TMT BreakoutCodex · 实盘不做把握 5/10
北京时间·🕒 发布 05:10:18时区推断·⟳ 抓取 12:45:32·⚡ 决策 08-05 12:47:29 +1m57s·83.1 天前
TMT Breakout盘后复盘汇总AMD、ANET、ALAB等科技股财报与AI板块异动
这是一篇盘后复盘型二级解读，核心财报事件并非首发，ANET和AMD相关催化已在近6.5小时内被系统处理过。
📄 原文TMT Breakout↗ 原文
TMTB EOD Wrap; AMD ANET ALAB BKNG LSCC First takes
How things can change in just a few days in AI/Tech land. With QQQs +3.4%, we’ve gone from the “Leopold Low” to QQQs up…
不做 Arista Networks（ANET.US）
ANET财报催化已被系统近时段单独处理，本条只是盘后复盘；同时公告前已连续上涨并接近52周高点，盘后跳升后不追。
不做 Astera Labs（ALAB.US）
ALAB业绩和指引偏强但财报前5日已上涨近39%，文中显示盘后冲高回落，拥挤交易下缺少可回测的正期望入场。
不做 Advanced Micro Devices（AMD.US）
AMD低于buyside bogey但仍为beat-and-raise，且相关财报催化已被系统近时段处理；负面不够强，不开空。
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
🔎 新鲜度:最早出现 2026-08-04 · 扩散 high · price-in 风险 high ↗
关键理由
- 这是一篇盘后复盘型二级解读，核心财报事件并非首发，ANET和AMD相关催化已在近6.5小时内被系统处理过。
- ANET财报明显优于预期但股价公告前已近5日上涨12.26%、靠近52周高点，盘后再跳升约11%，当前追高执行边界差。
- ALAB三季度指引显著高于Street和buyside bogey，但股价财报前5日已上涨38.98%，且文中显示盘后从+8%回落到-2%，拥挤和见光死风险高。
主要风险
- 若美股开盘后ANET或ALAB继续放量上行，说明市场仍在重定价，但本条复盘到达时缺少低风险入场价。
- AMD虽低于buyside bogey并盘后下跌，但仍为beat-and-raise，做空催化强度不足。
反向证据
- 做多前查反向 · 来源 search_processed_events · 日期 2026-08-05 · 影响 weakens
近时段已出现ANET财报超预期和AMD财报相关决策记录，本条作为复盘再推会重复打扰。
检索:search_processed_events hours=72 limit=200 - 做多前查反向 · 来源 stock_quote · 日期 2026-08-04 · 影响 weakens
ANET公告前5日已上涨12.26%、20日上涨14.45%，且贴近52周高位，财报后再跳升追高风险高。
检索:ANET.US stock_quote full - 做多前查反向 · 来源 stock_quote · 日期 2026-08-04 · 影响 weakens
ALAB公告前5日已上涨38.98%，文中还显示盘后冲高回落，强业绩已伴随拥挤交易。
检索:ALAB.US stock_quote full - 做空前查反向 · 来源 stock_quote · 日期 2026-08-04 · 影响 weakens
AMD虽低于buyside bogey，但仍高于Street且近期AI半导体beta强，负面不足以构成强空催化。
检索:AMD.US stock_quote full
工具调用 · 历史记录,结果未知(2 种)
- search_processed_events · 结果未知
- stock_quote · 结果未知
原文全文
展开原文 · 8,256 字
How things can change in just a few days in AI/Tech land. With QQQs +3.4%, we’ve gone from the “Leopold Low” to QQQs up ~10% in roughly four sessions, SOX up 15%+, and the AI vibes beginning to percolate again (eg, “Is RSI coming in the fall?”). Apparently, the market only knows how to travel at high speed; it just feels a lot more rational when the direction is up.
We’re now back where we were before the Leopold “unwind” began, which makes sense and was what we wanted to see as a first stop for semis as investors have gotten more confident throwing back on gross over the past couple days. That was the relatively easy trade, or at least the kind that becomes obvious ~15% after the bottom. From here, it gets tougher, and we shift back into a bit of “wait-and-see mode,” letting the market & nascent better AI vibes breathe and give us a better sense of where we’re headed next, but base case still remains choppy but higher over the next few weeks.
Post-close ANET/ALAB delivering the goods while fell AMD short of bogeys. These comments from Elon have NVDA +1% in the post:
The world we are in:
Overall, another very healthy day breadth wise with Internet, software and AI semis all rallying while hypersscalers digested the recent vertical moves since late last week.
