摩根士丹利Codex · 实盘不做把握 6/10
北京时间·🕒 发布 11:58:07时区推断·⟳ 抓取 12:40:20·⚡ 决策 07-28 12:42:07 +1m47s·91.2 天前
摩根士丹利维持 Navitas 低配并下调目标价,但财报指引利好与设计定点缺失相互抵消
Navitas 二季度收入 1050万美元、三季度收入指引中点 1350万美元,高于市场预期,AI基础设施收入预计四季度超过三分之一,方向上有增长确认。
📄 原文摩根士丹利↗ 原文
Navitas Semiconductor Corp: Progress, but proof points still pending
Progress on the Navitas 2.0 transformation is clear and running ~one quarter ahead of plan; what remains elusive is the…
不做 Navitas Semiconductor(NVTS.US)
财报指引利好与摩根士丹利低配/设计定点缺失相互抵消;目标价12.60美元仍高于现价11.41美元,做空催化不够强,做多又缺正式 design-win 证据和清晰未定价增量。
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
消息传播链
- Navitas 发布二季度业绩,确认收入环比增长、三季度收入指引继续增长、AI基础设施年末占比超过三分之一。
- 财经媒体转述财报与投资者材料,强调 AI 转型但盈利仍滞后,盘后股价回落。市场反应:报道显示盘后跌约3%。
- 摩根士丹利认为 Navitas 2.0 进展快于计划,但正式 design-win 证据仍缺失,因此下调估值倍数并维持低配。
关键理由
- Navitas 二季度收入 1050万美元、三季度收入指引中点 1350万美元,高于市场预期,AI基础设施收入预计四季度超过三分之一,方向上有增长确认。
- 摩根士丹利维持 Underweight 并把目标价从 13.70美元下调至 12.60美元,核心担忧是 800V/GaN 高价值插槽仍缺正式 design-win 证据。
- 公司近20日已跌约34%,但7月27日正股收涨4.49%、盘后又回落,说明市场对业绩利好和估值/兑现风险正在拉扯,短线同向 edge 不清晰。
主要风险
- 若后续披露明确 AI 数据中心 design win,当前低配逻辑会被快速证伪。
- 若盘前继续下跌,更多是财报后估值重估而非新增经营损害,不能用“跌多了”反手做多。
反向证据
- 做多前查反向 · 来源 kb_search · 日期 2026-07-28 · 影响 weakens
内部库命中 Navitas 财报会议纪要和 800V 电源行业纪要,但未见正式 design-win 证据,削弱追多。
检索:Navitas NVTS AI infrastructure 800V design wins Morgan Stanley underweight July 2026 原文 ↗ - 做空前查反向 · 来源 web · 日期 2026-07-27 · 影响 weakens
公司收入和三季度指引好于预期,AI基础设施占比提升,构成做空反证。
检索:Navitas Q2 2026 earnings AI infrastructure revenue guidance design win 原文 ↗
工具调用 · 历史记录,结果未知(3 种)
- search_processed_events · 结果未知
- stock_quote · 结果未知
- kb_search · 结果未知
原文全文
展开原文 · 4,379 字
Progress on the Navitas 2.0 transformation is clear and running ~one quarter ahead of plan; what remains elusive is the formal design-win evidence. JunQ beat at the high end of guidance and was broadly in-line with MSe/Street. SepQ guide of $13.5mn (+28% q/q at midpoint) is ahead of prior MSe, marking the first quarter of y/y revenue growth. Mobile and low-end consumer is on track to be insignificant by year-end, with AI infrastructure - defined as AI data centers+grid/energy - expected to exceed ⅓ of Q4 revenue. Near-term revenue is SiC-heavy in traditional AC/DC PSU designs, and the 800V sidecar opportunity - while real - is not expected to ramp meaningfully until 1H27, which raises the question of whether competitors have already secured the early sockets. The larger GaN step-up remains contingent on later 800V-class architectures, customer qualification, and production allocation, which is dependent on potential changes to Kyber.
While we acknowledge strength in the direction of travel for Navitas, the stock is pricing in a 2027 outcome that record book-to-bill and backlog commentary alone cannot de-risk. We lower our multiple modestly from 50x to 40x to reflect greater duration risk, which brings our PT down from $13.70 to $12.60 and remain UW.
What we learned from the call/callback/release:
1.) (+) Navitas 2.0 running ahead of plan as AI infrastructure mix building faster than expected. High-power markets grew more than 50% y/y in JunQ, with both GaN and SiC contributing to sequential growth and particular SiC acceleration. Mobile exit is tracking ~one quarter ahead of schedule, and AI infrastructure is expected to exceed one-third of Q4 revenue. Record book-to-bill and backlog extending beyond 2026 are constructive, and mgmt described the organizational restructuring as substantially complete with focus now shifting to execution and profitability.
2.) (=) 800V transition is multi-step; pre-800V SiC ramp is the near-term driver, GaN step-up remains 2027+. Mgmt articulated four inflection points, with Inflection 1 (higher-power AC/DC PSUs, SiC replacing silicon) already ramping in 2H26. The 800V sidecar (Inflection 2) is expected in 1H27, with native 800V compute trays (Inflection 3, GaN-critical) ramping through 2027 and accelerating in late 2027/early 2028. Mgmt emphasized the transition happens in multiple steps and that Navitas is not dependent on one NVIDIA platform, one GPU vendor, or one hyperscaler. We find the framework credible but note that formal design-win evidence remains absent and power customers typically dual- or triple-source.
