🐦 推特Tae KimCodex · 实盘不做把握 5/10不做原因:风险收益比不佳
北京时间·🕒 发布 06:45:52时区推断·⟳ 抓取 06:46:01·⚡ 决策 09-03 07:44:58 +58m57s·54.3 天前
Broadcom财报电话会披露AI半导体高增长与2027/2028远期指引，但短线交易信号偏混合
事件本身方向混合：Broadcom AI远期收入和OpenAI/Anthropic客户线索强，但2027年预期差被原帖明确标为低于买方bogey，2028收入指引兑现周期过长，公开媒体已在财报后集中报道且盘后反应不强；对AVGO做多缺少清晰未定价edge，对NVDA/MRVL也没有直接强空或强多催化。
📄 原文Tae Kim@firstadopter↗ 原文
Broadcom earnings call. Best hand-picked excerpts: “This was our Q3 AI semiconductor revenue, which grew 221% year on y…
不做 Broadcom（AVGO.US）
不追多：AI远期路径强但2027年相对买方预期偏失望，2028兑现太远，财报电话会信息已广泛传播且盘后反应偏弱，风险收益不足。
不做 Nvidia（NVDA.US）
不做空：Broadcom定制XPU叙事对GPU有竞争压力，但没有直接、强的Nvidia公司级利空，原帖还称HBM供应可能支持Nvidia。
不做 Marvell Technology（MRVL.US）
不交易：Marvell仅为竞争背景和先例参照，本事件没有给出其新增订单、指引或份额变化。
展开详情(成交明细 / 研究档案 / 原文全文 / 决策记录链)
研究档案
结论摘要
Broadcom官方财报和电话会确认AI半导体收入高速增长、Q4指引强、2027/2028远期收入路径上修，并点名Google、Anthropic、OpenAI等客户；但这不是低扩散独家，已被WSJ、Barron's、MarketWatch、IBD等财报报道同步覆盖。更关键的是，原帖自己的净评估是2027年AI收入指引低于买方预期、2028年强但需折现，盘后报道也显示AVGO未因利好顺畅上涨。因此当前不追AVGO，也不因定制XPU叙事去做空NVDA。
消息传播链
- Broadcom发布2026财年三季度正式业绩，确认收入、EPS、Q4收入指引和股息等基础财报事实。
- 电话会转录覆盖AI半导体收入、Google TPU、Anthropic、OpenAI Jalapeno及2027/2028 AI收入展望等原帖核心摘录。
- Tae Kim整理电话会精选摘录，并给出净评估：2027 AI收入指引低于买方bogey、2028强但需折现。
- IBD报道称Broadcom虽业绩和Q4指引超预期，但盘后股价下跌超过2%，显示短线市场未把电话会解读为单边利好。市场反应:AVGO盘后下跌超过2%，未形成利好顺畅重定价。
关键理由
- Broadcom 2026财年Q3收入29.6亿美元、非GAAP EPS 3.32美元，AI半导体收入16.7亿美元并给出2027年1150亿美元、2028年2300亿美元的远期AI收入路径。
- 若远期指引兑现，AI半导体收入将从2026年约580亿美元到2027年约1150亿美元、2028年约2300亿美元，量级相对当前收入基数很大。
- 但短线交易的核心不是名义增长，而是相对预期差；原帖已指出2027指引相对买方bogey偏失望，2028两年远期兑现风险高。
主要风险
- Broadcom盘后财报属于明牌高关注事件，媒体已广泛传播，若开盘高追容易被2027预期差和毛利/融资疑虑压制。
- OpenAI、Anthropic和Google部署仍受HBM、数据中心建设、融资结构和客户自研路线影响，两年远期指引不能等同于已确认短期收入。
反向证据
- 做多前查反向 · 来源 web · 日期 2026-09-02 · 影响 weakens
多家媒体称AVGO财报后盘后下跌或反应谨慎，原因包括2027预期差、Q4利润率、Google/Marvell竞争和融资敞口。
检索:Broadcom 2026 Q3 earnings stock falls despite upbeat earnings 2027 AI revenue guidance disappointing 原文 ↗ - 做空前查反向 · 来源 web · 日期 2026-09-02 · 影响 weakens
官方财报和电话会显示AI收入增长、客户部署和远期收入路径仍强，不能把预期差直接升级为强空催化。
