The New Silicon Era
The cycle we entered in 2023 is fundamentally different. It is not driven by consumer units shipped; it is driven by infrastructure capability. We are witnessing a capital expenditure (CapEx) super-cycle where the end customer is a trillion-dollar hyperscaler building the intelligence grid of the future. The collective CapEx of the "Hyperscale Five" is projected to exceed $200 billion annually, primarily focused on AI infrastructure.
This course serves as a comprehensive reference manual for the engineer who needs to understand global supply chain logistics, and for the supply chain manager who needs to understand the physics of 2nm manufacturing to make informed procurement decisions.
Course Modules
Module 1: The Convergence of AI and Silicon
- The End of General Purpose Computing: The breakdown of Dennard Scaling and the "Power Wall."
- CPU vs. GPU: Why the "Ferrari" (CPU) lost to the "Bus" (GPU) for matrix multiplication workloads.
- The Perfect Storm: The collision of Algorithmic scaling (Transformers), Manufacturing limits (EUV costs), and Geopolitical friction.
Module 2: The Physical Supply Chain
- Materials Tier: Purity risks (Neon, Photoresists) and the PFAS environmental regulations.
- Equipment Tier (Lithography): ASML's monopoly, EUV physics, and the $380M High-NA transition.
- The Foundry Tier: The "Big Three" (TSMC, Samsung, Intel) and the race to 2nm Gate-All-Around (GAA).
Module 3: The AI Demand Explosion (HBM & CoWoS)
- The Memory Wall: HBM3e vs HBM4, TSV stacking yields, and thermal challenges.
- Advanced Packaging (CoWoS): Why the "Package" is now the bottleneck, not the core.
- Energy Crisis: Managing 1200W TDP chips and the critical supply chain for liquid cooling (CDUs, Manifolds).
Module 4: AI in Manufacturing Operations
- Virtual Metrology: Using sensor data to predict wafer quality and reduce measurement steps.
- Automatic Defect Classification: Using Computer Vision and synthetic data (GANs) for defect detection.
- Generative AI for Recipes: Using Bayesian Optimization to reduce R&D time by 50%.
Module 5: AI-Driven Chip Design (EDA 2.0)
- Reinforcement Learning: How AI agents (like AlphaChip) optimize floorplanning for power and area.
- Verification: Using AI to predict redundant tests and accelerate "Time to Tape-out."
Module 6: Geopolitics & Security
- Export Controls: Understanding TPP (>4800 TOPS) thresholds and the Entity List.
- The "Gray Zone": How chips like the H20 are designed to comply with bans.
- Hardware Trojans: The risk of malicious dopant modifications in 3rd party IP.
Module 7: Future Outlook (2030)
- Glass Substrates: Solving warpage and interconnect density for massive packages.
- Silicon Photonics (CPO): Replacing copper with light to solve the data center energy crisis.
- The Trillion Dollar Forecast: Preparing for the industry to double revenue by 2030.
Who Should Attend
- Engineers: Who need to understand global supply chain logistics and the physical limits of manufacturing.
- Supply Chain Managers: Who need to understand the technical bottlenecks (CoWoS, HBM) driving delays.
- Investment Professionals: Who need to quantify the risks of the "Triopoly" market structure.