[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-openai-jalapeno-threatens-nvidia-realistically-en":3,"article-related-openai-jalapeno-threatens-nvidia-realistically-en":31,"series-model-release-81c51b29-6f78-43bb-a264-e6b208644d4f":79},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":23,"views":27,"created_at":28,"published_at":29,"topic_cluster_id":30},"81c51b29-6f78-43bb-a264-e6b208644d4f","openai-jalapeno-threatens-nvidia-realistically-en","OpenAI自研芯片不是秀肌肉，而是对英伟达的真实威胁","\u003Cp data-speakable=\"summary\">\u003Ca href=\"\u002Ftag\u002Fopenai\">OpenAI\u003C\u002Fa>的首颗自研推理芯片Jalapeño说明，英伟达在AI基础设施上的定价权开始松动。\u003C\u002Fp>\u003Cp>我认为，OpenAI这颗名为Jalapeño的自研芯片，不是一次公关式展示，而是英伟达最该警惕的信号。它从零到流片只用了九个月，目标也很明确：专门为大模型推理打造一颗“Intelligence Processor”，把最贵、最耗电、最依赖供应链的一段算力链条，直接收回到自己手里。\u003C\u002Fp>\u003Ch2>第一层威胁不是性能，而是议价权\u003C\u002Fh2>\u003Cp>真正让英伟达难受的，不是OpenAI有没有做出一颗能跑模型的芯片，而是OpenAI开始拥有替代选项。只要推理工作负载的一部分从\u003Ca href=\"\u002Ftag\u002Fgpu\">GPU\u003C\u002Fa>迁移到自研ASIC，英伟达就不再是唯一答案。对一个每年要烧掉巨量推理成本的公司来说，哪怕只把一小部分流量切出去，都足以在采购谈判里改变桌上的筹码。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782793065127-zluz.png\" alt=\"OpenAI自研芯片不是秀肌肉，而是对英伟达的真实威胁\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>这类变化在云厂商身上已经演过一遍。\u003Ca href=\"\u002Ftag\u002Faws\">AWS\u003C\u002Fa>有Graviton，\u003Ca href=\"\u002Ftag\u002Fgoogle\">Google\u003C\u002Fa>有TPU，微软也在推进自研加速器，逻辑都一样：先从最稳定、最可预测的负载下手，再慢慢扩大适用范围。OpenAI现在做的不是“取代GPU”，而是把GPU从绝对必要变成可替代，这一步本身就足够危险。\u003C\u002Fp>\u003Ch2>第二层威胁是推理经济学，而不是训练竞赛\u003C\u002Fh2>\u003Cp>Jalapeño的重点是推理，不是训练，这一点非常关键。训练前沿模型需要极端灵活的并行能力和成熟的软件生态，GPU仍然占优；但推理更看重单位成本、吞吐、延迟和功耗。大模型真正的商业化压力，往往不是训练一次要多少钱，而是上线后每一次回答、每一次检索、每一次工具调用都在持续消耗算力。\u003C\u002Fp>\u003Cp>如果OpenAI能把推理成本压下来，它得到的不是技术新闻，而是产品利润。举个最简单的例子，面向海量用户的聊天、摘要、检索增强和代理调用，都是高频推理场景。谁能把每千次请求的成本降下去，谁就能把更低价格、更高毛利和更激进的产品策略同时拿到手。英伟达卖的是通用算力，OpenAI要的是把算力变成可控的产品成本。\u003C\u002Fp>\u003Ch2>第三层威胁是供应链控制权\u003C\u002Fh2>\u003Cp>九个月从白纸到流片，这个速度说明OpenAI已经不满足于“买现成的最强芯片”，而是在搭建自己的基础设施主权。对前沿模型公司来说，最大风险从来不只是性能落后，而是供货、配额、交付周期和地缘政治限制。自研芯片的意义之一，就是把关键能力从外部供应商的排期表里拿回来。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782793063626-lse6.png\" alt=\"OpenAI自研芯片不是秀肌肉，而是对英伟达的真实威胁\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>这件事在规模上会越来越重要。AI公司一旦进入大规模服务阶段，芯片不是一次性采购，而是持续扩容、持续替换、持续优化的资产。自研ASIC哪怕只覆盖一部分推理集群，也能让公司在产能紧张、出口限制和价格波动时保留缓冲。英伟达最怕的不是一个客户少买几块卡，而是大客户开始把未来算力规划写进自己的芯片路线图。\u003C\u002Fp>\u003Ch2>The counter-argument\u003C\u002Fh2>\u003Cp>反对者会说，OpenAI这件事被夸大了。ASIC确实能在特定任务上更便宜、更高效，但它也更窄、更难迭代。大模型系统变化太快，今天流行的是某种注意力结构，明天就可能换成别的推理范式。GPU的优势就在于通用性和软件生态，\u003Ca href=\"\u002Ftag\u002Fcuda\">CUDA\u003C\u002Fa>和成熟开发工具链不是一颗新芯片三个月、九个月就能补齐的。\u003C\u002Fp>\u003Cp>还有一个现实问题：做芯片不等于做成芯片生意。流片只是开始，真正难的是良率、封装、供电、散热、驱动、编译器和大规模部署。很多自研硬件项目都在“能跑”和“能规模化赚钱”之间折戟。站在这个角度看，Jalapeño更像是一枚试探性的棋子，而不是立即改写行业格局的终局武器。\u003C\u002Fp>\u003Cp>但这个反驳只成立一半。