NVIDIA’s GTC 2026 (March 16–19) will likely be one of the most important AI infrastructure events of the year. It is expected to highlight the transition from AI training → AI inference → real-world AI systems. Here is how to think about the key points.
---
1. Breakthroughs and AI use cases that matter most
Three themes are likely to dominate GTC:
1. AI inference efficiency (the next phase of AI)
The industry is shifting from training large models to running them at scale. NVIDIA’s new architectures aim to drastically reduce the cost per token and power consumption for inference workloads.
Why this matters:
Chatbots, AI agents and copilots need millions of inference requests per second.
The winner will be the company that delivers the lowest cost per AI query.
2. Agentic AI and autonomous systems
GTC is expected to emphasise AI agents, robotics, and physical AI, where AI systems interact with the real world.
Key use cases:
autonomous factories
humanoid robots
autonomous vehicles
AI copilots that perform complex tasks
3. AI factories (data-centre scale computing)
The future of AI computing is rack-scale systems rather than individual GPUs. Rubin systems combine GPUs, CPUs, networking and memory into integrated AI supercomputers.
---
2. How Rubin and Feynman reshape the AI supply chain
Rubin (2026)
Rubin is NVIDIA’s next-generation AI GPU architecture launching in 2H 2026, using TSMC 3nm and HBM4 memory.
Major supply-chain impact:
1. Massive demand for HBM memory
Rubin relies heavily on HBM4 memory.
Beneficiaries: SK Hynix, Samsung, Micron.
2. Rack-scale AI systems Rubin integrates multiple components:
Vera CPU
Rubin GPU
NVLink 6 networking
BlueField DPUs
high-speed Ethernet fabric
This pushes the ecosystem toward AI supercomputer racks, benefiting:
Supermicro
Foxconn
Quanta
Dell
3. GPU density explosion A Rubin rack can deliver exaflop-level AI performance, massively increasing compute density.
---
Feynman (2028)
Feynman is the successor to Rubin, expected around 2028.
Its direction:
next-gen process nodes
more efficient AI inference
deeper integration with NVIDIA CPUs
The long-term goal is AI systems that can run billions of AI agents simultaneously.
---
3. Will GTC trigger another NVIDIA rally?
Historically, GTC often drives sentiment rallies for NVIDIA. But the reaction depends on expectations.
Bullish scenario
Surprise chip announcement
Strong inference roadmap
confirmation of massive hyperscaler demand
Jensen Huang has hinted that a new chip may “surprise the world” at the keynote.
Bearish scenario
No major surprise beyond Rubin roadmap
investors worry about AI capex sustainability
“sell the news” after strong run-up
---
✅ My base view
Short term:
GTC can trigger AI infrastructure hype again.
Long term:
The real catalyst is whether Rubin drives another wave of hyperscaler AI spending (2026–2027).
If hyperscalers keep building gigawatt-scale AI data centres, NVIDIA could still remain the core AI infrastructure monopoly.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
