520 lines
35 KiB
HTML
520 lines
35 KiB
HTML
<!DOCTYPE html>
|
||
<html lang="zh-CN" data-theme="night">
|
||
<head>
|
||
<meta charset="UTF-8">
|
||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||
<title>AI芯片 - 深度研究报告</title>
|
||
<script src="https://cdn.tailwindcss.com"></script>
|
||
<link href="https://cdn.jsdelivr.net/npm/daisyui@4.11.1/dist/full.min.css" rel="stylesheet" type="text/css" />
|
||
<script defer src="https://cdn.jsdelivr.net/npm/alpinejs@3.x.x/dist/cdn.min.js"></script>
|
||
<script src="https://cdn.jsdelivr.net/npm/echarts@5.5.0/dist/echarts.min.js"></script>
|
||
<style>
|
||
@import url('https://fonts.googleapis.com/css2?family=Exo+2:wght@300;400;500;700&display=swap');
|
||
|
||
body {
|
||
font-family: 'Exo 2', sans-serif;
|
||
background-color: #000213;
|
||
background-image:
|
||
radial-gradient(ellipse 50% 80% at 20% 40%, rgba(100, 10, 150, 0.3), transparent),
|
||
radial-gradient(ellipse 50% 80% at 80% 50%, rgba(10, 100, 150, 0.3), transparent);
|
||
background-attachment: fixed;
|
||
color: #e0e0e0;
|
||
}
|
||
|
||
.glass-card {
|
||
background: rgba(10, 10, 30, 0.4);
|
||
backdrop-filter: blur(16px) saturate(180%);
|
||
-webkit-backdrop-filter: blur(16px) saturate(180%);
|
||
border: 1px solid rgba(255, 255, 255, 0.1);
|
||
border-radius: 24px;
|
||
transition: all 0.3s ease;
|
||
}
|
||
|
||
.glass-card:hover {
|
||
transform: translateY(-5px);
|
||
box-shadow: 0 0 30px rgba(0, 191, 255, 0.2);
|
||
border-color: rgba(0, 191, 255, 0.3);
|
||
}
|
||
|
||
.bento-grid {
|
||
display: grid;
|
||
gap: 1.5rem;
|
||
}
|
||
|
||
.glow-text {
|
||
text-shadow: 0 0 8px rgba(0, 191, 255, 0.6), 0 0 12px rgba(0, 191, 255, 0.4);
|
||
}
|
||
|
||
.title-line {
|
||
position: relative;
|
||
padding-bottom: 0.5rem;
|
||
margin-bottom: 1rem;
|
||
}
|
||
|
||
.title-line::after {
|
||
content: '';
|
||
position: absolute;
|
||
bottom: 0;
|
||
left: 0;
|
||
width: 50px;
|
||
height: 3px;
|
||
background: linear-gradient(90deg, #00BFFF, #8A2BE2);
|
||
border-radius: 2px;
|
||
box-shadow: 0 0 8px #00BFFF;
|
||
}
|
||
|
||
.prose-custom {
|
||
color: #c0c0c0;
|
||
}
|
||
.prose-custom h1, .prose-custom h2, .prose-custom h3, .prose-custom h4, .prose-custom h5, .prose-custom h6 {
|
||
color: #ffffff;
|
||
}
|
||
.prose-custom strong {
|
||
color: #87CEEB; /* Sky Blue */
|
||
}
|
||
.prose-custom a {
|
||
color: #4682B4; /* Steel Blue */
|
||
}
|
||
.prose-custom blockquote {
|
||
border-left-color: #4682B4;
|
||
color: #a0a0a0;
|
||
}
|
||
.prose-custom code {
|
||
color: #DA70D6; /* Orchid */
|
||
}
|
||
</style>
|
||
</head>
|
||
<body class="min-h-screen">