因为OpenAI并不需要用Jalapeño全面打败GPU，它只需要在推理这个最赚钱、最重复、最稳定的环节里拿到结构性优势。只要它证明自研芯片能降低成本、稳定供给并支撑真实业务，英伟达的护城河就会从“不可替代”变成“仍然强，但不再绝对”。这已经足够让市场重新定价。\u003C\u002Fp>\u003Ch2>What to do with this\u003C\u002Fh2>\u003Cp>如果你是工程师，别把这件事理解成“芯片新闻”，而要理解成系统设计信号：未来的大模型栈会越来越垂直整合，模型、编译器、推理引擎和硬件会一起优化。如果你是PM，优先盯住推理成本、延迟和单位请求毛利，因为真正决定产品能否扩张的，不是模型参数，而是每次调用的经济账。如果你是创始人，这条新闻的启示更直接：当你的核心业务足够大时，供应商不会永远是供应商，你必须尽早把关键依赖变成自己的能力。\u003C\u002Fp>","OpenAI的首颗自研推理芯片Jalapeño说明，英伟达在AI基础设施上的定价权开始松动。","zhuanlan.zhihu.com","https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F2053473940031460150",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782793065127-zluz.png","model-release","en","76ca309d-9b8f-4595-a732-8cdb801b25e1",[17,18,19,20,21,22],"OpenAI","Jalapeño","NVIDIA","推理芯片","ASIC","大模型基础设施",[24,25,26],"OpenAI自研Jalapeño的核心意义是削弱英伟达的议价权，而不只是追求性能突破。","这颗芯片瞄准推理而非训练，因为推理才是大模型商业化里最持续、最昂贵的成本中心。","九个月流片说明AI公司正在把算力主权内建到自己的路线图中，供应链控制权正在重塑行业格局。",0,"2026-06-30T04:17:22.110402+00:00","2026-06-30T04:17:22.096+00:00","1bae1133-d241-4581-9332-fbf39690c319",{"tags":32,"relatedLang":38,"relatedPosts":42},[33,35],{"name":17,"slug":34},"openai",{"name":36,"slug":37},"Nvidia","nvidia",{"id":15,"slug":39,"title":40,"language":41},"openai-jalapeno-threatens-nvidia-realistically-zh","OpenAI自研芯片不是秀肌肉，而是英伟达的真实威胁","zh",[43,49,55,61,67,73],{"id":44,"slug":45,"title":46,"cover_image":47,"image_url":47,"created_at":48,"category":13},"bbc7f86f-2952-4aec-a003-1885ba544a22","k3s-v1-34-9-kubernetes-1-34-9-release-en","K3s v1.34.9 lands with Kubernetes 1.34.9","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782781394063-6jum.png","2026-06-30T01:02:52.014221+00:00",{"id":50,"slug":51,"title":52,"cover_image":53,"image_url":53,"created_at":54,"category":13},"ab62b837-c8ac-493d-a35a-4c454402fd12","kimi-2-7-price-coding-benchmark-en","Kimi 2.7 makes price the real coding benchmark","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782746269451-4jtb.png","2026-06-29T15:17:24.882797+00:00",{"id":56,"slug":57,"title":58,"cover_image":59,"image_url":59,"created_at":60,"category":13},"2b2e09ae-d63f-4d0d-88c9-ca494fc7cc3b","kimi-k26-open-source-coding-agentic-ai-benchmarks-en","Kimi K2.6 tops coding and agentic AI benchmarks","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782739081936-jpdb.png","2026-06-29T13:17:26.953686+00:00",{"id":62,"slug":63,"title":64,"cover_image":65,"image_url":65,"created_at":66,"category":13},"666962b5-ce8c-430c-9d07-8cdfd44ffd09","llama-legends-380-season-3-heroes-raids-en","Llama Legends 3.8.0 adds Season 3 heroes and raids","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782711179242-ednu.png","2026-06-29T05:32:33.398141+00:00",{"id":68,"slug":69,"title":70,"cover_image":71,"image_url":71,"created_at":72,"category":13},"b4840252-4311-4c44-9814-4a3d1666302f","omlx-045-dev1-glm52-minimax-m3-speedups-en","oMLX 