以下部分引用 AI 总结，现阶段 AI 仍然有幻觉。请以内容的原文为准。

数据新鲜度群聊 07-14 ✓卖方 07-14 ✓Wrap 断档15天 ✕News 07-15 ✓X 断档12天 ✕页面生成 07-15

Intel 18A

29 atoms · 跨 13 天 · 首见 2026-04-26 · 最近 2026-07-03

三色: 🟦 fact 19 · 🟥 take 10 · stance ▲15/▼3/◆3

叙事状态: 🍂 fading (退潮) · 💤 沉寂 · 策展近14d 0 atom · 🐦 X 近14d 2

状态只算低频策展源 (群/卖方); X firehose 仅作背景音量

来源: X 29

时态: fresh:28 · aging:1

标签: 好数字:11 · 好观点:10 · 好信源:7 · 好思考:3 · 好问题:2

🟨 AI 综合 · junior analyst 概览

展开 AI 综合 (灰色 · 非市场结论 · 点击数字溯源到原 atom)

model: deepseek-chat · 2026-07-12 · 默认折叠

Intel 18A 节点在截面图像上展现了极高的整洁度和工艺精度，RibbonFET与背面供电技术首次在同一截面中清晰集成 ¹ ¹ ¹。该节点实现了GAAFET和背面供电两大变革，被视为近20年来更新换代最大的节点 ¹，并已在数据中心凭借Intel Xeon 6+ 首发 ¹。工艺细节上，NMOS/PMOS宽度可独立调节，且power via精准对齐背面金属 ¹ ¹。良率方面，虽早期达标耗时长但正在爬坡，目前达到70%以上良率并进入全规模生产 ¹ ¹，不过6月报告仍指出良率未达目标，而7月报告称wafer-to-wafer良率问题已修复 ¹ ¹。然而，18A 产线因制造自家芯片已满载，可能无法为Apple A20腾出产能 ¹。

🧭 拥挤度 (一人一票): ▲ 6 位作者 (KOL6) vs ▼ 1 位作者 (KOL1)

⚖️ 多头 6/6 来自KOL

🟦 客观事实 (facts) (19)

