Browse
Research desk
IC corner
Radar
By Kristal Research Desk
Browse by channel
The next semiconductor battle is not inside the chip. It is around the package.
AI made power scarcity urgent. Bloom has the right machine, but the stock depends on whether related-party revenue becomes clean customer proof.
Power scarcity got Vistra noticed. Contracted cash flow is what gets it re-rated.
Meituan is becoming a different company than the one the market is selling
The market sees AI capex. Tencent is trying to build the tollbooth for China’s agent economy.
Micron is no longer just selling into a memory upcycle. It is trying to turn the cycle into a contract before China and new supply catch up.
We were right that AI would create more content. Q2 showed why turning that content into ARR is harder than we assumed.
Rubrik spent a decade mapping what the enterprise used to be. Q1 and Investor Day asked whether agents can now put it back together.
Stronger company, less forgiving stock, and why gross margin is now the only number that matters
I asked whether Zscaler’s inline architecture was valuable. Q3 forced a harder question: where is cybersecurity value moving?
Oracle found a bridge across the capital chasm. Q4 asked whether shareholders get across too.
The world moved from gate to graph. AI is now forcing the graph to operate at runtime.
We were early on the architecture and uncertain on the conversion. The data started catching up.
AI made Falcon’s data pipeline matter more.
The quarter disappointed the stock, but strengthened the case for Broadcom as AI’s industrial layer.