|
||
<div class="container mx-auto p-4 sm:p-6 lg:p-8">
|
||
|
||
<header class="text-center py-12">
|
||
<h1 class="text-4xl md:text-6xl font-bold glow-text tracking-widest">AI 芯片</h1>
|
||
<h2 class="text-2xl md:text-3xl font-light text-cyan-300 mt-2">深度研究报告</h2>
|
||
<p class="mt-4 text-sm text-gray-500 max-w-2xl mx-auto">北京价值前沿科技有限公司 AI投研agent：“价小前投研” 进行投研呈现，本报告为AI合成数据，投资需谨慎。</p>
|
||
</header>
|
||
|
||
<main class="space-y-12">
|
||
<!-- Insight Section -->
|
||
<section id="insight-analysis" class="glass-card p-6 md:p-8">
|
||
<h2 class="text-3xl font-bold mb-6 title-line glow-text">核心洞察 (Insight)</h2>
|
||
|
||
<div class="space-y-8 prose-custom max-w-none">
|
||
<div>
|
||
<h3 class="text-xl font-semibold text-sky-300">0. 概念事件时间轴</h3>
|
||
<p>AI芯片概念的核心叙事围绕着全球AI算力需求的爆发与中美科技博弈下的国产替代加速。</p>
|
||
<div id="timeline-chart" class="w-full h-96 mt-4 glass-card p-2"></div>
|
||
</div>
|
||
|
||
<div class="glass-card p-6 rounded-2xl">
|
||
<h3 class="text-xl font-semibold text-sky-300">1. 核心观点摘要</h3>
|
||
<p>AI芯片概念正处于由<strong>地缘政治驱动的强制性国产替代</strong>与<strong>AI应用爆发驱动的内生性需求增长</strong>双轮驱动的关键阶段。其核心逻辑在于，外部封锁为国产厂商创造了前所未有的市场准入机会，但技术性能、集群能力和软件生态仍是决定其能否将“窗口期”转化为长期市场份额的关键。未来，市场将从“有没有”的主题炒作，逐步过渡到“好不好用”的基本面验证，推理场景和信创领域将是国产芯片最先取得突破的阵地。</p>
|
||
</div>
|
||
|
||
<div class="bento-grid grid-cols-1 lg:grid-cols-2">
|
||
<div class="glass-card p-6 rounded-2xl lg:col-span-2">
|
||
<h3 class="text-xl font-semibold text-sky-300">2. 核心逻辑与市场认知分析</h3>
|
||
<div class="mt-4 space-y-4">
|
||
<div>
|
||
<h4 class="font-bold text-lg text-white">核心驱动力:</h4>
|
||
<ul class="list-disc list-inside space-y-1 mt-2">
|
||
<li><strong>政策与地缘政治（首要驱动力）</strong>：美国持续升级的出口管制是引爆国产AI芯片概念的根本原因，为国产芯片清除了最强大的竞争对手，创造了巨大的、受保护的国内市场。</li>
|
||
<li><strong>内生需求爆发</strong>：大模型应用蓬勃发展导致算力需求井喷。国内市场规模预估达<strong>1000-1500亿</strong>，为国产芯片提供了明确的商业化场景。</li>
|
||
<li><strong>成本与供应链安全</strong>：对于国内云厂商和智算中心，自研或采用国产芯片是实现“降本增效”和“保障供应链安全”的必然选择。</li>
|
||
</ul>
|
||
</div>
|
||
<div>
|
||
<h4 class="font-bold text-lg text-white">市场热度与情绪:</h4>
|
||
<p>当前市场对AI芯片概念抱有<strong>极高的关注度和乐观预期</strong>，但夹杂着对技术现实的<strong>潜在担忧</strong>。市场情绪整体偏向乐观，聚焦于“国产替代”的宏大叙事，但细节也透露出谨慎，呈现出“短期看题材、长期看落地”的分化特征。</p>
|
||
</div>
|
||
<div>
|
||
<h4 class="font-bold text-lg text-white">预期差分析:</h4>
|
||
<ul class="list-disc list-inside space-y-2 mt-2">
|
||
<li><strong>训练 vs. 推理</strong>: 市场可能高估了国产芯片在短期内于<strong>训练市场</strong>的替代能力，而低估了其在<strong>推理市场</strong>加速渗透的确定性。现实格局是“<strong>训练靠国外，推理看国产</strong>”。</li>
|
||
<li><strong>订单传闻 vs. 实际落地</strong>: 市场情绪易被未经证实的传闻驱动（如字节给寒武纪下单16万颗），而产业实际采购决策更为审慎，需要经过严格的验证周期。真实订单节奏可能不及最乐观预期。</li>
|
||
