Reward 500 Tiger-coinsNext week, NVIDIA GTC 2026 opens its doors. Jensen Huang will take the stage again. Over the past few years, each GTC has served as a major market catalyst.
What will he bring this time? Before the real answers are revealed, let's dive deep into Jensen's most important mental framework — the Five-Layer Cake Theory — and how it can guide us toward investment opportunities in this AI wave.
I. The Five-Layer Cake: From Energy to Applications
Jensen breaks down the AI industrial architecture into five layers, from bottom to top — like a five-layer cake:
Layer 1 · Energy
The foundation of everything. Data center electricity consumption is exploding — nuclear, natural gas, and renewables all benefit. Without stable, affordable, large-scale energy supply, everything else is just talk.
Energy stocks are the most underappreciated beneficiaries of this AI cycle. Nuclear power (CEG, OKLO, TLN) is highly sought after by data centers for its reliable baseload electricity. GE Vernova's power equipment orders are also accelerating.
Layer 2 · Chips
The translators that convert electricity into computation.
TSMC is the 'foundational bedrock' of the chip layer — virtually every advanced AI chip is manufactured there. Broadcom's custom ASIC route is being adopted by Google, Apple, and Meta, creating both a challenge and a complement to NVIDIA.
Layer 3 · Infrastructure
Data centers are AI's factories.
Data center construction is the most capital-intensive segment of this cycle. Oracle's deep partnership with NVIDIA and CoreWeave's rapid expansion both signal that the infrastructure layer opportunity window remains wide open.
Layer 4 · Models
The 'brains' running on the infrastructure. Not just language models — also protein AI, chemical AI, physical simulation models, and autonomous systems. Google, Meta, Microsoft, Amazon, and Alibaba are all heavily invested at this layer.
Layer 5 · Applications
The ultimate point of economic value creation. Drug discovery, autonomous driving, legal copilots, industrial robotics — Salesforce, Shopify, Palantir, and Tesla are all competing here. Every successful application pulls hard on every layer beneath it.
The application layer faces the end market most directly and is where AI commercialization is landing fastest. Tesla's FSD, Palantir's enterprise AI, Salesforce's Einstein AI Agent — all represent real product-market fit.
II. GTC Preview: How Should We Think About NVDA's Stock?
GTC has always been NVIDIA's home turf. Each year, Jensen's Keynote delivers a barrage of new products, new architectures, and new partnerships. Markets often exhibit a clear 'buy the rumor, sell the news' pattern around the event.
Key areas to watch this year:
Blackwell Ultra / Rubin Architecture Progress: The timeline and performance metrics for the next-generation GPU architecture are the most closely watched technical catalyst.
Physical AI & Robotics: Jensen has frequently mentioned embodied intelligence over the past year. Progress on Omniverse and the Isaac platform could create new narratives for the robotics sector.
Software & Services Revenue Mix: CUDA ecosystem, NIM microservices, AI Enterprise — whether NVIDIA's software flywheel is accelerating is critical for valuation re-rating
III. What Do You Think About These Investment Opportunities?
The Five-Layer Cake framework lets us examine the investment opportunity in this AI infrastructure revolution from a full-stack perspective, rather than fixating on the price movement of any single company.
• Which layer are you most heavily positioned in right now? NVDA in the chip layer? Tech giants in the model layer? Nuclear stocks in the energy layer?
• Which layer do you think the market is most underestimating? The energy layer is overlooked by many — yet it's the first principle Jensen emphasizes.
• What moves will you make after GTC? Waiting for the GTC catalyst, or already positioned ahead of it?
Share your thoughts in the comments.
Comments
The layer I think the market may be underestimating is energy. AI data centers require enormous electricity, and without reliable power the entire AI stack cannot scale. Companies like Constellation Energy, Vistra Energy & GE Vernova could quietly become major beneficiaries of the AI boom.
As for positioning ahead of GTC, I prefer to stay partially positioned rather than chase hype. Events like GTC often see “buy the rumor, sell the news,” so I focus more on the long-term trends Jensen highlights — especially whether AI expands further into robotics and physical AI. 🚀
@TigerStars @Tiger_comments @TigerClub
I am already positioned in the chip layer with NVIDIA. It is the beating heart of the AI compute layer. NVIDIA doesn't just sell chips. It is a vertically integrated AI compute company.
I am already positioned in the model layer with Alphabet. It is an outstanding key player in this layer as it is the architect of the Model Layer as Gemini is a vertically integrated AI engine.
The most underrated layer is Energy. XLU ETF represents utility companies that generate electricity, transmit it, distribute it & expand its capacity. This is the layer Jensen says that AI is requiring 10x more power.
The real wealth is built by owning the whole cake.
@Tiger_comments @TigerStars @TigerClub
I don’t think the market underestimates energy. It is just unpredictable and the supply and demand can be easily manipulated by the opec countries. It is not easily understood by most investors and so l prefer to stay out of it.
With the unpredictability of the war that makes prices of all stocks volatile and unpredictable, I don think the GTC will be the main reason for price moves. I would prefer to hold off any trades and watch for the effect of the war on stock prices.
@HelenJanet @SR050321 @Success88 @Universe宇宙 @Kaixiang @LuckyPiggie @DiAngel @Wayneqq @SPOT_ON @Fenger1188 come join
Netflix's $600M Move: Netflix acquired Ben Affleck’s AI startup (InterPositive). This isn't about replacing actors but using AI to automate post-production (e.g., removing wires, fixing lighting) to slash costs and speed up content delivery.
NVIDIA (NVDA) remains the dominant player, with the market focusing on the launch of the Rubin architecture and Vera CPUs at GTC 2026.
Hyperscalers like Alphabet, Amazon, Meta, and Microsoft are projected to invest a combined $650 billion in AI-related capital spending in 2026.
Nuclear Power: Large-scale deals, such as Meta's three new nuclear contracts in early 2026, have highlighted the reliance on baseload power.
Demand Shock: AI is driving a projected 25% increase in U.S. electricity demand by 2030, with grid operators struggling to keep up with data center requests.