检索:Broadcom AI semiconductor revenue 115 billion 2027 230 billion 2028 OpenAI Anthropic strong guidance 原文 ↗
先例如何影响了这次决策
对本次决策的影响:先例把AVGO从可小仓追AI远期上修降为AVOID，等待更清晰买方预期差或开盘承接确认；NVDA/MRVL也不因二阶叙事开仓。
方向:相似样本倾向说明AI芯片财报和客户大单叙事若只是低于高预期或兑现太远，短线容易冲高回落或下跌；当前不支持追多。
涨幅量级:没有8分以上高相似且锚点干净的样本，不能据此给AVGO明确tp/sl幅度。
形态:Broadcom 6月和Marvell 8月均显示AI定制芯片高预期下的见光死风险，尤其当长期故事强但近期指引不够超预期时。
历史先例(按相似度)
- 相似度 7/10 AVGO 2026-06-03 2026年6月Broadcom二季度收入和AI芯片指引低于预期且维持2027长期目标，盘后和次日出现大跌。 原文 ↗
事件后 +1日 -12.6% · +2日 -19.5% · +3日 -17.2% · +4日 -18.2% · +5日 -22.4%
⚠非首发(转载/聚合)·来源层级不符(声明wire/实为aggregator)·未核对目标相位 — 该行涨幅别用来定止盈止损
相似度依据:同为Broadcom财报与AI芯片远期预期差事件，机制高度相近，但当时是收入/AI指引低于预期，本次是远期强弱混合。
经验:Broadcom AI叙事即使长期强，只要低于高预期，短线可被买方bogey压制。 - 相似度 6/10 MRVL 2026-08-27 Marvell上调年度预测但未给出足够清晰的Google合作近期收入增量，盘后和次日下跌。 原文 ↗
事件后 +1日 -10.3% · +2日 -12.3% · +3日 -12.9%
⚠来源层级不符(声明wire/实为media) — 该行涨幅别用来定止盈止损
相似度依据:同属AI定制芯片和Google相关预期差，但标的是Marvell且事件为Google合作后的业绩说明，非Broadcom自身财报。
经验:AI定制芯片订单叙事若缺少近期收入兑现和更高指引，市场可能卖事实。 - 相似度 5/10 AVGO 2026-08-19 Marvell与Google达成AI芯片协议，引发市场担忧Broadcom在Google自研芯片生态中的份额。 原文 ↗
事件后 +1日 -0.4% · +2日 -0.0% · +3日 +1.2% · +4日 -1.5% · +5日 -2.0%
⚠非首发(转载/聚合)·来源层级不符(声明media/实为aggregator) — 该行涨幅别用来定止盈止损
相似度依据:与当前电话会中的Google TPU强关系相反，属于竞争份额担忧先例，可用于校准AVGO AI客户集中风险。
经验:Google相关定制芯片份额变化会快速影响Broadcom估值，但需要明确合同或客户转移才是强催化。
工具执行结果 · 成功 4 · 降级 0 · 失败 0 · 取消 0
- web_search 成功 ? · 0 ms
- web_search 成功 ? · 0 ms
- web_search 成功 ? · 0 ms
- web_search 成功 ? · 0 ms
原文全文
展开原文 · 2,531 字
Broadcom earnings call. Best hand-picked excerpts:
“This was our Q3 AI semiconductor revenue, which grew 221% year on year and up 54% sequentially.”
“Demand was simply hot and we're just getting started. Our six XPU customers are accelerating their adoption of custom accelerators and AI semiconductor revenue."
TPU v8i “is comparable, if not surpasses the Vera Rubin GPU.”
“Broadcom is now shipping the TPU version eight ahead of the Mediatek version”
“In Q3, we also shipped Jalapeno OpenAI first generation custom accelerator, which outperforms Grace Blackwell's GPU for inference workloads.”