<li><strong>单点性能 vs. 集群能力</strong>: 真正的瓶颈在于大规模<strong>集群</strong>的通信效率、稳定性和软件协同能力。市场的关注点应从单卡性能转向大规模集群的“可用性”和“易用性”。</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
|
||
<div class="glass-card p-6 rounded-2xl">
|
||
<h3 class="text-xl font-semibold text-sky-300">3. 关键催化剂与未来发展路径</h3>
|
||
<h4 class="font-bold text-lg text-white mt-4">近期催化剂（未来3-6个月）:</h4>
|
||
<ul class="list-disc list-inside space-y-1 mt-2">
|
||
<li><strong>美国政策明确化</strong>：关于H20芯片是否被正式禁运将是关键变量。</li>
|
||
<li><strong>标志性国产订单落地</strong>：头部互联网公司或运营商公布大规模集采结果。</li>
|
||
<li><strong>国产芯片新品发布</strong>：寒武纪690等新一代芯片的发布及评测。</li>
|
||
</ul>
|
||
<h4 class="font-bold text-lg text-white mt-4">长期发展路径:</h4>
|
||
<ol class="list-decimal list-inside space-y-1 mt-2">
|
||
<li><strong>阶段一 (当前-2026)</strong>: 推理场景突破与生态初步构建。</li>
|
||
<li><strong>阶段二 (2026-2028)</strong>: 中小规模训练集群的可用性验证。</li>
|
||
<li><strong>阶段三 (2028以后)</strong>: 向万卡级别训练集群和开放生态演进。</li>
|
||
</ol>
|
||
</div>
|
||
|
||
<div class="glass-card p-6 rounded-2xl">
|
||
<h3 class="text-xl font-semibold text-sky-300">5. 潜在风险与挑战</h3>
|
||
<ul class="list-disc list-inside space-y-2 mt-4">
|
||
<li><strong>技术风险</strong>: 性能与良率（代工能力差距）、集群能力瓶颈、CUDA生态壁垒。</li>
|
||
<li><strong>商业化风险</strong>: 客户验证周期长、价格竞争可能导致毛利率承压。</li>
|
||
<li><strong>政策与竞争风险</strong>: 美国政策摇摆是最大外部风险、国内众多玩家同台竞技，竞争激烈。</li>
|
||
<li><strong>信息交叉验证风险</strong>: 订单传闻不一致，政策口径模糊，需警惕基于单一传闻的决策风险。</li>
|
||
</ul>
|
||
</div>
|
||
|
||
<div class="glass-card p-6 rounded-2xl lg:col-span-2">
|
||
<h3 class="text-xl font-semibold text-sky-300">4. 产业链与核心公司深度剖析</h3>
|
||
<div class="overflow-x-auto mt-4">
|
||
<table class="table table-zebra table-sm">
|
||
<thead class="bg-slate-800 text-white">
|
||
<tr>
|
||
<th>公司</th>
|
||
<th>竞争优势</th>
|
||
<th>业务进展 / 风险</th>
|
||
<th>定位</th>
|
||
</tr>
|
||
</thead>
|
||
<tbody>
|
||
<tr>
|
||
<td class="font-bold">华为昇腾</td>
|
||
<td>全栈能力 (芯片到软件)，集群部署优势，市场份额领先 (约40%)</td>
|
||
<td>910C对标H系列，已实现4000卡集群。风险：稳定性不足，人力成本ROI偏低</td>
|
||
<td class="text-cyan-400">领导者</td>
|
||
</tr>
|
||
<tr>
|
||
<td class="font-bold">海光信息 (DCU)</td>
|
||
<td>GPGPU架构，兼容x86，信创根基深厚，性价比高，商务能力强</td>
|
||
<td>获5万颗订单，可做4000卡集群。风险：深度学习性能相对一般，生态待建</td>
|
||
<td class="text-green-400">核心追赶者</td>
|
||
</tr>
|
||
<tr>
|
||
<td class="font-bold">寒武纪 (MLU)</td>
|
||
<td>纯粹AI芯片设计商，软件优化能力强，产品线覆盖云边端</td>
|
||
<td>MLU590已商业化，下一代690备受期待。风险：订单不确定性，持续亏损</td>
|
||
<td class="text-yellow-400">技术型追赶者</td>
|
||
</tr>
|
||
<tr>
|
||
<td class="font-bold">百度昆仑芯</td>
|
||
<td>生态底蕴深厚，软硬件协同强，硬件架构近GPU，易用性好</td>
|
||
<td>获中移动大单，2025年销售目标20亿已基本达成。风险：通信性能待优化</td>
|
||
<td class="text-purple-400">生态型玩家</td>
|
||
</tr>
|
||
</tbody>
|