<a id="atom-61fd08998b174a9a"></a>🟦 2026-06-05 X·@bubbleboi ⭐ · NMOS sheets宽度 · 技术路线
They made the NMOS sheets ~50% wider than PMOS, trading sheet width to equalize n/p drive strength without burning area on extra fins.
tags: 好数字·好思考 · → daily
value: qty=~50% wider than PMOS · direction=wider
<a id="atom-e755913eb22f2f6f"></a>🟦 2026-06-05 X·@bubbleboi ⭐ · power via alignment · 技术路线
the power via. Landing clean on the backside metal, nanometers of alignment, no contention with M0 up top.
tags: 好数字·好思考 · → daily
value: qty=nanometers of alignment
<a id="atom-7f66f73d34605ba0"></a>🟦 2026-06-05 X·@BenBajarin · 良率正在提升 · 良率
yields took a LONG time to catch up but they are scaling now
tags: 好数字·好信源 · → daily
<a id="atom-76c55e457e2d797d"></a>🟦 2026-05-01 X·@DrFrederickChen · 良率状态 · 良率
18A yields improving but still not yet at target
tags: 好数字 · → daily
value: direction=improving
<a id="atom-d1740efeb6d7bf90"></a>🟦 2026-06-08 X·@intel · debuts in data center with Intel Xeon 6+ · 产品发布
Intel 18A debuts in the data center with the new Intel Xeon 6+
tags: 好信源 · → daily
value: qty=new · date=2026-06-08
<a id="atom-2195b1d7458a4995"></a>🟦 2026-07-03 X·@tomshardware · wafer-to-wafer yield issues fixed · 良率
Intel 18A wafer-to-wafer yield issues fixed, report claims
tags: 好数字 · → daily
value: direction=positive · date=2026-07-03
<a id="atom-5a58cf4ce6c6e94c"></a>🟦 2026-06-15 X·@Alex_Intel_ · D0 defect density at September 2024 · 良率
18A was 0.4 around September 24
tags: 好数字 · → daily
value: qty=0.4 · date=September 2024
<a id="atom-43ae8fa2131e2a2c"></a>🟦 2026-06-03 X·@meng59739449 · 良率达到 70% · 良率
18A process is now at full scale production that mean Yield at least 70% reach industry standard
tags: 好数字 · → daily
value: qty=70% · date=2026-06-03
<a id="atom-9351a93b418475f2"></a>🟦 2026-06-02 X·@IanCutress · 生产状态 · 供给产能/技术路线
Intel 18A is now at full scale.
tags: 好数字 · → daily
value: date=2026-06-02
<a id="atom-0dff22679b017a1a"></a>🟦 2026-05-31 X·@bubbleboi · yield improvement rate · 良率
It’s been improving 5-7% for the past 6 months meaning it’s between 60-65% now.
tags: 好数字 · → daily
value: qty=5-7% per month for past 6 months · date=2026-05-31
<a id="atom-4d641e18cd205148"></a>🟦 2026-05-31 X·@bubbleboi · current yield range · 良率
meaning it’s between 60-65% now.
tags: 好数字 · → daily
value: qty=60-65% · date=2026-05-31
<a id="atom-a109a911eb51a50f"></a>🟦 2026-06-05 X·@BenBajarin · 产能需求旺盛 · 订单需求/供给产能
working to add capacity given both their products and capacity are in demand
tags: 好信源 · → daily
<a id="atom-3f04957653fde611"></a>🟦 2026-06-03 X·@lithos_graphein · 使用 Applied Materials Centura Sculpta 工具 · 技术路线
TechInsights is saying Intel 18A is using Applied Materials' new EUV pattern-shaping tool (Centura Sculpta) to print contact(?).
tags: 好信源 · → daily
📷 原图
<a id="atom-fe576116eabebebb"></a>🟦 2026-05-25 X·@meng59739449 · High NA EUV version Panther Lake test chip · 技术路线/产品发布
if you see before flow looks like Blue Sky Creek 18A High NA EUV version Panther Lake test chip
tags: 好信源 · → daily
value: qty=18A · date=before H2/2027
📷 原图
<a id="atom-4b1f8c4e5b43e20e"></a>🟦 2026-05-01 X·@DrFrederickChen · 对毛利率的影响 · 毛利率
still a headwind for margins
tags: 好观点 · → daily
value: direction=headwind
<a id="atom-0b987b36b2d4177b"></a>🟦 2026-05-01 X·@MCH2024 · 工艺制程分类 · 技术路线
SF2 / i18A are 3nm class
tags: 好信源 · → daily
value: qty=3nm class · date=
<a id="atom-efda59d84a9a4726"></a>🟦 2026-04-26 X·@BenBajarin · 制程节点 · 技术路线/制造节点
18A is roughly 2nm ish.
tags: 好信源 · → daily
value: qty=2nm ish
<a id="atom-f04dee3b129ec083"></a>🟦 2026-05-31 X·@bubbleboi · design activity status · 订单需求/技术路线
No one is designing for 18A anyway only 14A & 18AP now
tags: 好观点 · → daily
value: direction=no one designing for 18A
<a id="atom-af1eade36d9214f9"></a>🟦 2026-06-05 X·@Michaelskywal · NMOS比PMOS宽 · 技术路线
That is so interesting that the NMOS is bigger than the PMOS. Most of the time you need effectively 2X as much PMOS to balance the NMOS!
tags: 好问题 · → daily

🟥 多头 takes (bullish) (8)