3.) (-) Competitive intensity remains an overhang. Competition from peers remains fierce across both SiC and GaN, and no formal design wins have been announced to confirm Navitas' positioning ahead of the 2027 ramp. Separately, the MagnaChip GeneSiC licensing deal and ongoing Wolfspeed and Renesas litigation introduce additional variables that are challenging to quantify.
Our estimates remain largely unchanged, though we lower our multiple given valuations overall under pressure around the power semi theme.
Key Takeaways
JunQ rev beat vs Street/MSe; SepQ guide beat ($13.5mn vs $11.4/$11.1mn MSe/Street).
Mobile declining as planned; AI Infra expected to exceed ⅓ of revenue by 4Q.
Duration risk for highest-value sockets in 800V remains; no design wins announced yet.
Result Details:
• Details on the June quarter: Revenue of $10.5mn was up 22.5% q/q and down 27.3% y/y, above MSe/Street of $10.0mn/$10.0mn. Non-GAAP GM/EPS of 39.5%/($0.04) were also in-line with MSe 39.3%/($0.04) and Street of 38.8%/($0.04).
• September quarter guide: The company guided to revenue of $13.5mn at the midpoint, up 28.2% q/q and up 33.5% y/y, well above MSe and Street of $11.4mn/$11.1mn. Non-GAAP GM guide of 39.7% at the midpoint was broadly in-line with MSe/Street of 39.6%/39.2%.
• Changes to our estimates: Our FY26-FY27 estimates come up slightly driven by increased revenue guide. Our FY26 estimates come up from $43.1mn/($0.15) to $47.6mn/($0.15) and our FY27 estimates come up from $65.5mn/($0.15) to $68.2mn/($0.10).
• Changes to our PT: We lower our multiple from 50x to 40x CY27 revenue to reflect duration risk, though still at a premium to reflect the potential growth opportunity. Applied to our updated CY27 revenue estimate of $68.2mn, our PT moves from $13.70 to $12.60.
While we acknowledge strength in the direction of travel for Navitas, the stock is pricing in a 2027 outcome that record book-to-bill and backlog commentary alone cannot de-risk. We lower our multiple modestly from 50x to 40x to reflect greater duration risk, which brings our PT down from $13.70 to $12.60 and remain UW.
What we learned from the call/callback/release:
1.) (+) Navitas 2.0 running ahead of plan as AI infrastructure mix building faster than expected. High-power markets grew more than 50% y/y in JunQ, with both GaN and SiC contributing to sequential growth and particular SiC acceleration. Mobile exit is tracking ~one quarter ahead of schedule, and AI infrastructure is expected to exceed one-third of Q4 revenue. Record book-to-bill and backlog extending beyond 2026 are constructive, and mgmt described the organizational restructuring as substantially complete with focus now shifting to execution and profitability.
2.) (=) 800V transition is multi-step; pre-800V SiC ramp is the near-term driver, GaN step-up remains 2027+. Mgmt articulated four inflection points, with Inflection 1 (higher-power AC/DC PSUs, SiC replacing silicon) already ramping in 2H26. The 800V sidecar (Inflection 2) is expected in 1H27, with native 800V compute trays (Inflection 3, GaN-critical) ramping through 2027 and accelerating in late 2027/early 2028. Mgmt emphasized the transition happens in multiple steps and that Navitas is not dependent on one NVIDIA platform, one GPU vendor, or one hyperscaler. We find the framework credible but note that formal design-win evidence remains absent and power customers typically dual- or triple-source.
3.) (-) Competitive intensity remains an overhang. Competition from peers remains fierce across both SiC and GaN, and no formal design wins have been announced to confirm Navitas' positioning ahead of the 2027 ramp. Separately, the MagnaChip GeneSiC licensing deal and ongoing Wolfspeed and Renesas litigation introduce additional variables that are challenging to quantify.
Our estimates remain largely unchanged, though we lower our multiple given valuations overall under pressure around the power semi theme.
Key Takeaways
JunQ rev beat vs Street/MSe; SepQ guide beat ($13.5mn vs $11.4/$11.1mn MSe/Street).
Mobile declining as planned; AI Infra expected to exceed ⅓ of revenue by 4Q.
Duration risk for highest-value sockets in 800V remains; no design wins announced yet.
Result Details:
• Details on the June quarter: Revenue of $10.5mn was up 22.5% q/q and down 27.3% y/y, above MSe/Street of $10.0mn/$10.0mn. Non-GAAP GM/EPS of 39.5%/($0.04) were also in-line with MSe 39.3%/($0.04) and Street of 38.8%/($0.04).
• September quarter guide: The company guided to revenue of $13.5mn at the midpoint, up 28.2% q/q and up 33.5% y/y, well above MSe and Street of $11.4mn/$11.1mn. Non-GAAP GM guide of 39.7% at the midpoint was broadly in-line with MSe/Street of 39.6%/39.2%.
• Changes to our estimates: Our FY26-FY27 estimates come up slightly driven by increased revenue guide. Our FY26 estimates come up from $43.1mn/($0.15) to $47.6mn/($0.15) and our FY27 estimates come up from $65.5mn/($0.15) to $68.2mn/($0.10).
• Changes to our PT: We lower our multiple from 50x to 40x CY27 revenue to reflect duration risk, though still at a premium to reflect the potential growth opportunity. Applied to our updated CY27 revenue estimate of $68.2mn, our PT moves from $13.70 to $12.60.
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。