Stanford MS&E435 Economics of the AI Supercycle | Spring 2026 | Building AI Factories

NVIDIA Corporation operates as a data center scale AI infrastructure company.

Direct beneficiary of continued hyperscaler AI compute capex. The seminar reinforces the supercycle framing and continued multi-year demand for GPUs/accelerators, though it provides no new procurement figures.

Vertiv Holdings Co. (VRT) — provider of thermal management, power and infrastructure systems for data centers.

Thermal and power infrastructure are primary constraints at gigawatt scale. Vertiv is directly exposed to spending on cooling and indirect power-management systems required by AI factories.

ANET is Arista Networks, Inc., a Technology-sector equity focused on networking solutions for data centers and enterprises.

High-speed switching and networking demand scale tightly with large-scale GPU clusters. Gigawatt AI factories imply increased networking spend per site, benefiting Arista's data-center switching and telemetry products.

Eaton Corporation plc operates as a power management company in the United States, Canada, Latin America, Europe, and the Asia Pacific.

Power distribution and electrical gear lead times and capacity constraints align with Eaton’s product exposure. Large AI loads increase demand for medium- and low-voltage distribution, UPS, and related electrical equipment.

Digital Realty Trust (DLR) — a data-center REIT with portfolio exposure to hyperscaler and co-location demand.

If AI-driven demand tightens capacity, large data-center REITs like DLR can benefit from higher leasing activity and buildout. However, REIT sensitivity to interest rates and project timing adds material uncertainty to the return profile.

GE Vernova / GEV (thematic exposure to power-generation and grid equipment).

If accelerated AI load growth drives additional generation and grid upgrades, companies supplying turbines and grid equipment become downstream beneficiaries. The linkage here is thematic and not directly evidenced by the provided excerpts.

Stanford seminar frames an “AI supercycle” centered on hyperscaler AI capex and the buildout of gigawatt-scale “AI factories” (data centers + power + cooling + networking). While the excerpt is introductory (few concrete numbers/ticker mentions), the investable implication is continued, multi-year demand for GPU/accelerator supply chains, AI networking, data-center power/cooling equipment, engineering & construction, and select data-center REITs/utilities—offset by cyclical/valuation and power-availability constraints.

Only a title/body were provided; no transcript, link, speaker names, or concrete technical claims to verify. From the topic (“AI in healthcare,” “open evidence,” “cyber risks”), the most plausible tradable implications are: (1) increased adoption of AI/LLMs in clinical workflow and imaging, (2) stronger demand for healthcare data infrastructure/interop tooling, and (3) heightened healthcare cybersecurity spend due to AI-enabled attack surface and regulatory scrutiny. All conclusions are high-uncertainty pending the actual video content.

Lecture summary (Altman @ Stanford CS153): argues scaling laws continue to deliver emergent capabilities; AI development pipeline (pre-train/post-train/RL) likely needs a rewrite potentially designed by AI; intelligence becomes a utility (like electricity); key risk fork is democratization vs concentration (~20% chance of concentrated outcome); near-term binding constraint is an underappreciated compute shortage, implying structurally rising demand for GPUs/ASICs, networking, data center buildouts, and power/grid capacity.

Stanford HCI seminar discusses modern “play” motivators in games: relaxation, immersion, PvP, and monetization mechanics. Content is academic and UX-focused; investable angle is gaming UX-driven monetization and live-services design rather than core AI infrastructure.

Transcript fragment discusses an “AI going to hyperscalers” thesis: enterprises prefer AWS/GCP/Azure-managed AI stacks versus newer GPU-cloud providers where customers must solve integration/ops. It implies strong forward demand for NVIDIA Blackwell B200 (mention of ~150k units needed in ~12–15 months) and highlights Google’s TPU path plus strong TSMC relationship. Content is partial and noisy; actionable signal centers on hyperscaler capture versus GPU-neocloud margin risk and continued NVDA/TSMC demand strength.

Lecture snippet focuses on LLM inference mechanics—especially KV-cache growth during long-context + tool-call workflows—and resulting systems bottlenecks. Key technical signal: inference scaling is increasingly constrained by memory capacity/bandwidth and storage hierarchy (GPU HBM → CPU DRAM → SSD), not just raw GPU FLOPs. Mentions industry “rumblings” (unverified) about OpenAI buying up SSD/DRAM, and references Nvidia plus emerging inference-focused chips.

Seminar discusses geometric inductive biases applied to robot learning and VLA-style models. Content is academic and architectural; no commercialization timeline or company/product link is given. Tradable implications are indirect — reinforcing demand for compute, edge inference silicon, and robotics stacks.

Seminar covers evolution to native multimodal models (text+vision+audio/video) and directions like sparsity (MoE/conditional compute). Reinforces a medium-term technical direction of larger context and higher-throughput inference, supporting demand for compute, memory bandwidth, networking, and inference-serving infrastructure; actionability is moderate-low and academic in nature.