Dean Pernas

SaaS companies are down 27% YTD, while the S&P 500 is roughly flat. The sector has become an increasing source of concern for investors. What we are witnessing is a transition into a new phase of the SaaS lifecycle. This sell-off marks the end of the halo period that long insulated SaaS from conventional competitive scrutiny. For more than a decade, capital flowed freely into software companies with little concern about existential risk. The business model appeared unusually insulated: long growth runways, recurring revenue, high gross margins, and high switching costs. This combination created ideal conditions for structural overvaluation. Our contention is that, for much of the time these companies have been public, they were overvalued because markets priced their economics as both durable and permanent.

This was originally intended to be a write-up on Zoom, as we believed the company held a strong strategic advantage in its core video communications franchise. Initially, we thought Zoom could leverage this position AI to expand its role within the enterprise. However, the lack of meaningful innovation to date, combined with increasing competitive pressure from Microsoft Teams, has led us to adopt a more neutral view. While we believe Zoom may experience modest growth going forward, we see the business as fundamentally growth-constrained, facing an expanding set of competitive threats and limited pricing power.

This piece is a state of play for investors. We summarize the current moment: the scale of capital being deployed, where real bottlenecks exist, how system-level advances like Blackwell are shifting constraints, how unit economics are evolving, and why enterprise adoption still lags consumer use. The aim is to give investors a coherent snapshot of where AI stands today, what is driving returns, and what factors will determine where value accrues next. A quick aside on valuations: At its peak AOL had a market cap of $220B with 25mm subs valuing each sub at $10k as everyone thought they would be the gateway to the internet. Giving Open AI the same value, it would be at a $10T valuation, it is about 1/20th of that.

公司名字

Seagate Technology（NAS:STX）

主要经营地

全球（数据中心存储供应链）

详细的生意模式

✅ HDD 龙头；核心客户包含 hyperscalers/数据中心产业链

✅ 技术路线推进 HAMR（站内明确点名）以提升容量/成本曲线

护城河

✅ 规模化制造与工程能力（资本密集/良率/供应链）

✅ 技术迭代领先（HAMR）与大客户验证壁垒

✅ 与 hyperscalers 的长期供需关系

估值水平（优先FCF/EV）

✅ FCF/EV（Adj. FCF Yield）：4.8%

✅ EV/EBITDA：18x；P/E：22x

EV/Market Cap

—

作者观点（Thesis 详尽总结）

✅ 核心判断：尽管 SSD 在消费端更强，但在 hyperscalers 的“成本/容量”约束下，HDD 仍将长期是企业级存储关键组件；站内提到 HDD 仍占数据中心存储约 80%

✅ AI 工作负载与机器生成数据推动需求，作者认为股价存在至少 50% 上行空间

⚠️ 该仓位已 Closed，站内披露收益 +52%

公司名字

Burberry（LSE:BRBY）

主要经营地

英国为核心，全球奢侈品经营

详细的生意模式

✅ 奢侈品品牌：以品牌溢价驱动，零售/批发等渠道组合

✅ 通过产品结构升级与品牌定位管理改善长期盈利能力（站内按奢侈品周期+品牌资产分析）

护城河

✅ 历史悠久的全球品牌资产（护城河核心）

✅ 渠道与供应链能力支撑毛利/品牌形象

✅ 定价权（奢侈品商业模式核心）

估值水平（优先FCF/EV）

✅ EV/EBITDA：6.5x；P/E：12.0x

EV/Market Cap

0.96

作者观点（Thesis 详尽总结）

✅ 核心判断：市场在宏观/行业逆风下压低估值，但作者认为公司“长期品牌价值与盈利能力”被低估，存在估值修复与经营改善带来的上行空间

✅ 站内给出 Upside 50%，并标注为 Open

公司名字

Natural Resource Partners（NYSE:NRP）

主要经营地

美国（Appalachia + Wyoming）

详细的生意模式

✅ MLP：持有约 1,300 万英亩矿权，向煤炭承租方收取矿权 royalty（站内估算约 5% 抽成）

✅ 合同含最低付款/最低开采义务等条款，形成下行保护

✅ 另持有 Wyoming 低成本 soda ash 资产股权，并拥有碳封存（sequestration）权利的期权属性

护城河

✅ “矿权 + 合同条款”带来的结构性现金流特征

✅ 低成本 soda ash 资产提供组合分散与估值支撑

✅ 碳封存租赁带来远期可选性

估值水平（优先FCF/EV）

✅ 站内以分部估值/DCF 方式给出：整体至少低估约 30%（并估算 2023 FCF 等）

❌ 未给出统一 FCF/EV 指标表

EV/Market Cap

—

作者观点（Thesis 详尽总结）

✅ 核心判断：煤炭行业虽长期衰退，但冶金煤存在更强韧性；在供给受限背景下，royalty 模式叠加合同最低付款使现金流更稳健

✅ soda ash 股权价值与分红提供安全垫；碳封存权利带来可选上行

✅ 作者给出“至少 30% 低估”的明确判断；后续仓位已 Closed（站内披露 +94%）

公司名字

Meta Platforms（NAS:META）

主要经营地

全球

详细的生意模式

✅ 社交平台矩阵（Facebook/Instagram/WhatsApp）以广告变现为主

✅ 同时投入 Metaverse/Reality Labs（站内把它列为关键变量之一）

护城河

✅ 全球级网络效应与用户规模

✅ 广告主生态与数据/产品迭代能力

✅ 多产品矩阵分散单一应用风险

估值水平（优先FCF/EV）

❌ 站内该页面片段未给出 FCF/EV 或 EV/EBITDA 表格

EV/Market Cap

—

作者观点（Thesis 详尽总结）

✅ 核心判断：尽管争议与声誉问题持续，作者认为当时股价对“高质量业务”定价过低（bargain levels）

✅ 作者把主要不确定性归为三条主线：Apple ATT 隐私变化、TikTok 崛起、Metaverse 投入与回报；若能处理好短期冲击，平台仍具备继续变现与杠杆化的空间

⚠️ 该仓位已 Closed，站内披露收益 +310%