Alibaba Qwen Takes Over Taobao Shopping.

Mazen Al Sabbagh

— 6 min read
Share

TL;DR

  • Agentic Commerce Scale: Alibaba is deploying its Qwen AI across Taobao and Tmall, creating the world's largest agentic commerce system with access to over 4 billion products for conversational shopping and end-to-end transactions. (Reuters, The Next Web)
  • Infrastructure Resistance: Fourteen US states are now weighing legislation to halt or pause new data center construction, reflecting intensifying grassroots opposition to AI infrastructure and creating tension between community interests and capital investment. (CNBC, Tom's Hardware)
  • Key Trial Testimony: The Musk v. Altman trial enters its critical third week, with Microsoft CEO Satya Nadella and OpenAI cofounder Ilya Sutskever slated to testify, preceding closing arguments and an anticipated liability verdict. (The Verge, The Guardian)
  • Labor-AI Profit Dispute: Samsung chip workers have rejected a substantial bonus offer and threaten an 18-day strike, seeking a greater share of the AI-driven profit surge that recently pushed Samsung's valuation past $1 trillion. (Tom's Hardware, Financial Times)
  • Job Redefinition Debate: A CNN analysis suggests AI is restructuring rather than replacing jobs, yet Anthropic's head of Claude Code foresees the "idea of software engineering to go away" by year-end, highlighting conflicting views on AI's labor market impact. (CNN)

Lead Story: Alibaba Qwen Takes Over Taobao Shopping

Alibaba is integrating its Qwen AI platform directly with Taobao and Tmall, its primary consumer marketplaces. This deployment, reported by Reuters, grants Qwen access to over four billion products, enabling agentic commerce where consumers transact conversationally rather than via traditional search. The Qwen-powered assistant within Taobao will facilitate virtual try-ons, price tracking, and tailored recommendations, indicating a shift in the core e-commerce interaction model. (Reuters, The Next Web)

This system extends beyond mere search enhancement. Qwen's "skills library" manages logistics, customer service, and post-sales processes. Native Alipay integration permits the agent to execute end-to-end transactions autonomously. A recent demonstration by Alibaba Group VP Wu Jia illustrated Qwen autonomously ordering and paying for goods, inclusive of loyalty discounts, with subsequent delivery. This showcases a complete transactional agency.

The distinguishing factor here is scale. Qwen currently serves 300 million monthly active users across Alibaba's platforms. The Next Web reported 140 million initial AI shopping experiences during the Chinese New Year campaign. Alibaba's strategic pivot, articulated as a move "from intelligence to agency," reflects the industry's broader transition from conversational interfaces to autonomous task execution.

A structural divergence in e-commerce AI is evident between Chinese and Western markets. Alibaba leverages a vertically integrated stack—model (Qwen), marketplace (Taobao/Tmall), payment (Alipay), and logistics (Cainiao). In contrast, Western platforms are fragmented. Amazon employs AI for internal search optimization but approaches full shopping autonomy cautiously. Efforts from Google (Remy), Meta (Hatch), and OpenAI remain in experimental phases, none commanding comparable transactional depth.

This launch precedes the imminent Trump-Xi summit, where AI governance features prominently on the diplomatic agenda. While regulatory debates in Washington often center on model pre-release controls, China's leading corporations are integrating advanced AI directly into consumer commerce at population scale. This highlights differing strategic velocities and regulatory philosophies. (Economic Times)

In Other News

The Anti-Data-Center Rebellion Goes National. The opposition to AI data center construction has evolved beyond localized disputes into a coordinated, bipartisan movement spanning 14 states. The National Conference of State Legislatures reports that states from Oklahoma to New York are considering bans or moratoriums on new builds. CNBC notes projected global AI data center spending could reach $7 trillion by 2030, yet host communities increasingly resist. Flashpoints include Maine's legislature passing a statewide ban, Michigan municipalities enacting moratoriums after a $16 billion Oracle/OpenAI facility proceeded despite local opposition, and Florida's Governor DeSantis signing SB 484 to regulate but not ban construction. Data center project cancellations are on track to surpass 2025 records. (CNBC, Tom's Hardware, Heatmap News, Boston Globe)

Musk-Altman Trial: Decisive Week Ahead. The Musk v. Altman trial resumes Monday in Oakland, featuring testimony from critical witnesses: Microsoft CEO Satya Nadella and OpenAI cofounder Ilya Sutskever. Closing arguments are anticipated by May 14, with a liability verdict expected around May 21. New Microsoft emails from 2018, surfaced by Winbuzzer, reveal internal skepticism about OpenAI's AGI claims and concerns about being treated as a "bucket of undifferentiated GPUs." Nadella's testimony will address the nature of Microsoft's relationship with OpenAI, while Sutskever's perspective as a former cofounder and key figure in Altman's 2023 ouster could significantly impact the mission-drift and safety arguments. Altman's potential testimony remains unconfirmed. (The Verge, Winbuzzer, MIT Technology Review, Local News Matters)