||
</table>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
|
||
<div class="glass-card p-6 rounded-2xl">
|
||
<h3 class="text-xl font-semibold text-sky-300">6. 综合结论与投资启示</h3>
|
||
<p><strong>最终看法</strong>：AI芯片概念已进入<strong>“政策护航下的基本面验证初期”</strong>。投资逻辑的重心应从“替代的可能性”转向“落地的确定性”和“进展的真实性”。</p>
|
||
<h4 class="font-bold text-lg text-white mt-4">最具投资价值的细分环节:</h4>
|
||
<ol class="list-decimal list-inside space-y-1 mt-2">
|
||
<li><strong>头部AI芯片厂商</strong>：海光信息（确定性高），寒武纪（弹性大）。</li>
|
||
<li><strong>ASIC/芯片定制服务</strong>：芯原股份等“卖水者”，受益于整个设计生态的繁荣。</li>
|
||
<li><strong>配套产业链“增量”环节</strong>：电源管理（杰华特）、散热、高速连接等，因国产芯片功耗更高而迎来确定性价值量提升。</li>
|
||
</ol>
|
||
<h4 class="font-bold text-lg text-white mt-4">需重点跟踪和验证的关键指标:</h4>
|
||
<ul class="list-disc list-inside space-y-1 mt-2">
|
||
<li>下游客户的<strong>实际中标份额和金额</strong>。</li>
|
||
<li>大规模集群的<strong>第三方性能评测报告</strong>。</li>
|
||
<li>核心芯片公司的<strong>收入确认速度和毛利率变化</strong>。</li>
|
||
<li>美国出口管制政策的<strong>官方文件</strong>。</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
</section>
|
||
|
||
<!-- Supporting Data Section -->
|
||
<section id="supporting-data" class="glass-card p-6 md:p-8" x-data="{ tab: 'news' }">
|
||
<h2 class="text-3xl font-bold mb-6 title-line glow-text">多维数据透视</h2>
|
||
<div class="tabs tabs-boxed bg-slate-900/50">
|
||
<a class="tab tab-lg" :class="{ 'tab-active': tab === 'news' }" @click="tab = 'news'">市场动态 (新闻)</a>
|
||
<a class="tab tab-lg" :class="{ 'tab-active': tab === 'roadshow' }" @click="tab = 'roadshow'">产业洞察 (路演)</a>
|
||
<a class="tab tab-lg" :class="{ 'tab-active': tab === 'report' }" @click="tab = 'report'">技术前沿 (研报)</a>
|
||
</div>
|
||
|
||
<div class="mt-6 prose-custom max-w-none space-y-4">
|
||
<!-- News Tab -->
|
||
<div x-show="tab === 'news'">
|
||
<div class="collapse collapse-plus glass-card">
|
||
<input type="radio" name="news-accordion" checked="checked" />
|
||
<div class="collapse-title text-xl font-medium text-white">宏观市场与政策环境</div>
|
||
<div class="collapse-content">
|
||
<ul class="list-disc list-inside">
|
||
<li><strong>美国出口管制</strong>: 传闻因能耗比问题限制英伟达H20进口 (90%概率为真)，若属实将重大利好国产芯片。同时，拜登政府计划在离任前实施新一轮分级(Tiered)出口管制，自主可控迫在眉睫。</li>
|
||
<li><strong>国产芯片机遇</strong>: 国内AI芯片市场空间预估达1000-1500亿。目前存在技术差距，特别是“训练环节无法使用”，呈现“训练靠国外，推理看国产”格局。</li>
|
||
<li><strong>地方政策支持</strong>: 深圳、苏州等地相继出台政策，加大对机器人AI芯片、GPU、ASIC等方向的攻关支持。</li>
|
||
</ul>
|
||
</div>
|
||
</div>
|
||
<div class="collapse collapse-plus glass-card">
|
||
<input type="radio" name="news-accordion" />
|
||
<div class="collapse-title text-xl font-medium text-white">国产AI芯片厂商动态</div>
|
||
<div class="collapse-content">
|
||
