<p>今年表现火爆的一只人工智能（AI）股票是社交媒体巨头 <strong>Meta Platforms</strong>(META 1.11%)。就在几周前，Meta 的领导层告诉投资者，2025 年将是一个充满活动的一年，因为公司希望在其 AI 路线图上继续积极推进。</p>
<p>具体来说，Meta 计划在今年投入高达 650 亿美元用于 AI 基础设施，这比去年增长了 65%。考虑到该公司与 <strong>Nvidia</strong> 和 <strong>Advanced Micro Devices</strong> 密切合作，并且还与 <strong>Broadcom</strong> 合作设计自己的定制推理芯片，Meta 计划如此大幅增加资本支出（capex）并不令人惊讶。</p>
<p>不过，令人惊讶的是，Meta 可能刚刚找到了一种新的资本来源——这可能会让 Nvidia 陷入困境。接下来，我将深入探讨一些与 Meta 相关的最新消息，并详细说明 Nvidia 可能受到的影响。</p>
<h2>Meta 可能如何超越 Nvidia</h2>
<p>尽管 Nvidia 在图形处理单元（GPU）市场占据主导地位，但该公司面临的竞争不仅限于 AMD 等老对手。具体来说，许多 Nvidia 最大的客户，如云计算巨头 <strong>Microsoft</strong>、<strong>Amazon</strong> 和 <strong>Alphabet</strong>，以及社交网络平台 Meta，正在投资自己的定制芯片产品。这些投资的理由是为了用新架构补充现有芯片组，以减少对 Nvidia 产品的完全依赖。</p>
<p>Meta 在 GPU 领域的探索围绕其 Meta 训练和推理加速器（MTIA）芯片展开。虽然 MTIA 芯片在 Meta 仍处于初期阶段，但该公司可能已经找到了比最初预期更快推进该项目的方法。</p>
<p>多家媒体报道，Meta 正在考虑收购来自韩国的芯片初创公司 FuriosaAI。FuriosaAI 声称其处理能力比 Nvidia 最新的 GPU 更具成本效益。如果这是真的，这可能会改变 Meta 在今年的资本支出计划以及其对更多数据中心芯片的关注。</p>
<p>理论上，FuriosaAI 带来的成本效益以及该公司的芯片开发流程可能为 Meta 的 AI 预算和产品路线图带来明显的协同效应，因为 MTIA 芯片的开发仍在继续。</p>
<p><img src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MDc2MTQvZ2V0dHlpbWFnZXMtMTc3MzU0MzgzMC0zLmpwZyZ3PTcwMA?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-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MDc2MTQvZ2V0dHlpbWFnZXMtMTc3MzU0MzgzMC0zLmpwZyZ3PTcwMA"/></p>
<p>图片来源：Getty Images。</p>
<h2>Meta 的交易是否会给 Nvidia 带来麻烦？</h2>
<p>从表面上看，MTIA 芯片的推出可能看起来像是 Nvidia 的一场危机。</p>
<p>然而，Meta 的芯片仅仅约一岁——因此不太可能在短短一年内开发出与 Nvidia 相媲美或更复杂的产品。</p>
<p>出于这个原因，我不会说 Nvidia 很快就会失去 Meta 作为客户。此外，目前关于 Meta 对 FuriosaAI 感兴趣的消息仍然只是传闻。</p>
<p>此外，即使 Meta 确实推进收购 FuriosaAI，整合和产品开发方面还有很多工作要做。将 FuriosaAI 的技术与 Meta 的 AI 相结合，对 Nvidia 来说并不一定是坏事。如果说有什么好处，增加的竞争可能会导致 GPU 领域更多的创新和尖端架构。</p>
<p>我认为半导体行业开始整合只是时间问题。许多 Nvidia 最大的客户已经开始关注自己的内部芯片产品。考虑到这一点，我不会感到惊讶，看到大型科技公司在未来几年开始收购或与 Nvidia 生态系统之外的其他芯片制造商合作。</p>
<p>归根结底，我并不认为 Meta 的潜在收购会对 Nvidia 构成致命打击。除非达成交易，否则 Nvidia 的股票因 Meta 的动机而产生的任何波动都根植于情绪驱动的投机。</p>

META

<p>Meta Platforms 计划在 2025 年投资 650 亿美元用于人工智能基础设施，比去年增加 65%。该公司正在探索收购芯片初创公司 FuriosaAI，以增强其定制芯片的开发，这可能会影响 Nvidia 的市场地位。然而，MTIA 芯片仍处于早期开发阶段，任何收购仍然是推测性的。虽然这可能引入竞争，但也可能促进 GPU 领域的创新，而不是对 Nvidia 的主导地位构成直接威胁</p>

