Serenity
AI & Semiconductor Supply-Chain Researcher, Trader
Anthropic's accusation that Alibaba's Qwen lab distilled its frontier models is optically bearish for Alibaba, though no penalties have been enforced. Per Anthropic, the lab used thousands of fake accounts and over 28.8 million exchanges. While penalties are absent, the rewards outweigh the risks for Alibaba if distillation lifts Qwen's market share, as many US AI startups already use Qwen for its cost-effectiveness versus Gemini and others. x.com x.com
Futu and UP Fintech look badly exposed as the Chinese government goes after their historical revenues. With no fair legal recourse in China, courts would be rigged against them and an appeal has little chance short of a hidden deal, while their local operations leave them unable to simply ignore penalties. The case, like Meta's Manus episode, shows why Chinese equities from Alibaba to PDD are best avoided even when cheap: the government can nuke or take over any business on a whim. x.com x.com
Alibaba is not a cheap Amazon, and cheap can stay cheap for a long time. It trades at 2015 prices, while since 2015 Amazon has gone from $34 to $250, Alphabet from $39 to $344 and QQQ from $100 to about $770, eleven years of missed compounding. Chinese companies should not be assumed to follow the playbook of American tech giants, just as Tencent is not Alphabet and BYD is not Tesla. x.com
Alibaba is no longer worth owning when better opportunities exist. The case for it as a China hedge was a macro feeling applied to a single company, and it ignored Jack Ma being yanked away and a lack of cloud growth while leaning on share buybacks and looking too cheap to ignore. Despite its discount to US companies, other stocks, such as GitLab and Airbnb, now look much more compelling. x.com
Alibaba is a sleeper AI winner: AliCloud external revenue grew 40% last quarter, and AI product revenue posted its 11th straight quarter of triple-digit growth. Its roughly 5% CXMT stake, bought for 7.6B yuan, is now worth over $20B. The main risk is distillation: Anthropic raised Qwen with Senate Banking in June, and the administration said in April it would act on Chinese labs pulling from US models. Alibaba denies it, but that is a headline waiting to hit. x.com x.com x.com
Alibaba's bull case ignores e-commerce margins being cut in half. A post the author replies to argues the stock is cheap at about 15x forward earnings, with cloud growth accelerating to 36% and management targeting $100 billion in annual cloud and AI revenue. x.com
Alibaba is tough to hold because it is splitting investment between e-commerce and AI, creating massive execution risk. AI is so capex-hungry that, as Western examples show, Alibaba cannot spread spending and still reach comparable results. The market sees China e-commerce spending, including the quick-commerce race, as having no ROI, and is right. Replies argue Alibaba has no choice; it might still make it, but the risk is that its AI push slows. x.com x.com x.com
Alibaba's Qwen models look able to beat costlier US AI: two weeks of testing showed no difference from US AI agents, at 80% less per token. US companies spending hundreds of billions and driving up compute costs risk losing the battle to cheaper Chinese models. A selloff to the low $100s would be a chance to buy much more. x.com
Alibaba is uninvestable as long as the Chinese Communist Party can touch its profits. That lesson came from owning it before, even as @janusrex argues China and Alibaba will deliver great returns in the coming years. x.com
Alibaba's $10.2 billion Hong Kong share placement shows its cash burn has caught up with it, and more dilution is likely. Core China ecommerce makes roughly $30 billion of flat EBIT, half of it reinvested in quick commerce to defend share against PDD. Cloud and AI earns $2 billion of EBIT growing 35% but burns about $10 billion a year in AI-model opex. Pro forma EBIT of about $7 billion collapses after $20-25 billion of annual capex, so Alibaba will keep tapping equity markets. x.com
Alibaba is a contrarian trade because accelerating demand for relatively cheaper compute is very underappreciated, and China's cloud market is still in its infancy versus the US. Alibaba is a consensus short as upside from closed-source AI looks illusory. x.com x.com
Alibaba's HK$80 billion share placement, the largest in Hong Kong history, dilutes shareholders to fund AI just as the market rewards free cash flow and buybacks. Q2 free cash flow was -44.7 billion with capex of 67.7 billion, and the food-delivery war burned 80-100 billion, about what the placement raises. Selling shares after buying back at ~$97/ADS recasts a value stock as a growth bet. The deal was oversubscribed at a modest discount; the test is whether per-share free cash flow rises. x.com … Show more