0.4.5.dev1 speeds up GLM-5.2 and MiniMax M3","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782709371396-mn9r.png","2026-06-29T05:02:28.770698+00:00",{"id":74,"slug":75,"title":76,"cover_image":77,"image_url":77,"created_at":78,"category":13},"1fe27411-ad64-4717-85c9-89b5c350253c","grok-45-private-beta-tesla-spacex-en","Grok 4.5 enters private beta at Tesla and SpaceX","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782687764199-vjto.png","2026-06-28T23:02:23.343104+00:00",[80,85,90,95,100,105,110,115,120,125],{"id":81,"slug":82,"title":83,"created_at":84},"d4cffde7-9b50-4cc7-bb68-8bc9e3b15477","nvidia-rubin-ai-supercomputer-en","NVIDIA Unveils Rubin: A Leap in AI Supercomputing","2026-03-25T16:24:35.155565+00:00",{"id":86,"slug":87,"title":88,"created_at":89},"eab919b9-fbac-4048-89fc-afad6749ccef","google-gemini-ai-innovations-2026-en","Google's AI Leap with Gemini Innovations in 2026","2026-03-25T16:27:18.841838+00:00",{"id":91,"slug":92,"title":93,"created_at":94},"5f5cfc67-3384-4816-a8f6-19e44d90113d","gap-google-gemini-ai-checkout-en","Gap Teams Up with Google Gemini for AI-Driven Checkout","2026-03-25T16:27:46.483272+00:00",{"id":96,"slug":97,"title":98,"created_at":99},"f6d04567-47f6-49ec-804c-52e61ab91225","ai-model-release-wave-march-2026-en","Navigating the AI Model Release Wave of March 2026","2026-03-25T16:28:45.409716+00:00",{"id":101,"slug":102,"title":103,"created_at":104},"895c150c-569e-4fdf-939d-dade785c990e","small-language-models-transform-ai-en","Small Language Models: Llama 3.2 and Phi-3 Transform AI","2026-03-25T16:30:26.688313+00:00",{"id":106,"slug":107,"title":108,"created_at":109},"38eb1d26-d961-4fd3-ae12-9c4089680f5f","midjourney-v8-alpha-features-pricing-en","Midjourney V8 Alpha: A Deep Dive into Its Features and Pricing","2026-03-26T01:25:36.387587+00:00",{"id":111,"slug":112,"title":113,"created_at":114},"bf36bb9e-3444-4fb8-ab19-0df6bc9d8271","rag-2026-indispensable-ai-bridge-en","RAG in 2026: The Indispensable AI Bridge","2026-03-26T01:28:34.472046+00:00",{"id":116,"slug":117,"title":118,"created_at":119},"60881d6d-2310-44ef-b1fb-7f98e9dd2f0e","xiaomi-mimo-trio-agents-robots-voice-en","Xiaomi’s MiMo trio targets agents, robots, and voice","2026-03-28T03:05:08.899895+00:00",{"id":121,"slug":122,"title":123,"created_at":124},"f063d8d1-41d1-4de4-8ebc-6c40511b9369","xiaomi-mimo-v2-pro-1t-moe-agents-en","Xiaomi MiMo-V2-Pro: 1T MoE Model for Agents","2026-03-28T03:06:19.238032+00:00",{"id":126,"slug":127,"title":128,"created_at":129},"a1379e9a-6785-4ff5-9b0a-8cff55f8264f","cursor-composer-2-started-from-kimi-en","Cursor’s Composer 2 started from Kimi","2026-03-28T03:11:59.132398+00:00"]