Samsung Chip Workers Demand AI Profit Share. Over 30,000 members of the National Samsung Electronics Union are threatening an 18-day general strike, scheduled from May 21 to June 7. This follows their rejection of a $340,000 one-time bonus. The union demands annual profit-linked payouts, citing Samsung's recent surge to a $1 trillion valuation driven by AI-accelerated demand for HBM memory chips. Samsung's Q1 operating profit exceeded its entire 2025 forecast in one quarter. A recent union rally already caused a 58% drop in night shift production. Tom's Hardware estimates a strike could cost Samsung up to $11.7 billion, exacerbating the critical memory chip shortage projected to last into 2027. (Tom's Hardware, TechCrunch, CNBC)

X / Social Pulse

CNN's weekend analysis, titled "AI isn't actually 'taking' your job. Here's what's happening instead," directly challenged recent job data indicating AI-attributed layoffs leading categories for a second consecutive month. Citing a Microsoft survey of 20,000 employees, the piece posited that AI automates job components rather than eliminating roles entirely. However, the most debated commentary emerged from Boris Cherny, head of Claude Code at Anthropic, who stated his expectation for "the idea of software engineering to go away" by year-end. This juxtaposed narrative—reassurance versus radical transformation—garnered significant social media traction.

The recent All-In Podcast episode, widely summarized across Substack, featured investor Brad Gerstner asserting that AI company revenue is primarily constrained by compute supply, not market demand. Gerstner contended that boundless power and data center access would enable "even more extreme" revenue trajectories for firms like Anthropic and OpenAI. This argument shifts the prevailing AI valuation discourse from demand-side uncertainty to foundational infrastructure scarcity.

The 2018 Microsoft emails, unveiled during the Musk-Altman trial, notably CTO Kevin Scott's assessment of OpenAI treating Microsoft "like a bucket of undifferentiated GPUs," circulated extensively among developer communities. The irony of this statement, now often applied to Nvidia's market dominance in compute infrastructure, was a recurring theme.

One to Watch

Google I/O: Pressure Mounts for Strategic Clarity. Google I/O on May 19 is poised to be a pivotal developer conference. Polymarket indicates high anticipation for a Gemini 3.2 announcement, with substantial trading volume on its release. Internally, Google has reportedly ceased its Mariner agent project to prioritize "Remy," anticipated to debut at I/O as a direct challenger to Claude Code and OpenAI's consumer agents. An "Auto Browse" agentic web feature is also in testing. The competitive landscape has demonstrably evolved since I/O 2025: Anthropic now leads OpenAI in Q1 2026 LLM revenue, Claude Code and Cowork accelerate enterprise adoption, and GPT-5.5 shows performance parity. Google's dual strategy, funding Anthropic via a $200 billion cloud deal while generating $20 billion quarterly from its own Cloud, suggests it finances both sides of the innovation race. A strong I/O showing, particularly in agentic and multimodal capabilities, would validate Google's full-stack advantage across model, cloud, devices, and search. Conversely, a subdued performance would perpetuate the narrative of Google's AI endeavors consistently trailing. (Polymarket, Business Insider)

Quick Hits

  • AI Note-Taker Liability: AI-powered meeting transcription services are generating significant legal liability concerns. Attorneys caution that recorded conversations, including privileged or informal remarks, may be subject to subpoena, with no AI firm yet issuing transparency reports on law enforcement data requests. (NYT)
  • Florida Data Center Regulation: Florida Governor DeSantis signed SB 484, implementing data center regulations requiring environmental and utility disclosures. This measure stops short of an outright ban, contrasting with the failure of his more expansive AI Bill of Rights proposal to pass legislative review. (Florida Phoenix)
  • Software Engineering Evolution: Anthropic's head of Claude Code posited to CNN that the foundational concept of software engineering will diminish by year-end. He anticipates AI will reorient the role from direct coding to orchestration and management of AI agents for code generation, testing, and iteration. (CNN)
  • Extended Chip Shortages: Both Samsung and SK Hynix forecast AI-driven memory chip shortages extending into 2027 or beyond. The lead time for new semiconductor capacity, typically two to three years, lags significantly behind accelerating AI demand. (CNBC)
  • Compute as Revenue Bottleneck: Investor Brad Gerstner argued on the All-In Podcast that AI revenue is fundamentally supply-constrained, not demand-limited. He suggested that with infinite compute, frontier labs could achieve "even more extreme" revenue growth, reframing the AI valuation discussion toward infrastructure scarcity. (Dr. Wil's Ruminations)

The weekend's key developments consistently highlight AI's transition from an anticipated technology to an active force structurally reshaping global markets. Friction points are emerging across multiple layers: Alibaba's aggressive deployment of agentic commerce at an unprecedented scale; the nationwide pushback against the data center infrastructure powering these systems; labor disputes over the distribution of AI-generated profits among essential semiconductor workers; and critical testimony scheduled in the Musk-Altman trial, which will scrutinize foundational strategic decisions made at the genesis of a dominant AI entity. The simultaneous construction, contestation, and litigation of the AI era's core architecture underscore its rapid and complex evolution.

Sources

Lock in. M. mazen@thorterminal.com

Read more