<p>One artificial intelligence (AI) stock that got off to a scorching start this year is social media juggernaut <strong>Meta Platforms</strong>(META 1.11%). Just a couple of weeks ago, Meta's leadership told investors that 2025 is going to be a jam-packed year as the company seeks to continue moving aggressively on its AI roadmap.</p>
<div class="lb-trans"><p>今年表现火爆的一只人工智能（AI）股票是社交媒体巨头 <strong>Meta Platforms</strong>(META 1.11%)。就在几周前，Meta 的领导层告诉投资者，2025 年将是一个充满活动的一年，因为公司希望在其 AI 路线图上继续积极推进。</p>
</div><p>To be specific, Meta is preparing to spend up to $65 billion in AI infrastructure this year -- representing a 65% increase year over year. Given the company works closely with <strong>Nvidia</strong> and <strong>Advanced Micro Devices</strong>, and is also working with <strong>Broadcom</strong> to design its own custom inferencing chips, it's not too surprising that Meta is planning for such a robust increase in capital expenditures (capex).</p>
<div class="lb-trans"><p>具体来说，Meta 计划在今年投入高达 650 亿美元用于 AI 基础设施，这比去年增长了 65%。考虑到该公司与 <strong>Nvidia</strong> 和 <strong>Advanced Micro Devices</strong> 密切合作，并且还与 <strong>Broadcom</strong> 合作设计自己的定制推理芯片，Meta 计划如此大幅增加资本支出（capex）并不令人惊讶。</p>
</div><p>What may be surprising, though, is that Meta may have just found a new source for its capital -- and it could put Nvidia in a jam. Below, I'll dig into some recent news involving Meta right now and detail how Nvidia could be impacted.</p>
<div class="lb-trans"><p>不过，令人惊讶的是，Meta 可能刚刚找到了一种新的资本来源——这可能会让 Nvidia 陷入困境。接下来，我将深入探讨一些与 Meta 相关的最新消息，并详细说明 Nvidia 可能受到的影响。</p>
</div><h2>How Meta might be trying to leapfrog Nvidia</h2>
<div class="lb-trans"><h2>Meta 可能如何超越 Nvidia</h2>
</div><p>Even though Nvidia dominates the graphics processing unit (GPU) market, the company has been facing some competition beyond the usual suspect of AMD. Namely, many of Nvidia's largest customers, such as cloud hyperscalers <strong>Microsoft</strong>, <strong>Amazon</strong>, and <strong>Alphabet</strong>, as well as social networking platform Meta, are investing in their own custom chip products. The rationale behind these investments is to complement existing chipsets with new architectures in an effort to migrate away from a full reliance on Nvidia's stack.</p>
<div class="lb-trans"><p>尽管 Nvidia 在图形处理单元（GPU）市场占据主导地位，但该公司面临的竞争不仅限于 AMD 等老对手。具体来说，许多 Nvidia 最大的客户，如云计算巨头 <strong>Microsoft</strong>、<strong>Amazon</strong> 和 <strong>Alphabet</strong>，以及社交网络平台 Meta，正在投资自己的定制芯片产品。这些投资的理由是为了用新架构补充现有芯片组，以减少对 Nvidia 产品的完全依赖。</p>
</div><p>Meta's foray into the GPU realm revolves around its Meta Training and Inference Accelerator (MTIA) chip. While the MTIA chip is still a nascent initiative at Meta, the company may have identified a way to push the project forward much faster than initially anticipated.</p>
<div class="lb-trans"><p>Meta 在 GPU 领域的探索围绕其 Meta 训练和推理加速器（MTIA）芯片展开。虽然 MTIA 芯片在 Meta 仍处于初期阶段，但该公司可能已经找到了比最初预期更快推进该项目的方法。</p>
</div><p>Several media outlets are reporting that Meta is considering acquiring FuriosaAI, a chip start-up out of Korea. FuriosaAI claims that its processing power is more cost efficient compared to Nvidia's newest GPUs. If this is true, that could change the game for Meta as it pertains to its capex plans this year and its focus on acquiring more data center chips.</p>