<p><strong>寒武纪</strong>: 690芯片采用双Die设计，预计2025年底推出，软件优化能力是优势。</p>
|
||
<p><strong>昆仑芯</strong>: 生态底蕴深厚，易用性强，下一代产品2026年发布。</p>
|
||
<p><strong>海光信息</strong>: DCU聚焦信创与HPC，可做4000卡集群，性价比高。</p>
|
||
<p><strong>华为(昇腾)</strong>: 提供完整数据中心解决方案，多机通信有优势，可做4000卡集群。</p>
|
||
</div>
|
||
</div>
|
||
<div class="collapse collapse-plus glass-card">
|
||
<input type="radio" name="news-accordion" />
|
||
<div class="collapse-title text-xl font-medium text-white">国际厂商与采购动态</div>
|
||
<div class="collapse-content">
|
||
<p><strong>英伟达</strong>: H100遇冷，对华特供的H20价格下调。字节跳动已下单56万颗H20。</p>
|
||
<p><strong>云巨头自研</strong>: 亚马逊发布Trainium3，微软调整自研路线图并与英特尔合作，谷歌与Meta商讨供应定制芯片。全球AI芯片市场呈现“一超多强+区域替代”格局。</p>
|
||
<p><strong>字节跳动</strong>: 预计2025年采购30万片，会额外采购10-20万张国产卡（海光、沐曦、昆仑芯为主）。</p>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
|
||
<!-- Roadshow Tab -->
|
||
<div x-show="tab === 'roadshow'">
|
||
<div class="collapse collapse-plus glass-card">
|
||
<input type="radio" name="roadshow-accordion" checked="checked" />
|
||
<div class="collapse-title text-xl font-medium text-white">市场规模与格局</div>
|
||
<div class="collapse-content">
|
||
<p>全球AI芯片市场规模预计从2023年的406亿美金增长至2030年的1666亿美金。英伟达占据数据中心GPU市场92%份额，毛利率高达70%+。云厂商自研ASIC是重要趋势，未来份额将提升。</p>
|
||
</div>
|
||
</div>
|
||
<div class="collapse collapse-plus glass-card">
|
||
<input type="radio" name="roadshow-accordion" />
|
||
<div class="collapse-title text-xl font-medium text-white">国产替代与挑战</div>
|
||
<div class="collapse-content">
|
||
<p>美国出口限制为国产厂商（华为、寒武纪、海光）提供替代机会。华为昇腾910B算力接近A100，占据国内AI服务器芯片约40%份额。国产芯片主要挑战在于生态系统、数据交换带宽和先进制程代工能力。</p>
|
||
</div>
|
||
</div>
|
||
<div class="collapse collapse-plus glass-card">
|
||
<input type="radio" name="roadshow-accordion" />
|
||
<div class="collapse-title text-xl font-medium text-white">大厂自研动态</div>
|
||
<div class="collapse-content">
|
||
<p><strong>海外</strong>: Google (TPU v5p), Amazon (Trainium), Microsoft (Maia100), Meta (MTIA) 均在推进自研芯片，以实现降本增效和供应链安全。博通是谷歌、Meta等的核心ASIC设计服务商。</p>
|
||
<p><strong>国内</strong>: 百度（昆仑芯三代对标H100）、阿里（含光900）、腾讯（投资燧原）也在布局。主要瓶颈是7nm以下先进制程获取困难。</p>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
|
||
<!-- Research Report Tab -->
|
||
<div x-show="tab === 'report'">
|
||
<div class="bento-grid grid-cols-1 md:grid-cols-2">
|
||
<div class="glass-card p-4">
|
||
<h4 class="font-bold text-lg text-white">端侧AI芯片</h4>
|
||
<p>炬芯科技、瑞芯微等是国内AIoT SoC领军企业。端侧AI推动NPU、AI-ISP等技术集成，应用场景覆盖智能座舱、机器人、智能穿戴等。瑞芯微下一代RK3688 NPU算力将达16TOPS。</p>
|
||
</div>
|
||
<div class="glass-card p-4">
|
||
<h4 class="font-bold text-lg text-white">AI手机芯片</h4>
|
||
<p>有望成为最大端侧芯片市场，渗透率预计从2024年的16%增至2028年的54%。高通骁龙8Gen3、联发科天玑9300、苹果M4等芯片均内置强大NPU，专为加速端侧生成式AI任务。</p>
