Table01

Q1 FY2027 results — $81.6B revenue (+85% YoY) and $91B Q2 guidance — forcefully answered the growth skeptics. The framework shifts from "pricing for quality" to "re-pricing for high-quality growth": upward earnings revisions compress the forward P/E to ~23.5x, so valuation appeal actually improves rather than deteriorates. The thesis has upgraded from "AI chip monopolist" to "AI infrastructure platform company," with the TAM boundary extending from GPUs toward CPUs (Vera), physical AI, and software.

Full-year FY2026 results fully verify the core logic of an "AI + cloud-driven valuation regime switch": cloud revenue accelerated each quarter (26%→34%→36%→38%), AI products grew triple-digit for 11 straight quarters with annualized ARR surpassing RMB 35.8B, international turned near break-even, and AI commercialization entered a "returns-at-scale" phase. But the original report underestimated two things: the scale of investment (the CEO said it will "far exceed RMB 380B," possibly RMB 460-823B over six years) and the profit pressure (FY2026 adjusted EBITA -56%, free cash flow -RMB 46.6B, sharply lower Non-GAAP net income), pushing profit release and FCF turning positive out by 12-18 months. The framework remains a barbell of "AI-transition option + defensive cash flow": Taobao/Tmall cash flow as defense, AI + cloud growth as offense. Composite scores: short-term 5.73 / mid-term 7.08 / long-term 8.23 — i.e., short-term hold/buy-the-dip, mid-term buy, long-term strong buy. 12-month base target $180 (probability-weighted $179); accumulation zone lowered to $110-150.

A pure AI-infrastructure play: revenue up 25x in three years, ARR up 5x in 12 months to $1.25B, a ~$46B Meta+Microsoft backlog providing extreme revenue visibility, plus NVIDIA's $2B strategic investment and a full-stack TCO advantage. But ~92% customer concentration, an ISS governance score of 10/10, the CEO's 52% voting power, extreme 2026 capex of $16-20B, and ongoing dilution make it high-risk. At a TTM P/S of 87x, the valuation depends on executing revenue from $530M to ~$10B within two years. H1 2026 is the key verification window; the strategy is staged, small probing positions — accumulate on a pullback to $95-120 or ahead of catalysts, hard stop $75.

Circle is the leader with the deepest compliance moat and most forward-looking tech roadmap in stablecoins. In Q1 2026 USDC transaction volume surpassed USDT for the first time and RLDC margin improved to 41%, verifying usage depth and margin recovery. But the model is highly dependent on reserve interest (96% of revenue) and the Coinbase distribution agreement (~55% reserve-revenue share, renewing August 2026), for a composite risk score of 7.7/10. The current forward P/E of 95-108x already fully prices optimistic expectations, mismatched with the risk. Rating Hold/cautious accumulate; suggested buy zone $80-90, waiting for a 20-30% pullback to build a margin of safety.

Corning is one of the most fundamentally certain beneficiaries of the AI infrastructure wave: optical communications grew +35% to $6.27B in 2025, the Meta $6B deal and the NVIDIA agreement lock in long-term visibility, and the Springboard plan hit its targets early and was upgraded to add $11B by 2028. But a TTM P/E of 87-90x already fully prices the best case, and with insiders selling 17 times and buying zero, the near-term risk/reward is severely asymmetric. Strategy: avoid short term; mid term wait for the forward P/E to fall back to 30-35x; long term scale in after a pullback (target position 5-8%).

Tencent is the "highest-certainty, lowest-valuation, most generous shareholder returns" name among China's internet giants. 2025 revenue was RMB 751.8B (+14%), Non-IFRS net income RMB 259.6B (+17%), and gross margin rose to a record 56% — high-quality "slower-growth, faster-profit" growth; AI entered a commercialization harvest phase (higher ad CTR, Yuanbao over 100M, Tencent Cloud's first profit at scale). At a TTM P/E of 17-18x it is at the ~20th historical percentile, a 40%-50% discount to global peers, with HK$121B of annual shareholder returns as a floor. The biggest long-term call option is that WeChat's 1.385B MAU has not yet deeply integrated an AI Agent; the core risks are uncertain returns from doubling AI investment and CFIUS/1260H geopolitical review. Suggested core holding 15-20%, hold 3+ years.