<div class="lb-trans"><p>多家媒体报道，Meta 正在考虑收购来自韩国的芯片初创公司 FuriosaAI。FuriosaAI 声称其处理能力比 Nvidia 最新的 GPU 更具成本效益。如果这是真的，这可能会改变 Meta 在今年的资本支出计划以及其对更多数据中心芯片的关注。</p>
</div><p>In theory, the cost efficiencies brought from FuriosaAI as well as the company's chip development processes could bring some obvious synergies to Meta's AI budget and its product roadmap as development for the MTIA chips continues.</p>
<div class="lb-trans"><p>理论上，FuriosaAI 带来的成本效益以及该公司的芯片开发流程可能为 Meta 的 AI 预算和产品路线图带来明显的协同效应，因为 MTIA 芯片的开发仍在继续。</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MDc2MTQvZ2V0dHlpbWFnZXMtMTc3MzU0MzgzMC0zLmpwZyZ3PTcwMA" alt="" original-src="https://g.foolcdn.com/image/?url=https%3A%2F%2Fg.foolcdn.com%2Feditorial%2Fimages%2F807614%2Fgettyimages-1773543830-3.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>Could Meta's deal spell trouble for Nvidia?</h2>
<div class="lb-trans"><h2>Meta 的交易是否会给 Nvidia 带来麻烦？</h2>
</div><p>On the surface, the introduction of the MTIA chip might seem like a crisis for Nvidia.</p>
<div class="lb-trans"><p>从表面上看，MTIA 芯片的推出可能看起来像是 Nvidia 的一场危机。</p>
</div><p>However, Meta's chip is only about a year old -- so it's unlikely that the company developed something as sophisticated or more when compared to Nvidia in just one year.</p>
<div class="lb-trans"><p>然而，Meta 的芯片仅仅约一岁——因此不太可能在短短一年内开发出与 Nvidia 相媲美或更复杂的产品。</p>
</div><p>For this reason, I wouldn't say that Nvidia is at risk of losing Meta as a customer soon. Moreover, the news around Meta's interest in FuriosaAI is, for now, still a rumor.</p>
<div class="lb-trans"><p>出于这个原因，我不会说 Nvidia 很快就会失去 Meta 作为客户。此外，目前关于 Meta 对 FuriosaAI 感兴趣的消息仍然只是传闻。</p>
</div><p>Furthermore, even if Meta does move forward with acquiring FuriosaAI, there will be a lot of work to do on the integration and product development sides of the house. Augmenting FuriosaAI's technology with Meta's AI isn't necessarily a bad thing for Nvidia. If anything, the added competition could lead to more innovation and cutting-edge architectures in the GPU landscape.</p>
<div class="lb-trans"><p>此外，即使 Meta 确实推进收购 FuriosaAI，整合和产品开发方面还有很多工作要做。将 FuriosaAI 的技术与 Meta 的 AI 相结合，对 Nvidia 来说并不一定是坏事。如果说有什么好处，增加的竞争可能会导致 GPU 领域更多的创新和尖端架构。</p>
</div><p>I think it's a matter of time before the semiconductor sector begins to consolidate anyway. Many of Nvidia's largest customers are already focusing on their own in-house chipware. With that in mind, I wouldn't be surprised to see big tech start acquiring or partnering with other chipmakers outside of the Nvidia ecosystem over the next few years.</p>
<div class="lb-trans"><p>我认为半导体行业开始整合只是时间问题。许多 Nvidia 最大的客户已经开始关注自己的内部芯片产品。考虑到这一点，我不会感到惊讶，看到大型科技公司在未来几年开始收购或与 Nvidia 生态系统之外的其他芯片制造商合作。</p>
</div><p>At the end of the day, I do not see a potential acquisition by Meta as a checkmate move against Nvidia. Unless and until a deal is formed, any movement in Nvidia's stock based on Meta's motives is rooted in emotion-driven speculation.</p>
<div class="lb-trans"><p>归根结底，我并不认为 Meta 的潜在收购会对 Nvidia 构成致命打击。除非达成交易，否则 Nvidia 的股票因 Meta 的动机而产生的任何波动都根植于情绪驱动的投机。</p>
</div>

Did Meta Platforms Just Say "Checkmate" to Nvidia?

- Meta 计划在 2025 年投入 65 亿美元用于 AI 基础设施，增幅达 65%。  
- Meta 可能收购 FuriosaAI，以提高其 MTIA 芯片的成本效益。  
- 此举可能对 Nvidia 构成竞争压力，但短期内 Nvidia 仍是 Meta 的主要客户。

Meta Platforms 刚刚对 Nvidia 说 “将死” 了吗？