Dawson Lee | Silicon & Infra Alpha

30-yr Optics & Semiconductor Engineer. turned 8x Founder. Now Canadian buy-side investor hunting physical bottlenecks in AI infrastructure. Silicon,

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

https://t.co/K9lsxpQhdi

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

Chroma ($2360.TW) SLT revenue ramping from NT$4B (2025) to NT$9B (2027E) marks the structural shift from single-GPU dependency to platform diversification:

Multi-program ramp: Google/Marvell inference ASIC + Vera CPU + Rubin Ultra in 2H27
Physics driver: 1500W+ module TDP mandates high-precision Active Thermal Control (ATC) liquid handlers
Extended test duration per package directly multiplies tool unit requirements.

[LIWEI_TW Capital @LIWEI_TWCapital]

$2360.TW: Bernstein maps a 2027 CPU and GPU testing ramp

Bernstein expects Chroma ($2360.TW) to see 2027 system-level testing demand from $GOOGL/ $MRVL inference projects and $NVDA’s Vera CPU. It also forecasts Rubin Ultra GPU testing to ramp in 2H27.

That gives $2360.TW several potential program ramps to track, rather than a single AI-chip catalyst

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

The durability argument holds up because hyperscaler HBM 3:1 wafer cannibalization leaves zero margin for the Big 3 to re-enter DDR3/trailing DDR4.

At 1.6x runrate P/E, the key risk isn't pricing, it's PSMC wafer allocation ceilings. If their 19nm tape-out ramps smoothly into 2027, gross margin expansion survives even if foundry input costs climb further. $ESMT

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

6/6 [KV Cache & Physical AI] Beyond training logs, WDC is sampling dual-actuator high-bandwidth HDDs to hyperscalers to evaluate substituting expensive DRAM/SSD tiers for inference KV cache workloads.

Top 7 customers provide demand visibility to 2030, but near-term margin debates and CapEx creep will test valuation multiples in a risk-off tech tape. $WDC

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

1/6 Western Digital ($WDC) reported Q1 FY27: Rev +55% YoY, EPS +161%, gross margin 58.1%, and $1.4B FCF. Yet shares fell -12% across regular and after-hours trading.

Behind the sell-the-news tape: an Exabyte growth gap vs peers, head wafer bottlenecks, and CapEx intensity inflections. Key Q&A takeaways below. 🧵👇

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

5/6 [Head Wafer Physics & CapEx] Management reiterated zero incremental unit capacity additions: bit growth is 100% areal density driven.

However, HAMR magnetic heads require significantly more process steps and longer fab cycle times, creating throughput loss. To support HAMR heads and future 14-platter drives, CapEx will track above the historical 4-6% revenue baseline for the next two years. $WDC

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

@tuolaji2024 熔断倒不至于，但开盘挨锤是定局。昨晚纳指杀的是 AI CapEx 回报率预期，海力士和三星刚打完几百亿美元回购，就看今天本土机构资金在半年线附近敢不敢硬接外资的程序化卖盘。

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

While P&L was strong ($1.24B rev / $0.94 EPS), OCF turned negative this quarter as unbilled inventories and working capital expanded to support hyperscaler deployments.

On-site fuel cell EPC cycles take time to convert into cash. In a crowded trade where whisper numbers were already elevated, institutions simply used the post-earnings liquidity window to de-risk duration exposure. Cash conversion matters more than sell-side EPS beats.

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

彭博下方那行小字才是真正的重磅信号：“OpenAI 的 AI 发现了 PHP 漏洞更深层的完整变体”。

前沿大模型已经具备了在大规模经典开源底座里自主挖掘未知零日漏洞的能力。全球超过 75% 的 Web 站点至今还在跑 PHP，绝大多数根本做不到第一时间升级环境。

前沿实验室在几个月前将漏洞机密通报给 Cloudflare，由其在边缘通过 WAF 部署虚拟补丁（Virtual Patching），直接证明了集中式边缘网络在应对 AI 自动化武器扫描时不可替代的防御价值。

[P Equity Research 📰 @pequityresearch]

Cloudflare CEO, Matthew Prince, on emerging cyber attacks:

"Hackers right now are exploiting this flaw at just incredible rates. Literally, we're seeing millions of attempted attacks using this flaw now across all of our network, every single minute of every single day. The flaws now known and if you are running PHP on your servers, you need to update PHP to the latest version. It's free, you can get it updated immediately, and that's the only thing that's going to keep you safe."

"Most of the frontier labs are working with companies like cloudflare in this way that when they found this flaw a couple months ago, they reached out to us and they said we've found this. They disclose it to us in a in a confidential way."

