‌$英伟达(NVDA)$  ‌supply commitments exploded from $119B to $279B in a single quarter with memory procurement driving a lot of that increase. That’s the world’s largest AI chip buyer is effectively signaling this memory cycle could have far more duration than the market still expects.

$先科电子(SMTC)$  Earnings: The AI Data Center Inflection Is Accelerating $SMTC delivered a much stronger quarter than many expected. Revenue came in at $341.9M vs $328M guidance, with EPS at $0.71 vs $0.61 expected. But the real story is not just the beat — it is the acceleration in Data Center. Data Center revenue trajectory: • Q1: $71.6M • Q2: $100M+ • Q3 guidance: ~$145M That represents roughly 45% sequential growth and a significant year-over-year acceleration. Management also highlighted that FY27 Data Center capacity is already fully booked, while FY28 demand visibility has exceeded 70%. This shows hyperscaler demand for AI infrastructure continues to expand. The biggest catalyst: 1.6T optical connectivity. $SMTC reached 1.6T qualification

Over the past two years, investors have focused primarily on AI models, large language models, and software applications. However, a deeper challenge is becoming increasingly clear: The biggest bottleneck for AI is no longer algorithms — it is physical infrastructure. The future of AI depends not only on smarter models, but on the ability to provide massive amounts of computing power, memory capacity, networking speed, and energy. AI is entering a new infrastructure investment cycle, similar to the early stages of the internet. The entire AI ecosystem can be divided into several critical layers: 1. AI Compute — The Engine of the AI Era Training and running advanced AI models require enormous computing power. Key players: ‌$英伟达(NVDA)$  ‌ The l