|
||
</div>
|
||
<div class="glass-card p-4 md:col-span-2">
|
||
<h4 class="font-bold text-lg text-white">市场趋势：从训练到推理</h4>
|
||
<p>AI训练成本正急剧下降，市场焦点正从训练转向推理。英伟达CEO黄仁勋表示“未来推理市场将会是现在训练市场的百万倍”。IDC预计2028年中国AI服务器用于推理的负载占比将达73%，这为国产芯片提供了更广阔的替代空间。</p>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
</section>
|
||
|
||
<!-- Stocks Section -->
|
||
<section id="stock-data" class="glass-card p-6 md:p-8">
|
||
<h2 class="text-3xl font-bold mb-6 title-line glow-text">相关概念股</h2>
|
||
|
||
<h3 class="text-2xl font-semibold mb-4 text-sky-300">AI芯片核心标的</h3>
|
||
<div class="overflow-x-auto">
|
||
<table class="table w-full">
|
||
<thead>
|
||
<tr class="text-white">
|
||
<th>股票名称</th>
|
||
<th>股票代码</th>
|
||
<th>核心逻辑</th>
|
||
<th>标签</th>
|
||
</tr>
|
||
</thead>
|
||
<tbody>
|
||
<!-- GPU -->
|
||
<tr class="hover">
|
||
<td>景嘉微</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=300474" target="_blank" class="link link-hover">300474</a></td>
|
||
<td>国内专用GPU龙头，核心产品JM11系列芯片及景宏系列AI算力产品</td>
|
||
<td><div class="badge badge-info">GPU</div></td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>海光信息</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=688041" target="_blank" class="link link-hover">688041</a></td>
|
||
<td>DCU系列基于GPGPU架构，技术国内领先；CPU兼容x86指令集</td>
|
||
<td><div class="badge badge-info">GPU</div><div class="badge badge-accent ml-1">CPU</div></td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>国芯科技</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=688262" target="_blank" class="link link-hover">688262</a></td>
|
||
<td>GPU芯片IDM929已内测成功；RISC-V GPGPU研发结束</td>
|
||
<td><div class="badge badge-info">GPU</div></td>
|
||
</tr>
|
||
<!-- MLU -->
|
||
<tr class="hover">
|
||
<td>寒武纪</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=688256" target="_blank" class="link link-hover">688256</a></td>
|
||
<td>国内稀缺的云端AI芯片厂商，云端产品线已迭代至思元590系列</td>
|
||
<td><div class="badge badge-primary">MLU</div></td>
|
||
</tr>
|
||
<!-- ASIC & Customization -->
|
||
<tr class="hover">
|
||
<td>芯原股份</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=688521" target="_blank" class="link link-hover">688521</a></td>
|
||
<td>国内领先SOC及ASIC设计服务提供商，拥有一体化平台式定制芯片设计能力</td>
|
||
<td><div class="badge badge-secondary">ASIC</div> <div class="badge badge-outline">IP/定制</div></td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>翱捷科技</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=688220" target="_blank" class="link link-hover">688220</a></td>
|
||
<td>定制化ASIC能力国内领先</td>
|
||