On Google TPUs: “We are planning to deliver multi tens of billions of dollars of TPUs annually over the next several years. We expect this growing demand in 28 and 29 to be fulfilled through successive generations of the increasingly complex TPUs.”
“We expect Anthropic to deploy another five gigawatts of TPU version eight in 2027. And in 2028, we have clear line of sight to deliver another and incremental ten gigawatts even as we expect Google to grow for us.”
“Anthropic is on track to become our largest XPU customer in 2027.”
“OpenAI .. on track for the planned deployment of 1.3GW in 2027 .. we have a line of sight for OpenAI to deploy over five gigawatts of Jalapeno and its successors [in 2028] .. would make OpenAI our second largest .. customer”
"Here is our outlook for our AI semiconductor revenue. In 2027, we have secured the supply to again double AI revenue to approximately $115 billion. Our demand actually exceeds this outlook and we will work to improve supply. In 2028, we expect the trajectory of growth to continue. We have line of sight for fiscal 2028 a semiconductor revenue growth to again double to $230 billion. Here again, we have secured the supply to meet this outlook.”
“As a result, I got to say, we are very much on target to exceed $30 in earnings per share in fiscal 2028.”
OpenAI and Anthropic: “They are using cloud services third party services to deploy their models. Long term, we see these guys to be no different from a hyperscaler and will run their own data centers and be first party to offer generative AI APIs .. models to the world. So we see that happening. It's not speculation. It's actually happening. And we are in the midst of enabling that”
Net-net: Looks like Jensen secured the HBM for 2027! Broadcom's 2027 AI revenue guidance is disappointing relative to buyside bogeys, but the 2028 outlook is strong. That said, you have to discount a two year out guide.
“This was our Q3 AI semiconductor revenue, which grew 221% year on year and up 54% sequentially.”
“Demand was simply hot and we're just getting started. Our six XPU customers are accelerating their adoption of custom accelerators and AI semiconductor revenue."
TPU v8i “is comparable, if not surpasses the Vera Rubin GPU.”
“Broadcom is now shipping the TPU version eight ahead of the Mediatek version”
“In Q3, we also shipped Jalapeno OpenAI first generation custom accelerator, which outperforms Grace Blackwell's GPU for inference workloads.”
On Google TPUs: “We are planning to deliver multi tens of billions of dollars of TPUs annually over the next several years. We expect this growing demand in 28 and 29 to be fulfilled through successive generations of the increasingly complex TPUs.”
“We expect Anthropic to deploy another five gigawatts of TPU version eight in 2027. And in 2028, we have clear line of sight to deliver another and incremental ten gigawatts even as we expect Google to grow for us.”
“Anthropic is on track to become our largest XPU customer in 2027.”
“OpenAI .. on track for the planned deployment of 1.3GW in 2027 .. we have a line of sight for OpenAI to deploy over five gigawatts of Jalapeno and its successors [in 2028] .. would make OpenAI our second largest .. customer”
"Here is our outlook for our AI semiconductor revenue. In 2027, we have secured the supply to again double AI revenue to approximately $115 billion. Our demand actually exceeds this outlook and we will work to improve supply. In 2028, we expect the trajectory of growth to continue. We have line of sight for fiscal 2028 a semiconductor revenue growth to again double to $230 billion. Here again, we have secured the supply to meet this outlook.”
“As a result, I got to say, we are very much on target to exceed $30 in earnings per share in fiscal 2028.”
OpenAI and Anthropic: “They are using cloud services third party services to deploy their models. Long term, we see these guys to be no different from a hyperscaler and will run their own data centers and be first party to offer generative AI APIs .. models to the world. So we see that happening. It's not speculation. It's actually happening. And we are in the midst of enabling that”
Net-net: Looks like Jensen secured the HBM for 2027! Broadcom's 2027 AI revenue guidance is disappointing relative to buyside bogeys, but the 2028 outlook is strong. That said, you have to discount a two year out guide.
决策记录链 · 只追加
每次领取产生一条记录,重新决策追加一条并以 supersedes 指向被替代的记录(ADR 0014)。
这条决策早于决策账本(2026-10-08 起逐条追加),没有追加记录。