<a id="atom-a389beea9aafd048"></a>🟥 2026-06-05 X·@CostaV02 · 截面图像质量 · 技术路线/产品发布
Am I wrong or is this the first public picture of 18A that looks really clean across the entire Cutaway? cutaways from earlier years had given me the the vague sense of … ruggedness, although it's hard to pinpoint exactly. This looks more symmetric and uniform to me
tags: 好观点·好问题 · → daily
<a id="atom-6f0c5f3fa87ed9be"></a>🟥 2026-06-05 X·@bubbleboi · 独立调整NMOS和PMOS宽度能力 · 技术路线/竞争格局
The more interesting aspect that I don't think many have realized is that you can tune both independently in a Gate All Around Transistor. On finFETs you need to add whole fins one at a time, so you couldn't fine-tune each device like this.
tags: 好思考 · → daily
<a id="atom-6830a7cd5a266395"></a>🟥 2026-06-05 X·@FundyAnalysis · RibbonFET与背面供电集成 · 技术路线/产品发布
it's nice seeing RibbonFET + backside power delivery show up this cleanly in the same cross-section. Patterning, alignment, isolation, via placement…
tags: 好观点 · → daily
<a id="atom-68846eb2978475d2"></a>🟥 2026-06-05 X·@lithos_graphein · 截面图像整洁度 · 技术路线
You're right. It's clean af.
tags: 好观点 · → daily
<a id="atom-33a6491f431a7504"></a>🟥 2026-06-05 X·@bubbleboi · 拥有背面供电 · 技术路线
No you have not. You still don’t have backside power delivery.
tags: 好观点 · → daily
value: direction=has
<a id="atom-15ebb1fe47a93ef2"></a>🟥 2026-06-05 X·@BenBajarin · 20年来节点更新换代最大 · 技术路线/技术路线
18A has two monumental changes, GAAFET and BSP, its rare to make one change like this let alone 2 in one process node.
tags: 好观点 · → daily
<a id="atom-08b5fe64deb14350"></a>🟥 2026-06-04 X·@lithos_graphein · TEM图像质量 · 技术路线/产品发布
Intel 18A is a game changer. This is an absolutely amazing TEM image of the nanosheet & power via patterning at play here in this device. Flawless.
tags: 好观点 · → daily
📷 原图
<a id="atom-9e947135d1f4a6ed"></a>🟥 2026-05-05 X·@RYUNPON000 [老化中] · 假设全面生产下的盈利状态 · 盈利能力
あのまま突っ走ってたら生産力不足もなく18Aフル生産で今頃大黒字だったんじゃ…
tags: 好观点 · → daily
value: direction=positive

🟥 空头 takes (bearish) (1)

<a id="atom-c1ea6142c16cc3f4"></a>🟥 2026-07-03 X·@mweinbach · 没有产能给 Apple A20 · 供给产能/供应链
Intel's 18A node is at capacity for Intel making it's own chips. They won't have ANY capacity at all for Apple within that timeline
tags: 好观点 · → daily
value: qty=0 · date=2026

⏱ 时间轴 (近 20)

🟦 2026-07-03 · fact · wafer-to-wafer yield issues fixed · → daily
🟥 2026-07-03 · forecast · 没有产能给 Apple A20 · → daily
🟦 2026-06-15 · fact · D0 defect density at September 2024 · → daily
🟥 2026-06-11 · forecast · 产能 · → daily
🟦 2026-06-08 · fact · debuts in data center with Intel Xeon 6+ · → daily
🟦 2026-06-05 · narrative · NMOS比PMOS宽 · → daily
🟥 2026-06-05 · narrative · 独立调整NMOS和PMOS宽度能力 · → daily
🟥 2026-06-05 · narrative · RibbonFET与背面供电集成 · → daily
🟥 2026-06-05 · narrative · 截面图像质量 · → daily
🟥 2026-06-05 · narrative · 截面图像整洁度 · → daily
🟦 2026-06-05 · fact · NMOS sheets宽度 · → daily
🟦 2026-06-05 · fact · power via alignment · → daily
🟥 2026-06-05 · refute · 拥有背面供电 · → daily
🟥 2026-06-05 · narrative · 20年来节点更新换代最大 · → daily
🟦 2026-06-05 · fact · 良率正在提升 · → daily
🟦 2026-06-05 · fact · 产能需求旺盛 · → daily
🟥 2026-06-04 · narrative · TEM图像质量 · → daily
🟦 2026-06-03 · fact · 良率达到 70% · → daily
🟦 2026-06-03 · fact · 使用 Applied Materials Centura Sculpta 工具 · → daily
🟦 2026-06-02 · fact · 生产状态 · → daily

← 实体目录 · 系统日志