$微软(MSFT)$   This wasn’t a “one-quarter surprise” story. Azure and Copilot demand stayed strong, supply is the real constraint. The target price was trimmed for one reason: AI capex is higher than expected. But the conclusion didn’t change — more spending doesn’t mean weaker business, it means AI is moving deeper into the core.

$阿里巴巴-W(09988)$  $阿里巴巴(BABA)$   This research note focuses on Alibaba’s latest AI progress, but the real story sits underneath. As Qwen moves from a standalone model into Alibaba’s cloud, maps, and core apps, AI is starting to translate into actual workload demand. The report is less about short-term numbers, and more about whether AI can structurally lift cloud margins over time.

The quarter itself was clean. Growth, margins and cash flow stayed on track. The real variable now is the WBD acquisition — not growth, but whether the deal reshapes the structure.$奈飞(NFLX)$  

$百度(BIDU)$   $百度集团-SW(09888)$   Many international investors still associate Baidu mainly with search and online advertising. Today, the company operates a much broader full-stack AI platform, spanning AI cloud infrastructure, in-house AI chips, large language models, and robotaxi operations. AI cloud demand is accelerating, internal chips are helping lower inference costs, and Baidu’s robotaxi service has surpassed 10 million rides, with improving unit economics.

$Adobe(ADBE)$   FY26E op margin is guided down to 45%, the lowest in years. Is this a sign of real AI monetization… or just heavy spending masked as innovation? (One Chart to Understand below 👇)

$老铺黄金(06181)$   Laopu Gold is still unfamiliar to many investors outside China. Unlike traditional gold jewelers, Laopu sells craftsmanship and brand value, not gold weight. Its fixed-price model reduces exposure to gold price swings and supports premium pricing. The company deliberately limits store expansion and focuses on top-tier locations. Long queues are treated as part of the brand experience, not an operational issue. Despite having far fewer stores than mass-market peers, Laopu delivers very high revenue per store, suggesting strong pricing power. This is a first-coverage view on the business model and long-term logic, rather than a short-term earnings story.

$泡泡玛特(09992)$   Overseas revenue is scaling faster than domestic, margins are higher, and IPs are being replicated across regions and channels. @Tiger_SG 

$MINIMAX-WP(00100)$   MiniMax has just gone public, but this research note isn’t about the IPO hype. What stands out is an AI company that already runs a dual-engine model: consumer apps for scale and data, enterprise services for margins — with most users coming from overseas. The report looks past the listing event and asks a more practical question: how an AI company turns global usage into sustainable revenue after going public.

$小米集团ADR(XIACY)$   $小米集团-W(01810)$   While everyone was watching Tesla and BYD, Xiaomi quietly turned a profit on 109,000 car deliveries in Q3—with a 25.5% gross margin. But the real shift isn’t just in cars. Its internet business hit a record 76.9% margin. Smartphone premiumization held despite memory cost spikes. And its new Miloco AI home system signals a deeper play: not just selling gadgets, but orchestrating an entire “Human-Car-Home” ecosystem. At 22x forward P/E—well below its historical average—the market still seems stuck in the “budget hardware” mindset. Maybe it’s time to see Xiaomi as what it’s becoming: a vertically integrated tech stack with wheels. (One Chart to Under