China Desk — 20 Aug 2026, 08:50 UTC

Investigation of comedian exposes China's narrowing cultural space

# China Digital Times 4h

Chinese online reaction to an investigation of comedian Guo Degang treats a joke about a revolutionary song as political overreach and a return of red-line censorship.

Wuhan's culture and tourism bureau is investigating Guo and his theater group after he improvised lyrics to a song associated with a 1956 revolutionary film.
The reported trigger was changing a lyric about Weishan Lake to one about the Forbidden City and adding a mangled Buddhist mantra for comic effect.
Officials cited rules requiring performance content to be submitted for prior approval, showing how administrative licensing can reach improvisational speech inside a theatrical performance.
Commenters reportedly described the move as a politically motivated overreaction and compared it with earlier punishment of comedians, but these are online reactions rather than verified public opinion data.
The episode shows how censorship works through anticipatory compliance: performers must avoid not only explicit political criticism but also irreverent treatment of state-sanctioned cultural memory.
Guo's fame makes the case unusually visible, yet the investigation's chilling effect may be felt most by smaller theaters that lack his audience and bargaining power.
The key uncertainty is whether the bureau issues a formal penalty or quietly closes the case; either outcome will signal how far cultural regulators intend to police parody.

Open at China Digital Times Ask Claude ChatGPT

China publishes a trust standard for instant-retail agents

# Leiphone 3h

China's national communications institute and Taobao Flash Sale published a sector standard that sets measurable accuracy, interface and security targets for shopping agents.

The standard covers functional completeness, execution accuracy and efficiency, plus data, access and operational security, vulnerability management, service management and ethics.
It calls for accuracy above 95% on simple tasks, above 90% on complex tasks and interface success above 99%, with at least 100 test cases for each evaluation class.
The document was co-written with companies including Ant, Baidu, MiniMax and several security firms, making it an industry-coordination instrument rather than a statute.
It follows a May national opinion calling for agent standards and protocols, showing how China is moving agent governance into sector-specific benchmarks.
The practical significance is that agents ordering food, goods or services will be judged on execution reliability and auditability, not only conversational quality.
The targets are self-regulatory technical requirements until regulators attach them to licensing, procurement or enforcement.

Open at Leiphone Ask Claude ChatGPT

China's Zhuque-3 first stage completes a land landing

# Solidot 1d

LandSpace successfully launched Zhuque-3, placed a satellite in orbit and landed its first stage on land, making it China's first orbital-class stage recovered by landing legs.

The rocket lifted off at 7:35 and landed its first stage six minutes later at the Zhuque-3 landing site in Gansu, according to the report.
The vehicle is 66 meters tall and 4.5 meters wide, using nine Tianque-12 engines on the first stage and one Tianque-15 on the second.
The mission was Zhuque-3's second flight and reportedly delivered Hongqing Technology's Honghu-03 satellite to its planned orbit.
The landing makes Zhuque-3 the fourth flight-proven partially reusable orbital rocket type after Falcon 9, New Glenn and China's Long March 10B.
The engineering significance is repeatability: a second flight and a controlled land recovery move China closer to a high-cadence launch model rather than a one-off demonstration.
The next proof points are stage refurbishment, reflight, turnaround time and whether the recovery penalty can be justified by launch economics.
This is a strategic infrastructure milestone because reusable launch lowers the marginal cost of deploying large satellite constellations and increases pressure on China's launch ecosystem to industrialize.

Open at Solidot Ask Claude ChatGPT

Beijing blocks Chinese firms from helping the EU probe JD.com

# SCMP 19h

China's Ministry of Justice ordered Chinese entities not to assist an EU foreign-subsidy investigation into JD.com's proposed acquisition, turning a competition case into a sovereignty dispute.

Beijing called the EU investigation an improper extraterritorial measure and said no organization or individual may execute or assist it.
The EU probe concerns JD.com's proposed takeover of German electronics retailer Ceconomy, according to the report.
The order creates a direct compliance conflict for companies, auditors and advisers that may face obligations under both Chinese law and EU investigative rules.
The measure signals that Beijing is willing to use blocking authority to defend Chinese companies against foreign subsidy scrutiny, not only to manage data or export controls.
For future Chinese acquisitions in Europe, due diligence will have to model regulatory discovery and evidence-sharing risk as part of the transaction itself.
The immediate practical question is whether the EU escalates against JD.com or intermediaries, and whether Beijing's order becomes a template for other foreign investigations.
This is a structural clash over whose regulator can compel information from a cross-border deal, not merely a disagreement about JD.com's market conduct.

Open at SCMP Ask Claude ChatGPT

Researchers observe vacuum fluctuations enhancing superconductivity

# Solidot 24m

A Chinese-led experiment used a terahertz dark cavity to raise the superconducting transition temperature of layered niobium diselenide by up to 5.4%, turning vacuum fluctuations into a tunable material control knob.

