的产品。</p>
</div><p>There are only so many companies out there that can afford to spend billions of dollars on a consistent basis to buy Nvidia's hardware, and so it shouldn't come as a surprise that the company boasts the likes of &#34;Magnificent Seven&#34; members <strong>Microsoft</strong>, <strong>Meta Platforms</strong>, <strong>Tesla</strong>, and <strong>Alphabet</strong> as some of its largest customers.</p>
<div class="lb-trans"><p>能够持续花费数十亿美元购买 Nvidia 硬件的公司屈指可数，因此，Nvidia 拥有像 <strong>Microsoft</strong>、<strong>Meta Platforms</strong>、<strong>Tesla</strong> 和 <strong>Alphabet</strong> 这样的 “大七” 成员作为其最大客户也就不足为奇了。</p>
</div><p>Debating whether Nvidia or AMD has the superior chipware is more of a subjective argument. With that said, the introduction of AMD's MI300X accelerators comes with more than just an alternative GPU on the market. AMD can compete with Nvidia on price, and recent trends indicate that some big tech companies may be looking for ways to build their AI infrastructure in a more cost-efficient manner.</p>
<div class="lb-trans"><p>争论 Nvidia 和 AMD 哪个芯片更优秀更像是一个主观的争论。话虽如此，AMD 的 MI300X 加速器的推出不仅仅是市场上的另一款 GPU。AMD 可以在价格上与 Nvidia 竞争，最近的趋势表明，一些大型科技公司可能正在寻找更具成本效益的方式来构建其 AI 基础设施。</p>
</div><p>For example, AMD's MI300X accelerator chips have garnered both Microsoft and Meta as customers in recent quarters. In addition to Microsoft and Meta, <strong>Oracle</strong> recently shared with investors that it &#34;signed a multi-billion dollar contract with AMD to build a cluster of 30,000 of their latest MI355X GPUs.&#34;</p>
<div class="lb-trans"><p>例如，AMD 的 MI300X 加速器芯片在最近几个季度获得了 Microsoft 和 Meta 的客户。除了 Microsoft 和 Meta，<strong>Oracle</strong> 最近向投资者透露，它 “与 AMD 签署了一份价值数十亿美元的合同，以建立一个包含 30,000 个最新 MI355X GPU 的集群。”</p>
</div><p>While AMD's foray into the GPU realm is still nascent, I think the company has done an impressive job acquiring leading AI enterprises as customers -- especially ones that have been working closely with Nvidia over the last two years.</p>
<div class="lb-trans"><p>虽然 AMD 进入 GPU 领域仍处于初期阶段，但我认为该公司在获取领先 AI 企业作为客户方面做得相当出色，尤其是那些在过去两年与 Nvidia 密切合作的企业。</p>
</div><p>Given AMD is scheduled to release next-generation chipsets, I find the company's early wins with its current architecture as an encouraging sign of what could be in store over the next couple of years.</p>
<div class="lb-trans"><p>鉴于 AMD 计划发布下一代芯片，我认为该公司在当前架构下的早期胜利是未来几年可能发生的积极迹象。</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MTEzNDkvZ2V0dHlpbWFnZXMtMTQ4MDYzMzI0MC01LmpwZyZ3PTcwMA" alt="" original-src="https://g.foolcdn.com/image/?url=https%3A%2F%2Fg.foolcdn.com%2Feditorial%2Fimages%2F811349%2Fgettyimages-1480633240-5.jpg&amp;op=resize&amp;w=700"/></p>
<p>Image source: Getty Images.</p>
<div class="lb-trans"><p>图片来源：Getty Images。</p>
</div><h2>AMD's valuation is a bargain</h2>
<div class="lb-trans"><h2>AMD 的估值是个便宜货</h2>
</div><p>Unlike its counterpart, AMD's share price has not received anywhere near the same level of enthusiasm compared to Nvidia. But with shares down 47% over the last year, AMD stock has become too cheap to ignore.</p>
<div class="lb-trans"><p>与其同行相比，AMD 的股价并没有受到与 Nvidia 相同程度的热情追捧。但在过去一年中，AMD 股价下跌了 47%，因此该股票变得便宜到不容忽视。</p>
</div><p>As of this writing, AMD is trading at a forward price-to-earnings (P/E) multiple of 22 -- its lowest level in well over a year.</p>
<div class="lb-trans"><p>截至目前，AMD 的前瞻市盈率（P/E）为 22，是一年多以来的最低水平。</p>
</div><p>I think investors aren't giving AMD's GPU business enough credit and view the company as a laggard compared to Nvidia. But in my eyes, AMD need not outperform Nvidia's data center business in order to be seen as a worthy investment.</p>
<div class="lb-trans"><p>我认为投资者没有给予 AMD 的 GPU 业务足够的认可，并将该公司视为落后于 Nvidia 的企业。但在我看来，AMD 不需要在数据中心业务上超越 Nvidia，才能被视为值得投资的公司。</p>
</div><p>Rather, I think that AMD's current trajectory and impressive customer wins could lead to a prolonged period of sustained accelerated growth -- underscored by the company's relentless pursuit of incremental market share in the GPU realm.</p>
<div class="lb-trans"><p>相反，我认为 AMD 当前的轨迹和令人印象深刻的客户胜利可能会导致持续加速增长的长期时期，这一切都得益于该公司在 GPU 领域对增量市场份额的不懈追求。</p>
</div><p>In other words, AMD may not become the No. 1 player in the chip space. But if the company can demonstrate its ability to continue winning large contracts and grow at a pace commensurate to or even faster than Nvidia (already a mature player in the GPU industry) then I could see growth investors flocking to AMD stock as an alternative to Nvidia.</p>
<div class="lb-trans"><p>换句话说，AMD 可能不会成为芯片领域的第一玩家。但如果该公司能够展示其继续赢得大型合同的能力，并以与 Nvidia（已经是 GPU 行业的成熟玩家）相当甚至更快的速度增长，那么我可以看到增长型投资者涌向 AMD 股票，作为 Nvidia 的替代品。</p>
</div><p>Despite an underwhelming performance over the last year, I think AMD stock could be on its way to witnessing a similar path to that of Nvidia following the initial jolt from the AI boom.</p>
<div class="lb-trans"><p>尽管在过去一年表现平平，我认为 AMD 股票可能正朝着与 Nvidia 在 AI 热潮初期相似的路径发展。</p>
</div>

Has AMD's "Nvidia Moment" Finally Arrived?

- LLMs 的崛起使得信息获取变得迅速，Nvidia 在 GPU 市场占据 90% 份额。  
- AMD 通过与 Oracle 等大客户签订合同，开始进入数据中心 GPU 市场。  
- AMD 股价下跌 47%，当前估值吸引投资者关注，未来有望实现增长。  

AMD 的 “英伟达时刻” 是否终于到来了？