Let’s get to it…
### **Post-close earnings:**
**AMD -7%:** Beat-and-raise, but below elevated buyside expectations for a crowded stock which leaned long into the print. Q2 came in at $11.5B vs bogeys closer to $12B and Q3 guide $13B at midpoint vs bogeys closer to $13.2B. Nothing here changes the bull debate as narrative centered around ‘27 and beyond. Definitely a miss of expectations, but imagine Lisa sounds bulled up on future product cycle (DC and GPU) and want to hear about Helios/GPU ramp in Q4 and CY27.
**2Q Results**
Revenue: $11.54B vs Street at $11.31B vs. bogeys of $12B
GM: 56.2% vs Street at 55.8%
EPS: $1.66 vs Street at $1.62
Data Center: $6.72B, or 58% of revenue, vs Street at $6.55B
Client: $3.06B vs Street at $3.00B
Gaming: $779M vs Street at $814M
Embedded: $977M vs Street at $978M
**3Q Guide**
Revenue: $13.0B vs Street at $12.5B vs bogeys of $13.2B
GM: 56.0% vs Street at 56.25%
**ANET 11%:** Looks really good with a giant beat and raise. Revs +38% y/y vs Street at 28%, accelerating from 35% in 1Q, well above bogeys while operating margins reached close to 50%. The 3Q revenue guide implies another acceleration to 43% growth and came in roughly 12% above Street and about 4-5% above bogeys with EPS ~17% above. ANET’s biggest revenue beat since October 2022, and bulls love the acceleration in revs. Stock leaned long into the print and was of only a few semis hitting 52wk highs, but very much delivered here.
**2Q Results**
Revenue: $3.04B vs Street at $2.83B
Operating margin: 49.9% vs Street at 46.5%
EPS: $1.02 vs Street at $0.89
**3Q Guide**
Revenue: $3.30B vs Street at $2.95B
Operating margin: 48.5% vs Street at 46.5%
EPS: $1.07 vs Street at $0.92
**ALAB -2%:** Looks like a solid beat and raise with Q3 much better than bogeys at $550M vs $475M and street at $417M. Expectations high and leaned long into the print, but #s seemed above buyside and margins look fine as well. Not sure why the stock selling off from +8% to -2%, but we’ll listen to the call and have more tomorrow.
**2Q Results**
Revenue: $392M vs Street at $361M
Gross margin: 73.7% vs Street at 73.0%
EPS: $0.80 vs Street at $0.69
**3Q Guide**
Revenue: $550M vs Street at $417M
Gross margin: 72.0% vs Street at 71.9%
EPS: $1.185 vs Street at $0.81
**SPCX (-7%):** Looks solid with much better revs but Starlink slightly light. Stock was up 9% today so just giving some of that back. Tough to tell where bogeys were for this one given first q out of the gate. AI benefited from roughly $1.6B tied to the new cloud deal. Elon also laid out a very aggressive xAI compute ramp, targeting 2GW by year-end and closer to 10GW by the end of next year, **while reiterating that xAI remains exclusive to NVDA** because it views NVDA as the best AI machine. With earnings now cleared and attention shifting back toward the lockup expiration, not surprising stock to be down.
2Q revenue: $7.81B vs Street at $7.11B
2Q AI revenue: $2.56B vs Street at $2.26B
Starlink subscribers: 12M vs Street at 12M
Starlink ARPU: $66 vs Street at $66
**BKNG (+5%):** Looks solid with room nights, bookings, revenue and EBITDA all ahead. The 3Q guide was softer, however, with nights, bookings and EBITDA growth all below Street but was expected and inline to slightly better than bogeys. The offset is the FY bookings outlook, which was raised to high-single-digit growth from low-to-high single digits, although the cc outlook remains unchanged at high single digits. Positioning mixed into the print so should be good enough
2Q nights: 325M, +5%, vs Street at 320M
2Q bookings: $51.0B, +9%, vs Street at $49.7B
2Q revenue: $7.4B, +8%, vs Street at $7.18B
2Q EBITDA: $2.6B vs Street at $2.56B
3Q nights growth guide: +3–5% vs Street at +6%
3Q bookings growth guide: +4–6% vs Street at +8.75%
3Q EBITDA growth guide: +4–6% vs Street at +8.6%
FY bookings growth: raised to high single digits from low-to-high single digits
FY constant-currency bookings growth: maintained at high single digits
**LSCC (-2%):** stronger top-line numbers than expected. Investors were looking for a low-single-digit beat-and-raise excluding the AMI deal; instead, LSCC beat quarterly revenue by 9% and guided 11% above Street. Mgmt pointed to strong demand, accelerating backlog and design-win momentum, with the FPGA business tracking toward 65% y/y growth and the company reaching a $1B revenue run rate ahead of expectations.
