肖志斌 / Xiao Zhibin

肖志斌 appears in E230｜1万亿收入预期背后：英伟达的巅峰与软肋 as a semiconductor and AI infrastructure guest. He separates Jensen Huang’s $1 trillion demand-side order frame from the harder question of whether [[SemiconductorSupplyChain|the semiconductor supply chain]] can support [[NvidiaBlackwellPlatform|Blackwell]] and [[NvidiaVeraRubinPlatform|Vera Rubin]] volume by 2027.

His source-specific view is that wafer nodes, CoWoS-style Advanced Packaging, High Bandwidth Memory, interconnect, switches, and data-center deployment all matter before demand can become delivered systems. He also argues that single-point inference-chip startups face model and software-ecosystem risk, so durable opportunities may sit in heterogeneous systems, neutral infrastructure, interconnect, and system-level optimization.

Connections

• Nvidia, Nvidia Blackwell Platform, and Nvidia Vera Rubin Platform - order and platform execution context.
• Advanced Packaging, High Bandwidth Memory, Memory Wall, and AI Hardware Supply Chain Pressure - constraints he emphasizes.
• AI Chip Specialization, Low-Latency Inference Chip, and Inference Chip Startup Narrowing - challenger and startup opportunity frame.
• TPU, Google, and Strategic AI Infrastructure Dependence - competitive and vertical-integration context.