Ren
Head of AI, Ex-Product Manager, AI Supply-Chain Bottlenecks
Agentic AI burns 10-100x more compute per session than a chatbot, lifting the whole chip supply chain. Anthropic's ARR rose from $9B to $30B in five months. Alphabet's TPU v8, with Broadcom and Marvell exposure, runs through TSMC, the unavoidable foundry chokepoint, and CoWoS shortages make Intel's EMIB packaging relevant. Micron, SanDisk and SK hynix gain on on-package memory demand, and Applied Optoelectronics, Lumentum, Coherent, Sivers and POET on interconnect demand. x.comÂ
AI infrastructure bottlenecks could give eight stocks as much as 18x upside. Nvidia supplies frontier-model GPUs; Broadcom and Marvell supply custom silicon and networking; TSMC makes advanced chips; Micron supplies sold-out HBM; Vertiv provides power and cooling; Palantir monetizes enterprise AI; Alphabet spans TPUs, research, cloud and apps. The list tracks where AI infrastructure spending is flowing. x.com x.comÂ
Google TPU adoption could broaden the AI hardware opportunity beyond Nvidia, with demand reaching custom ASICs, HBM, packaging, optics and testing. Per The Information, OpenAI reportedly plans to buy Google's TPUs and is hiring compiler engineers, but no purchase is confirmed. Broadcom, SK hynix, Samsung, Micron, TSMC, Coherent, Lumentum, FormFactor and Teradyne could benefit. x.comÂ
Intel and AMD face a real shift in competition as NVIDIA moves into CPUs faster and more directly than expected. On NVIDIA's earnings call, Jensen Huang said its new Vera CPU platform will contribute $20 billion of standalone revenue this year. The pressure on Intel and AMD is not just fund rotation but a genuine change in the competitive landscape. x.comÂ
AI data-center spending flows beyond GPUs and operators into networking, optics, power and cooling. Nvidia, AMD, Broadcom and Marvell supply chips; Arista, Nokia and Credo connectivity; Applied Optoelectronics and Lumentum optics; Vistra, Constellation and Bloom power; Vertiv and Modine cooling. Meta's $140B spending this year illustrates the scale, per a commenter with whom the point is agreed. x.com x.comÂ
A rising value of compute per watt should benefit AI cloud, chip, connectivity and power suppliers. Elon Musk says that value will rise over time. Nebius has Microsoft and Meta deals; Marvell benefits from custom silicon demand; Credo connects GPUs; Vistra has long-term nuclear supply deals with Amazon and Meta; Nokia's Nvidia partnership could help it capture data-center networking growth. x.com x.comÂ
Broadcom's integrated optics, electronics and packaging could lower custom-chip system costs because separate suppliers defend their margins. At similar valuations, it is preferred to Nvidia and has more room to beat expectations. Google TPU v8 shipments are estimated at 4.0M in 2026 and 9.1M in 2027, assuming Broadcom supplies the remainder after MediaTek's allocation. Anthropic financing adds value only if chip profit exceeds likely support payments and capacity costs. x.com x.comÂ
Broader adoption of custom AI silicon could benefit Broadcom and Marvell. OpenAI reportedly plans to expand its use of Google's TPUs beyond Nvidia GPUs, and Google TPUs are coming to Oracle Cloud in 2027. Both are potential tailwinds for custom-chip suppliers. x.comÂ
Architecture changes could lift TSMC wafer demand. Shifting from 12-high to 8-high lifts volumes; an on-device AI phone upgrade cycle would be enormously bullish, while ASIC gains over Nvidia help because ASICs use more wafers. TSMC is arguably more vital to AI buildout than MAG stocks and a reasonable extended-AI-cycle play at a not-crazy multiple, barring big macro events. Margins may not expand as capex rises; a Terrafab partnership, if it happens, would somewhat ease Intel competition risk. … Show more
Inference's move toward edge devices and dense, memory-hungry racks should drive demand for specialized chips, links, power delivery and cooling. McKinsey projects inference at 60% of AI computing demand by 2030. Ambarella makes on-device AI chips; Astera Labs links GPUs, memory and networks; Credo supplies short copper connections; Rambus earns from memory interfaces; Navitas reduces power waste; Modine cools dense racks; and Super Micro builds inference-specific server racks at volume. x.com … Show more
TSMC's capex plans suggest the AI chip trade is not over for semiconductors, Nvidia, AMD and Micron. CFO Wendell Huang cites robust multi-year demand, including agentic AI, behind $60-64 billion of 2026 capex; TSMC expects 2026-28 spending to be even more significantly higher than in the prior three years. Hyperscaler earnings should reinforce the higher AI capex trajectory. CoreWeave-driven fears feel overdone, though the neocloud model does not appeal long term. x.com x.comÂ
