Community Discussion · Policy

EastVoice · AI · 2026-08-13 · Issue 35

EastVoiceAug 132026/08/12 308 views

EastVoice · AI · 2026-08-13 · Issue #35

Editor's Note: Alibaba open-sourced its flagship Qwen3.8-2.4T weights overnight, DeepSeek shipped V4 Pro to production APIs, and Tencent's earnings showed GPU supply loosening by year-end. China's model race is no longer about who demos better — it's about who ships, at scale, to global developers. OpenRouter's weekly chart says it plainly: Chinese models took the top four slots for the 15th straight week.

Commentary: The quiet story of this cycle is agent infrastructure. ByteDance spun up an "AI Data & Security" first-level department, a former Alibaba P10 founded Pragmatik Labs for next-gen agents, and Huawei Cloud re-positioned its office AI as a general-purpose agent entry point. Model quality was last year's battleground; orchestration is this year's.

I. AI Foundation Models

1. Alibaba Open-Sources Qwen3.8-2.4T-A95B: First Max-Level Weights Ever Released

Summary: Alibaba's Qwen team released the weights for Qwen3.8-2.4T-A95B — a 2.4T-parameter MoE model activating 95B per token, with native 256K context. It's the first time Alibaba has open-sourced a "Max" flagship-class model, letting anyone self-host what was previously only available via API.
Source: ITHome · 2026-08-13
Editor's Take: The open-weight war just escalated. When a 2.4T frontier-class model becomes downloadable, the moat shifts to tooling, alignment, and deployment expertise — not parameters.

2. DeepSeek Ships V4 Pro to Production APIs Overnight

Summary: DeepSeek rolled out the official V4 Pro release to its API (model name unchanged, new DeepSeek-V4-Pro-0813 entry in docs). The update beefs up agent capabilities, adds Responses API support and Codex integration, with internal benchmarks approaching Anthropic's Fable 5 on several tests.
Source: ITHome · 2026-08-13
Editor's Take: DeepSeek's cadence is now "ship to API first, announce later." Pragmatic — and a reminder that in this market, distribution beats spectacle.

3. Tencent Earnings: GPU Supply Set to Loosen by Year-End, Cloud Push Continues

Summary: In its Q2 2026 earnings call, Tencent said cloud GPU resources should become notably more abundant by late this year or early next, lifting cloud service capacity and MaaS opportunities. The company is coordinating compute allocation between internal models and external cloud demand.
Source: ITHome · 2026-08-12
Editor's Take: Tencent's tone suggests China's GPU crunch — the bottleneck that gave us "compute nationalism" — is passing. Cheaper inference across the board is coming.

4. Meta Open-Sources 30B "Lightning" Model, Zuckerberg Namedrops DeepSeek, Qwen, Kimi

Summary: Days after four US states' attorneys general sought $1.4 trillion in damages against Meta, the company open-sourced a 30B model overnight. Zuckerberg explicitly cited DeepSeek, Qwen and Kimi as the open-weight rivals to beat, framing open source as the only way to stay relevant.
Source: Leiphone · 2026-08-12
Editor's Take: Meta's legal nightmare and its open-source pivot are two sides of one strategy: own the open layer, because that's where developer gravity has moved. China's open models forced the move.

5. Alibaba Cloud Launches First Domestic Super-Node Running 2T-Parameter Models

Summary: Alibaba Cloud debuted the Lingjun Zhenwu M890 super-node instance in Ulanqab, claiming it's China's first cloud product that can run 2T+ parameter MoE inference. Each unit is a 64-GPU high-speed interconnect slice, theoretically scalable to 100T-parameter models, no self-built cluster required.
Source: ITHome · 2026-08-12
Editor's Take: "Rent a supercomputer by the hour" is now real. This collapses the infrastructure barrier for Chinese startups trying to serve frontier-scale models.

6. ByteDance Creates "AI Data & Security" First-Level Department — Before It Even Has a 10T Model

Summary: ByteDance reportedly spun up a new first-level department, "AI Data & Security", on par with Seed and Flow, led by former TikTok platform trust chief Wang Yinglei. The move signals ByteDance is betting big on data governance and safety as differentiators — even though its rumored 10T-parameter flagship hasn't shipped yet.
Source: Intelligent Emergence · 2026-08-12
Editor's Take: ByteDance is building the "trust layer" before the model — a striking inversion of the usual order, and a sign that data strategy is now a board-level priority at Chinese AI leaders.

II. AI Software & Applications

1. Huawei Cloud Upgrades OfficeClaw to "OfficeAce": Targeting the Enterprise Agent Entry Point

Summary: Huawei Cloud's office AI agent product OfficeClaw has been upgraded and repositioned as OfficeAce, a general-purpose agent spanning personal and enterprise office work — content generation, file processing, knowledge search, and PPT building. Huawei is betting on agents as the entry point into enterprise AI.
Source: TMTPost · 2026-08-12
Editor's Take: Every Chinese cloud giant is now racing for the same doorway: the enterprise desktop agent. The war isn't about models anymore, it's about which assistant lives in your daily workflow.

2. Midea Joins Qwen App as Exclusive Home-Appliance Partner via A2A Protocol

Summary: Midea became the exclusive home-appliance partner on Alibaba's Qwen app, integrated through the A2A (Agent-to-Agent) protocol. Users can bind Midea's AI agent for product inquiries, home-environment monitoring, energy queries, and maintenance management — all inside Qwen.
Source: TMTPost · 2026-08-12
Editor's Take: A2A is moving from white paper to consumer reality. When your fridge's agent can talk to your AI assistant, the "everything app" starts to mean everything.