Tesla is at a key inflection from "EV pioneer" to "physical AI company," but the core auto business keeps deteriorating: 2025 deliveries were 1.636M (-8.6%, two straight years of decline, losing the global BEV sales crown to BYD), revenue $94.83B declined for the first time annually, GAAP net income $3.79B (-46.5%), and gross margin compressed to 17-19%. The only bright spot is energy (storage +48.7%, revenue +27%, gross margin 29.8%). With a brand-reputation collapse (Axios Harris #8→#95, Yale NBER quantifying 1.0-1.26M lost sales) and 2026 capex raised to $25B+ turning FCF negative, near-term headwinds are multiple. The current market cap is $1.41T at a P/E of 365x, with the auto business only 3-4% of market cap and the other 96% AI/robotics option value; Wall Street targets range $145-$4,600, consensus "Hold." Conclusion: only suitable for high-risk-tolerant investors who believe in AI commercialization to make a small option-like allocation (≤5-10%); value investors should wait for a better $200-250 entry. Composite rating HOLD, cautious wait-and-see.

TSMC is one of the highest-certainty investments of the AI era, but the certainty has a time boundary (2028 is the key inflection). 2026Q1 gross margin was 66.2%, net margin 50.5%, and advanced process 74% of wafer revenue, in a "super window" overlapping AI demand exploding + 2nm effective monopoly + CoWoS sold out through 2027: 2nm wafer ASP >$30K, capacity only 1/3 of demand, Samsung 3nm a yield disaster, Intel 18A external revenue only ~$300M — unprecedented pricing power. A forward P/E of ~27x implies a ~3-5pp geopolitical discount (PEG ~0.6). Core risks: US-fab ramp dilutes gross margin 2-4pp, Samsung/Intel catch up in 2028, and Taiwan-Strait tail risk. Strategy: overweight in 2026-2027 (15-25% of a tech portfolio) to enjoy the certainty, adjust dynamically from 2028 based on the competitive landscape, with gross-margin trends and NVIDIA orders as core exit signals.

Intel is a turnaround bet carrying a unique "geopolitical put," with a highly binary value-realization path: if 18A hits its yield target (~70%) and wins NVIDIA/Apple-grade external customers, it could return to $100+; if 18A slips or yield falls short, it could fall back to $35-50. Fundamentals have bottomed — FY2025 net loss narrowed from $18.8B to $300M, free cash flow improved from -$15.7B to -$4.9B, and DCAI hit a record $5.1B in Q1 2026 (31% margin) on AI inference. But Intel Foundry still loses $10B+/yr with external revenue under 0.2% of the total, x86 share keeps eroding (server revenue share 58.7%, nearly overtaken by AMD), a forward P/E of ~60x already prices a lot of optimism, and at $82.54 the stock is ~20% above the probability-weighted fair value of $68.75. The US government's $11.1B of support (incl. a 9.9% stake) provides a structural floor. Suggested as a 3-5% satellite position, managed with conditional triggers — wait for 18A yield confirmation or a pullback to $60-70 to add.

POP MART is at a key inflection from "high-speed growth" to "high-quality growth." Over 2022-2025, revenue compounded ~100% and net income ~207%, with gross margin reaching 72.1%, adjusted net margin 35.2%, zero interest-bearing debt, and ROE ~77.6%. After the March 2026 annual report, a "no less than 20%" growth guidance triggered a one-week crash of over 35%, compressing the forward P/E to ~11x and PEG to 0.72 — the 5%-10% percentile since listing, below Bandai/Sanrio/Disney. Core view: the valuation over-prices the pessimistic "IP cycle peak" scenario, conflating "deliberate operational rest (an F1 pit stop)" with "demand exhaustion." The biggest call option is overseas revenue crossing 50%, triggering a "Chinese consumer stock → global IP platform" re-rating; the core risks are Labubu's single-IP concentration of 38.1% and new-IP success rate, cultural mismatch in the West, tariffs, and a governance discount from founder selling. Rated Overweight/Buy, 12-month target HK$250 (~+56% upside); accumulate in stages, target position 5%-8%.