Alibaba's CEO says he doesn't see 'much of an issue' with an AI bubble and plans to invest 'aggressively'

Alibaba reported September-quarter revenue of 247.8 billion yuan ($34.8 billion), up 5% year-over-year, while net income plunged 53% to 20.6 billion yuan as heavy AI and commerce-related spending and more-than-doubled sales & marketing costs weighed on profits. The cloud unit—home to the new Qwen platform—grew 34% to 39.8 billion yuan and Qwen exceeded 10 million downloads in its first week; CEO Eddie Wu said AI demand outstrips chip and server supply and signaled plans to invest aggressively, suggesting the previously announced 380 billion yuan AI infrastructure commitment may be understated. The results underscore strong AI-driven revenue growth but near-term margin pressure from large capital and operating expenditures, amid a stock rally of over 86% year-to-date and mixed leadership views on an AI 'bubble.'

Analysis

Market structure: Winners are hyperscale cloud/software owners and AI-hardware suppliers — Alibaba (BABA) and NVDA top the list — plus energy and copper suppliers used in data center builds. Losers include smaller Chinese internet retailers and pure-play data-center builders who face heavy capex and margin pressure; expect pricing power to shift to software/IP owners and GPU suppliers over the next 12–36 months as Big Tech spends ~$320bn this year on capex.

Risk assessment: Tail risks include a regulatory shock (China/US export controls or data-security bans) or a demand unwind that compresses ROI on multi-year capex (380bn CNY+ at Alibaba). Immediate (days) risk = earnings volatility; short-term (3–9 months) = execution and supply shortages (GPU/server lead times); long-term (3+ years) = monetization risk of AI services and stranded assets if models or standards change.

Trade implications: Favor concentrated exposure to AI hardware and cloud platform winners while hedging execution/regulatory risk. Expect tight GPU supply to sustain NVDA margins; cloud revenue growth (BABA cloud +34% YoY) supports a tactical long in BABA with downside protection. Commodities (copper/industrial power) should see a 6–18 month demand boost from buildouts.

Contrarian angles: Consensus underestimates persistent supply-side constraint benefits — hardware vendors likely enjoy above-normal margins for 12–36 months; consensus overestimates return predictability for aggressive capex spenders without differentiated AI IP. Historical parallel: 2000 telecom capex created stranded assets, but today's cloud has stronger software monetization — still, beware malinvestments and sovereign policy shocks.

Key Decisions for Investors

• Establish a 2–3% long position in BABA (equity) sized to portfolio risk with a 12-month horizon to capture cloud/AI monetization; hedge by buying 12-month 15% OTM puts sized at 30% of the position cost or use a 9–12 month call spread to cap premium. Target: +20–40% upside in 12 months; cut position if BABA drops >18% from entry or reports worse-than-expected cloud ARPU next quarter.
• Allocate 1–2% to NVDA via 9–18 month call spreads (buy LEAP call, sell higher strike) to capture continued AI-hardware tightness while limiting premium; take profits on a +50–80% move or roll if NVDA falls >25% intraday. Rationale: GPU demand outstripping supply and industry capex of ~$320bn supports pricing for 12–36 months.
• Initiate a 1–2% pair trade: long BABA vs short KWEB (equal notional) over 6–12 months to isolate idiosyncratic cloud upside vs broader China internet risk. Rebalance if KWEB outperforms/underperforms by >15% or if Chinese regulatory guidance changes within 30–90 days.
• Add 0.5–1% exposure to copper/energy via FCX or a copper ETF (e.g., COPX) with a 6–18 month horizon to capture incremental demand from data center builds; take profits if LME copper rises >15% or if global manufacturing softens materially.
• Monitor (and set alerts for) China export-control announcements, Alibaba cloud ARPU and server deployment cadence, and US chip export policy over the next 30–90 days; if an export-control or large regulatory restriction emerges, reduce China AI/cloud exposures (BABA/KWEB) by 50% within 5 trading days.