**2Q Results**
Revenue: $201M vs Street at $185.5M
Gross margin: 71.7% vs Street at 70.0%
EPS: $0.53 vs Street at $0.45
**3Q Guide**
Revenue excluding AMI: $220M vs Street at $199M
Gross margin: 69.5% vs Street at 70.4%
Revenue including two months of AMI: $255M at the midpoint
### **Roundup**
Earnings calls ongoing so recap short today. In *internet*, **AMZN -2**% after Bezos announced he would sell $4Bn worth of stock. In *software*, **PLTR +30%** the big standout where accelerating revs and implied lack of decel in the guide (given typical magnitude of beats), had bulls believing the Karp “sovereign” story and saying stock should re-rate back to levels at least inline with **SNOW, DDOG,** etc. Stock was shorted going in on the back of weaker 3p checks and “AI loser” narrative so just another sign of how much positioning can play in a big move, especially if accelerating #s help shift the numbers in a positive direction.
In AI semis, **ARM +18%** and **MRVL +13%** the big movers. Optical **(COHR +12% / LITE +9%)** got a bid after Reuters reported Trump was considering banning new optical transceivers. Lots of pushback and confusion about what exactly this means and how and if it gets implemented. MS had a good note out mid day explaining the push and pull saying COHR is the clearest scaled beneficiary given its vertical integration, with AAOI and FN also positioned to absorb incremental demand, while LITE should benefit from tighter EML supply but faces potential InP substrate risk from China. MS said the bigger issue is feasibility: non-Chinese suppliers cannot currently support total AI demand, making access to Chinese InP supply critical, noting clouds have spent years qualifying alternatives, and believes a broader US-China agreement involving reciprocal component purchases could ultimately be part of the solution. Key question is also whether mfg in China is part of the ban or just a corporate domiciled ban, which would affect how much it impacts LITE and COHR and others.
Otherwise strength around the board with most names up 5%+…
We liked this from resident technical HBF specialist Randy in TMTB Slack around **SNDK (+11%)/**SK Hynix’s NAND announcements:
That’s it for today - see you tomorrow morning with full earnings recaps.
We’re now back where we were before the Leopold “unwind” began, which makes sense and was what we wanted to see as a first stop for semis as investors have gotten more confident throwing back on gross over the past couple days. That was the relatively easy trade, or at least the kind that becomes obvious ~15% after the bottom. From here, it gets tougher, and we shift back into a bit of “wait-and-see mode,” letting the market & nascent better AI vibes breathe and give us a better sense of where we’re headed next, but base case still remains choppy but higher over the next few weeks.
Post-close ANET/ALAB delivering the goods while fell AMD short of bogeys. These comments from Elon have NVDA +1% in the post:
The world we are in:
Overall, another very healthy day breadth wise with Internet, software and AI semis all rallying while hypersscalers digested the recent vertical moves since late last week.
Let’s get to it…
### **Post-close earnings:**
**AMD -7%:** Beat-and-raise, but below elevated buyside expectations for a crowded stock which leaned long into the print. Q2 came in at $11.5B vs bogeys closer to $12B and Q3 guide $13B at midpoint vs bogeys closer to $13.2B. Nothing here changes the bull debate as narrative centered around ‘27 and beyond. Definitely a miss of expectations, but imagine Lisa sounds bulled up on future product cycle (DC and GPU) and want to hear about Helios/GPU ramp in Q4 and CY27.
**2Q Results**
Revenue: $11.54B vs Street at $11.31B vs. bogeys of $12B
GM: 56.2% vs Street at 55.8%
EPS: $1.66 vs Street at $1.62
Data Center: $6.72B, or 58% of revenue, vs Street at $6.55B
Client: $3.06B vs Street at $3.00B
Gaming: $779M vs Street at $814M
Embedded: $977M vs Street at $978M
**3Q Guide**
Revenue: $13.0B vs Street at $12.5B vs bogeys of $13.2B
GM: 56.0% vs Street at 56.25%
**ANET 11%:** Looks really good with a giant beat and raise. Revs +38% y/y vs Street at 28%, accelerating from 35% in 1Q, well above bogeys while operating margins reached close to 50%. The 3Q revenue guide implies another acceleration to 43% growth and came in roughly 12% above Street and about 4-5% above bogeys with EPS ~17% above. ANET’s biggest revenue beat since October 2022, and bulls love the acceleration in revs. Stock leaned long into the print and was of only a few semis hitting 52wk highs, but very much delivered here.
**2Q Results**
Revenue: $3.04B vs Street at $2.83B
Operating margin: 49.9% vs Street at 46.5%
EPS: $1.02 vs Street at $0.89
**3Q Guide**
Revenue: $3.30B vs Street at $2.95B
Operating margin: 48.5% vs Street at 46.5%
EPS: $1.07 vs Street at $0.92
**ALAB -2%:** Looks like a solid beat and raise with Q3 much better than bogeys at $550M vs $475M and street at $417M. Expectations high and leaned long into the print, but #s seemed above buyside and margins look fine as well. Not sure why the stock selling off from +8% to -2%, but we’ll listen to the call and have more tomorrow.