<td><div class="badge badge-secondary">ASIC</div> <div class="badge badge-outline">定制</div></td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>灿芯股份</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=688691" target="_blank" class="link link-hover">688691</a></td>
|
||
<td>公司拥有一站式芯片定制服务，在ASIC定制芯片领域有成功案例</td>
|
||
<td><div class="badge badge-secondary">ASIC</div> <div class="badge badge-outline">定制</div></td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>科德教育</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=300192" target="_blank" class="link link-hover">300192</a></td>
|
||
<td>中昊芯英（持股5.52%）掌握TPU架构AI芯片核心研发技术并实现TPU芯片量产</td>
|
||
<td><div class="badge badge-secondary">ASIC</div> <div class="badge badge-warning">TPU</div></td>
|
||
</tr>
|
||
<!-- PCIe -->
|
||
<tr class="hover">
|
||
<td>万通发展</td>
|
||
<td><a href="https://valuefrontier.cn/company?scode=600246" target="_blank" class="link link-hover">600246</a></td>
|
||
<td>数渡科技（拟收购）核心产品为PCIe高速交换芯片，是AI服务器必不可少的关键芯片</td>
|
||
<td><div class="badge badge-success">PCIe交换芯片</div></td>
|
||
</tr>
|
||
</tbody>
|
||
</table>
|
||
</div>
|
||
|
||
<h3 class="text-2xl font-semibold mb-4 mt-8 text-sky-300">涨跌幅异动分析</h3>
|
||
<div class="overflow-x-auto">
|
||
<table class="table w-full">
|
||
<thead>
|
||
<tr class="text-white">
|
||
<th>股票名称</th>
|
||
<th>涨跌幅</th>
|
||
<th>日期</th>
|
||
<th>异动原因分析</th>
|
||
</tr>
|
||
</thead>
|
||
<tbody>
|
||
<tr class="hover">
|
||
<td>利扬芯片 (<a href="https://valuefrontier.cn/company?scode=688135" target="_blank" class="link">688135</a>)</td>
|
||
<td><span class="text-red-400">+20.01%</span></td>
|
||
<td>2025-09-17</td>
|
||
<td>主因市场对“Tesla AI Day”发布Dojo 2训练芯片的产业链误读，将其第三方测试业务与Tesla自研芯片强行关联，形成情绪型炒作。公司尚未具备5nm量产测试能力。</td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>电科芯片 (<a href="https://valuefrontier.cn/company?scode=600877" target="_blank" class="link">600877</a>)</td>
|
||
<td><span class="text-red-400">+5.72%</span></td>
|
||
<td>2025-08-28</td>
|
||
<td>受益于半导体芯片板块全面走强、AI算力概念热度提升、国产替代预期增强及卫星互联网概念活跃等多重因素共同作用。</td>
|
||
</tr>
|
||
<tr class="hover">
|
||
<td>泰凌微 (<a href="https://valuefrontier.cn/company?scode=688591" target="_blank" class="link">688591</a>)</td>
|
||
<td><span class="text-red-400">+20.01%</span></td>
|
||
<td>2025-06-25</td>
|
||
<td>受业绩高增长(H1净利+267%)和小米AR眼镜发布预期双重驱动，叠加公司端侧AI芯片放量带来的想象空间。</td>
|
||
</tr>
|
||
</tbody>
|
||
</table>
|
||
</div>
|
||
|
||
</section>
|
||
</main>
|
||
|
||
<footer class="text-center py-8 mt-12">
|
||
<p class="text-sm text-gray-600">报告生成时间: <span id="generation-time"></span></p>
|
||
<p class="text-xs text-gray-700 mt-2">免责声明：本报告由AI生成，内容仅供参考，不构成任何投资建议。投资者据此操作，风险自担。</p>
|
||
