In our last Weekly edition, we discussed why memory and CPU names have been relentlessly bid over the past couple of weeks and highlighted that near-term buying demand may be exhausting in these most heated segments. As we observed over the past week, these names have paused as fast money rotated into less crowded positions. Neocloud and power semis were the two most-discussed segments among investors over the past week. Meanwhile, NVDA was also consistently bid ahead of earnings, which had previously lagged the Mega Cap stocks. The rotation within AI semi/hardware sub-segments has accelerated meaningfully, amid increasing macro headwinds; we believe investors should exercise greater caution in the near term. However, any meaningful near-term pullback or unwind should present opportunities in the long run.
On the Premium Research side, this week, we previewed and reviewed TSEM/AMAT/ASTS earnings and also published the Kioxia preview ahead of Friday’s print. We think the throughline across all of these is the same: AI infra bottlenecks (optics, server CPU, SiPho substrates, NAND ASP, semi equipment) are translating into multi-year demand visibility that the sell-side consensus has not yet caught up to. Please contact sales@funda.ai for institutional access.
This Week’s Reports
Deep|QCOM: Better Late Than Never – Massive CPU TAM Growth Sufficient to Accommodate New Entrants. Agentic AI is shifting the CPU-to-GPU ratio toward 1:1, server CPU lead times have stretched from 1–2 weeks to 8–12 weeks, and Intel/AMD have raised prices — creating the optimal window for QCOM’s Nuvia-derived Oryon core plus the Alphawave acquisition to enter the data center CPU market with both standard and semi-custom CSP offerings. Our channel checks suggest QCOM could begin shipping an AI ASIC similar to LPU to an Asian Internet giant by the end of 2026, and could start shipping general-purpose CPUs to a US client in 2027. QCOM may also be collaborating with two other U.S. CSPs for potential deals.
Deep|QCOM: Better Late Than Never – Massive CPU TAM Growth Sufficient to Accommodate New Entrants
Changes in CPU-to-GPU Ratios and TAM Forecast
Research|LITE: From Cyclical Optical Name to AI Datacenter Bottleneck. LITE’s Nasdaq-100 inclusion looked like a passive flow story on the surface, but last week’s print already validated 1.6T, InP/EML, and OCS demand; today’s move is really the market repricing LITE from a legacy optical cycle stock to a critical AI datacenter interconnect bottleneck asset, with 2.4T / coherent-lite as the next leg.
Research|LITE: From Cyclical Optical Name to AI Datacenter Bottleneck
LITE rallied today, and on the surface it looks like passive flows tied to the Nasdaq-100 inclusion. But the real story runs deeper. Last week’s print already validated the demand intensity in 1.6T, InP/EML, and OCS — revenue, gross margin, and operating leverage all started to come through. So today’s index flow is more of an catalyst; the substance is that the market is beginning to reprice LITE from a legacy optical components cycle stock into a critical bottleneck asset in AI datacenter optical interconnect.