**2Q Results**
Revenue: $392M vs Street at $361M
Gross margin: 73.7% vs Street at 73.0%
EPS: $0.80 vs Street at $0.69
**3Q Guide**
Revenue: $550M vs Street at $417M
Gross margin: 72.0% vs Street at 71.9%
EPS: $1.185 vs Street at $0.81
**SPCX (-7%):** Looks solid with much better revs but Starlink slightly light. Stock was up 9% today so just giving some of that back. Tough to tell where bogeys were for this one given first q out of the gate. AI benefited from roughly $1.6B tied to the new cloud deal. Elon also laid out a very aggressive xAI compute ramp, targeting 2GW by year-end and closer to 10GW by the end of next year, **while reiterating that xAI remains exclusive to NVDA** because it views NVDA as the best AI machine. With earnings now cleared and attention shifting back toward the lockup expiration, not surprising stock to be down.
2Q revenue: $7.81B vs Street at $7.11B
2Q AI revenue: $2.56B vs Street at $2.26B
Starlink subscribers: 12M vs Street at 12M
Starlink ARPU: $66 vs Street at $66
**BKNG (+5%):** Looks solid with room nights, bookings, revenue and EBITDA all ahead. The 3Q guide was softer, however, with nights, bookings and EBITDA growth all below Street but was expected and inline to slightly better than bogeys. The offset is the FY bookings outlook, which was raised to high-single-digit growth from low-to-high single digits, although the cc outlook remains unchanged at high single digits. Positioning mixed into the print so should be good enough
2Q nights: 325M, +5%, vs Street at 320M
2Q bookings: $51.0B, +9%, vs Street at $49.7B
2Q revenue: $7.4B, +8%, vs Street at $7.18B
2Q EBITDA: $2.6B vs Street at $2.56B
3Q nights growth guide: +3–5% vs Street at +6%
3Q bookings growth guide: +4–6% vs Street at +8.75%
3Q EBITDA growth guide: +4–6% vs Street at +8.6%
FY bookings growth: raised to high single digits from low-to-high single digits
FY constant-currency bookings growth: maintained at high single digits
**LSCC (-2%):** stronger top-line numbers than expected. Investors were looking for a low-single-digit beat-and-raise excluding the AMI deal; instead, LSCC beat quarterly revenue by 9% and guided 11% above Street. Mgmt pointed to strong demand, accelerating backlog and design-win momentum, with the FPGA business tracking toward 65% y/y growth and the company reaching a $1B revenue run rate ahead of expectations.
**2Q Results**
Revenue: $201M vs Street at $185.5M
Gross margin: 71.7% vs Street at 70.0%
EPS: $0.53 vs Street at $0.45
**3Q Guide**
Revenue excluding AMI: $220M vs Street at $199M
Gross margin: 69.5% vs Street at 70.4%
Revenue including two months of AMI: $255M at the midpoint
### **Roundup**
Earnings calls ongoing so recap short today. In *internet*, **AMZN -2**% after Bezos announced he would sell $4Bn worth of stock. In *software*, **PLTR +30%** the big standout where accelerating revs and implied lack of decel in the guide (given typical magnitude of beats), had bulls believing the Karp “sovereign” story and saying stock should re-rate back to levels at least inline with **SNOW, DDOG,** etc. Stock was shorted going in on the back of weaker 3p checks and “AI loser” narrative so just another sign of how much positioning can play in a big move, especially if accelerating #s help shift the numbers in a positive direction.
In AI semis, **ARM +18%** and **MRVL +13%** the big movers. Optical **(COHR +12% / LITE +9%)** got a bid after Reuters reported Trump was considering banning new optical transceivers. Lots of pushback and confusion about what exactly this means and how and if it gets implemented. MS had a good note out mid day explaining the push and pull saying COHR is the clearest scaled beneficiary given its vertical integration, with AAOI and FN also positioned to absorb incremental demand, while LITE should benefit from tighter EML supply but faces potential InP substrate risk from China. MS said the bigger issue is feasibility: non-Chinese suppliers cannot currently support total AI demand, making access to Chinese InP supply critical, noting clouds have spent years qualifying alternatives, and believes a broader US-China agreement involving reciprocal component purchases could ultimately be part of the solution. Key question is also whether mfg in China is part of the ban or just a corporate domiciled ban, which would affect how much it impacts LITE and COHR and others.
Otherwise strength around the board with most names up 5%+…
We liked this from resident technical HBF specialist Randy in TMTB Slack around **SNDK (+11%)/**SK Hynix’s NAND announcements:
That’s it for today - see you tomorrow morning with full earnings recaps.
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。