3. ByteDance's Doubao Launches Student Plan: 2.5x Free Quota, Pro at ¥38/Month

Summary: Doubao rolled out a verified-student program: 2.5x the free usage quota, plus the ¥68/month Pro plan at ¥38/month. It's a direct shot at the China campus market as AI assistants compete to become students' default tool.
Source: ITHome · 2026-08-12
Editor's Take: Pricing students cheap now = default habit later. The AI subscription playbook is copying the streaming playbook.

4. Unitree Unitree: 18,000 Humanoid Robots Off the Line

Summary: Unitree announced cumulative production of roughly 18,000 humanoid robots — counting only bipedal humanoids, excluding wheeled-base variants. It's a scale number no Western humanoid maker comes close to.
Source: ITHome · 2026-08-12
Editor's Take: The "China can't build humanoids at scale" thesis is quietly dead. 18,000 units means supply chains, quality systems, and real-world failure data Western labs only dream of.

5. Former DJI/Anker Exec Launches Pet Robot Startup, Yuanhe Leads Angel Round

Summary: Liu Liyuan, a former core product lead at DJI and Anker, has launched a pet-companion robot startup, closing an angel round led by Yuanhe Capital. Pet robots are emerging as one of the few consumer hardware categories with real willingness to pay.
Source: Leiphone · 2026-08-12
Editor's Take: Consumer robotics' next act isn't a humanoid in the living room — it's a cheaper, cuter machine that does something people already pay for. Pets are that wedge.

6. Lin Junyang, Ex-Alibaba Youngest P10, Found Pragmatik Labs for Next-Gen Agents

Summary: Lin Junyang — Alibaba's youngest-ever P10 — announced Pragmatik Labs (Yu Yong Ke Ji, "p7k") in Shanghai, backed by Tencent. The startup targets next-generation agents spanning both the digital and physical worlds. Lin left Alibaba five months ago.
Source: ITHome · 2026-08-12
Editor's Take: The "agents that cross into the physical world" pitch is becoming the standard founding story for China's top AI talent. Watch this space — Lin's team built China's first trillion-parameter open models at Alibaba.

III. Humanoid Robots

1. Honor Enters World Humanoid Robot Games, Its Bots Swept Top-6 in a Half-Marathon

Summary: Honor announced its robots will compete in the 2nd World Humanoid Robot Games (Aug 22–26, Beijing Ice Ribbon). Honor's self-developed humanoids previously swept the top six in a half-marathon event. The games feature 50 events: 29 competitive, 21 scenario-based.
Source: ITHome · 2026-08-13
Editor's Take: Phone makers racing humanoids is no longer a curiosity — Honor treats it as brand engineering. Expect more consumer-tech giants to enter the arena.

2. Unitree IPO in Full Frenzy: 0.018% Lottery Odds, 536 Billion Shares Subscribed

Summary: Unitree's Shanghai IPO opened subscription with a 150.80 CNY/share price. 9.78 million retail accounts subscribed for 53.6 billion shares — an over-subscription exceeding 8,000x, triggering the clawback mechanism; the final lottery rate was 0.018%.
Source: Mirror Studio · 2026-08-12
Editor's Take: 8,000x oversubscription for a robot maker tells you everything about the retail appetite for China's embodied-AI story — and about how crowded that trade already is.

3. Bridges: AutoNavi Vets Keep Feeding the Humanoid Supply Chain

Summary: BridgeDP Robotics (Qiao Jie Shu Wu), a general-purpose robot operating system developer, closed a Pre-A+++ round led by China Mobile's industry fund, with Foresight Fund and Shenzhen Capital adding. Funding goes to its RoboCraft AI motion platform and OS commercialization.
Source: 36Kr · 2026-08-12
Editor's Take: The "picks-and-shovels" layer of humanoids — operating systems, motion platforms — is quietly raising big rounds while consumer humanoids grab headlines.

IV. Autonomous Driving

1. Changan & Huawei Sign Strategic Pact on AI Products and Vertical Vehicle Models

Summary: Changan Auto and Huawei signed a strategic framework agreement in Shenzhen, covering AI products, general and automotive vertical foundation models, and scene collaboration. Both chairmen attended — a signal this is a board-level alliance, not a technical trial.
Source: ITHome · 2026-08-12
Editor's Take: Huawei's model business is now a car-industry strategy weapon. Every major OEM signing with Huawei tightens the grip of its full-stack playbook.

V. World Models / Physical AI

1. Shanghai Targets ¥4 Trillion Software & Info-Services Industry by 2030

Summary: Shanghai's 15th Five-Year plan for software and information services sets a 2030 goal of ¥4 trillion in industry scale, positioning the city as an AI-empowerment "main battlefield" and global competition bridgehead, with heavy emphasis on foundational model applications and data economy.
Source: TMTPost · 2026-08-12
Editor's Take: Provincial-scale industrial policy is calibrating around AI consumption, not just production. The next decade of China's AI story will be measured in deployment numbers, not demos.

2. Samsung Adopts Claude for Chip Design Verification: One-Month Cycle Cut to Two Days

Summary: Samsung confirmed it has brought Anthropic's Claude into semiconductor R&D: after ~3 months of Claude Code access for software developers, design verification cycles that took a month now reportedly finish in two days — a 15x speedup at the hardest layer of chip development.
Source: ITHome · 2026-08-13
Editor's Take: AI eating into chip-design verification is the quietest, most consequential adoption story of the year. Whoever owns this workflow owns the semiconductor roadmap.

Frontier of the Physical World | EastVoice

Policy