Via @technology

$NET

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

@pequityresearch OpenAI's models uncovering PHP zero-day variants marks a structural shift in offensive security. Because PHP still powers over 75% of web backends, automated bot scans hit millions of attempts per minute. Cloudflare's edge WAF is serving as the global virtual patch.

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

讲透了自有资金的负债端优势。

对冲基金减仓是因为有每月净值考核、VaR 限额和 LP 赎回压力，必须机械性压降总敞口（De-grossing）；个人自有资金没有被强平的风险，在估值被砸到历史低位时盲目抄对冲基金的作业，反而丢掉了长线的时间套利优势。

卖耐克加谷歌是漂亮的防御升级，只要标的资产负债表和自由现金流足够硬，耐得住波动，反弹时的赔率自然会奖赏有耐心的钱。

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

2027 年即使供给缺口从 20% 收窄到 10%，依然是绝对的卖方市场。

存储芯片的需求价格弹性极低，历史上只要有 5% 的真实缺口，合约价就能一路跳涨、毛利率稳在 50% 以上。海力士 M15X 和美光新厂的大规模产能在 2027 年底之前根本放不出来，再加上 ASML 光刻机百台缺口的物理压制，三大原厂在 2027 年全年的定价权非常稳固。

[P Equity Research 📰 @pequityresearch]

Expert Takeaways: Memory

> S/D gap of at least 20% in second half of 2026.

> Big vendors have just begun forecasting since demand for 2027 has not settled yet.

> S/D gap for 2027 is possibly around 10%. A pessimistic take believing in balance by end of next year.

$MU $DRAM $EWY $SKHY

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

@pequityresearch A 10% S/D deficit in 2027 is still an aggressive seller's market. Because memory demand is price-inelastic, even a 5% shortfall maintains 50%+ gross margins. With cleanroom additions gated until 2028 and DUV backlogged, suppliers retain structural pricing power.

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

Patrick Moorhead 点出了晶圆厂资本开支最关键的账本：钢架外壳（Truss）花不了多少钱，真正的资本绞肉机是 High-NA EUV 机台进场后的折旧计提。

俄亥俄直接跳过 18A 切入 14A 是务实的，但这背后也是代工商业化的死结：外部 Fabless 巨头（高通、苹果、英伟达）要看稳定的 D0 缺陷率和实机良率才敢签大单，而英特尔要等大客户预付款才敢装机台。2026 下半年到 2027 上半年这个 14A 决策窗口，决定了俄亥俄到底是一座世界级先进制程基地，还是只是一座昂贵的钢结构机房。

[Patrick Moorhead @PatrickMoorhead]

. @Intel_Foundry has the first major truss in at Ohio One, and Arizona Fab 62 tool install is fully underway for operations by the end of 2028.

The headlines say Ohio is five years late. I said back in 2024 that the original date was overly ambitious for a first fab in a new state. Pacing it to customer demand is the right discipline.

What matters now: Ohio is designated for 14A and beyond, and @intel says it can pull the schedule in if customers commit. We are inside the 14A decision window @LipBuTan1 laid out, the second half of 2026 into the first half of 2027.

The truss is the shell. Customers set the clock. $INTC

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

@PatrickMoorhead @Intel_Foundry Building the shell while gating tool install to commitments avoids idle EUV depreciation. But it also highlights the foundry dilemma: fabless customers hesitate to commit without high D0 yield data, while Intel waits for prepayments to tool the cleanroom. 14A is the real test.

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

8/8
Summary and Valuation Divergence:

The market is pricing Nokia through the lens of a legacy, low-growth telecom equipment vendor facing component inflation.

Underneath, the company is building a captive Western supply chain for 6-inch indium phosphide optical engines, backed by a €5B hyperscaler order book expanding at 25% to 30%.

As Fab 2 reaches steady-state yields in 2027, operating leverage will decouple high-speed optical hardware earnings from the broader telco cycle.

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

1/8
Nokia $NOK dropped 9.72% post-Q3 2026 earnings.

The tape reacted to gross margin dilution in Network Infrastructure (43% down to 41%) and component inflation.

Behind the quarterly miss lies an inflection in custom optical silicon and AI datacenter interconnect. Here is the institutional breakdown of the Q3 transcript. 🧵👇

[Dawson Lee | Silicon & Infra Alpha @LiDaosen]

7/8
Telco Stagnation and Portfolio Rationalization:

Traditional radio access network (RAN) demand remains flat globally.

Management confirmed no evidence of inventory pre-buying from wireless carriers. Cash generation is supported by internal restructuring:
- Reaching the upper ceiling of the €800M to €1.2B structural cost reduction target.
- Accelerating the integration of the China joint venture from 36 months down to 12-24 months to cut overhead in low-share geographies.
- Channeling savings directly into Fab 2 InP expansion and Co-Packaged Optics R&D.