</footer>
|
||
|
||
</div>
|
||
|
||
<script>
|
||
document.getElementById('generation-time').textContent = new Date().toLocaleString('zh-CN');
|
||
|
||
var timelineChart = echarts.init(document.getElementById('timeline-chart'), 'dark');
|
||
var option = {
|
||
backgroundColor: 'transparent',
|
||
tooltip: {
|
||
trigger: 'axis',
|
||
formatter: function (params) {
|
||
return params[0].name + '<br/>' + params[0].value[2];
|
||
}
|
||
},
|
||
grid: {
|
||
left: '3%',
|
||
right: '4%',
|
||
bottom: '12%',
|
||
containLabel: true
|
||
},
|
||
xAxis: {
|
||
type: 'time',
|
||
axisLine: { lineStyle: { color: 'rgba(255,255,255,0.3)' } }
|
||
},
|
||
yAxis: {
|
||
type: 'value',
|
||
splitLine: { show: true, lineStyle: { color: 'rgba(255,255,255,0.1)' } },
|
||
axisLine: { lineStyle: { color: 'rgba(255,255,255,0.3)' } }
|
||
},
|
||
dataZoom: [{
|
||
type: 'inside',
|
||
start: 0,
|
||
end: 100
|
||
}, {
|
||
start: 0,
|
||
end: 20,
|
||
handleIcon: 'M10.7,11.9v-1.3H9.3v1.3c-4.9,0.3-8.8,4.4-8.8,9.4c0,5,3.9,9.1,8.8,9.4v1.3h1.3v-1.3c4.9-0.3,8.8-4.4,8.8-9.4C19.5,16.3,15.6,12.2,10.7,11.9z M13.3,24.4H6.7V23h6.6V24.4z M13.3,19.6H6.7v-1.4h6.6V19.6z',
|
||
handleSize: '80%',
|
||
handleStyle: {
|
||
color: '#fff',
|
||
shadowBlur: 3,
|
||
shadowColor: 'rgba(0, 0, 0, 0.6)',
|
||
shadowOffsetX: 2,
|
||
shadowOffsetY: 2
|
||
},
|
||
bottom: 10
|
||
}],
|
||
series: [{
|
||
name: '事件',
|
||
type: 'line',
|
||
showSymbol: true,
|
||
symbol: 'circle',
|
||
symbolSize: 10,
|
||
hoverAnimation: false,
|
||
data: [
|
||
{ name: '2023-10-17', value: ['2023-10-17', 1, '美国BIS限制A100/H100出口，打开国产替代窗口。'] },
|
||
{ name: '2025-01-10', value: ['2025-01-10', 2, '彭博社报道美国计划实施新一轮分级出口管制。'] },
|
||
{ name: '2025-03-01', value: ['2025-03-01', 3, '深圳、苏州等地出台政策支持AI芯片产业。'] },
|
||
{ name: '2025-04-28', value: ['2025-04-28', 4, '字节计划采购10-20万张国产卡，昇腾/海光具备4000卡集群能力。'] },
|
||
{ name: '2025-05-08', value: ['2025-05-08', 3, '传闻特朗普政府可能取消拜登时代AI芯片限制，带来不确定性。'] },
|
||
{ name: '2025-07-24', value: ['2025-07-24', 5, '传闻工信部可能因能耗比限制H20进口，但国产训练能力仍是痛点。'] },
|
||
{ name: '2025-08-04', value: ['2025-08-04', 6, '字节下单56万颗H20，华为/海光也获大单，市场需求旺盛。'] },
|
||
{ name: '2025-10-04', value: ['2025-10-04', 7, '寒武纪690/昆仑芯下一代产品计划公布，国产进入产品力竞争深水区。'] },
|
||
{ name: '2025-12-03', value: ['2025-12-03', 8, '亚马逊发布Trainium3，国际巨头自研加速。'] }
|
||
],
|
||
itemStyle: {
|
||
color: '#00BFFF'
|
||
},
|
||
lineStyle: {
|
||
color: '#8A2BE2',
|
||
width: 2
|
||
}
|
||
}]
|
||
};
|
||
timelineChart.setOption(option);
|
||
|
||
window.addEventListener('resize', function() {
|
||
timelineChart.resize();
|
||
});
|
||
</script>
|
||
</body>
|
||
</html> |