The team placed six-layer niobium diselenide inside a split-ring terahertz resonator that reshaped the electromagnetic environment.
The cavity increased the superconducting critical temperature by as much as 5.4% and also strengthened critical current and critical magnetic field near the transition.
The result is significant because it observes an effect usually treated as a quantum-electrodynamic subtlety at a macroscopic condensed-matter scale.
A cavity is not a new bulk material or a room-temperature route; the immediate achievement is controlled environmental engineering of a quantum phase.
The main scientific question now is whether the effect survives across materials, geometries and larger devices.
If it does, engineered electromagnetic environments could become a practical method for tuning superconducting electronics rather than merely studying them.
The report calls this the first experimental observation of vacuum-fluctuation-enhanced superconductivity, but independent replication will determine how durable the claim is.

Open at Solidot Ask Claude ChatGPT

Chinese team entangles quantum memories across 420 kilometers

# Solidot 22h

A University of Science and Technology of China team reports entanglement between cold-atom quantum memories over 420 km of fiber, extending the building blocks of an intercity quantum network.

The experiment exceeded 230 km, the report's stated theoretical limit for direct entanglement distribution without a repeater.
The memories are physical quantum nodes rather than only photons, making the result relevant to repeater-based communication and distributed quantum computing.
The team reports a progression from 50 km in 2020 to a three-node metropolitan network in 2024 and high-fidelity hundred-kilometer links in 2026.
The distance is a network milestone, not a complete quantum internet: fidelity, rate, memory lifetime and repeater operations determine practical usefulness.
The result strengthens China's position in long-distance quantum networking, an area where sustained system engineering matters as much as a single laboratory distance record.

Open at Solidot Ask Claude ChatGPT

Jefferies test ranks Alibaba's Qwen Office first among eight agents

# AI QbitAI 7h

A Jefferies office-task test put Alibaba's Qwen Office ahead of Claude Cowork and Codex, emphasizing harness quality and cost per completed task over model branding.

Also: Leiphone

The test used five tasks covering document synthesis, web research, browser control, English presentation creation and image-based marketing design.
Qwen Office was the only product reported to score above 90 in every category, with particular strength in browser control and multimodal generation.
The report separates model capability from the harness around it: prompts, context, tools, guardrails, feedback, correction and governance.
It says Qwen 3.8 Max pricing is materially below some overseas models, making cost per task a more relevant enterprise metric than token price alone.
The test is a small, sponsor-sensitive sample rather than a universal ranking, but it captures a real shift toward workflow integration and execution reliability.

Jefferies ranks Alibaba's Qwen Office ahead of Claude Cowork and Codex Leiphone

Open at QbitAI Ask Claude ChatGPT

MORPHI demonstrates a 15-minute household robot task loop

# AI QbitAI 7h

MORPHI's KINO robot completed cleaning, refrigerator restocking and laundry tasks in one continuous demonstration, while its MoRA architecture pushes memory and progress tracking into the action model.

The demonstration required navigation, object recognition, force-controlled grasping, appliance buttons, doors and multi-step task sequencing in a constrained home layout.
MORPHI argues that the usual high-level reasoning model plus reactive policy split leaves the action layer without enough memory, goal persistence or recovery behavior.
Its Agentic-Native policy model is designed to accept text or image goals, maintain short-, medium- and long-term memory and emit progress and future-state estimates alongside actions.
The useful technical claim is architectural: long-horizon reliability may require the low-level policy to understand task state rather than wait for a large model to replan every subtask.
As with most conference demonstrations, the missing evidence is failure rate across homes, object variation and unattended operation.

Open at QbitAI Ask Claude ChatGPT

Embodied-AI data platform goes on sale for about $720

# AI QbitAI 1d

51WORLD launched a hardware-software platform that captures human demonstrations, filters bad data at the edge and feeds validated trajectories into robot training and simulation.

The first AperData kit is priced at CNY 5,100, about $720, and combines a multi-camera capture device with a data-governance platform.
Its four-camera configuration reports more than 270 degrees of horizontal view, hardware timestamp synchronization and trajectory physical consistency up to 99%.
The important workflow change is capturing people directly rather than occupying expensive robot bodies with teleoperation during every demonstration.
The platform screens blur, exposure, dropped frames, synchronization and incomplete tasks at the edge, reducing wasted upload, storage and cloud-compute costs.
The vendor claims more than a tenfold production-efficiency improvement over traditional robot teleoperation at equivalent cost, a figure that needs independent field validation.
A second platform links reconstruction, training, evaluation, deployment and operations, showing the emerging market for embodied-AI infrastructure below the model layer.

Open at QbitAI Ask Claude ChatGPT

World Robot Conference asks who can follow Unitree to market

# AI QbitAI 2h

A second Chinese report argues that robot valuations are moving toward commercial delivery, but its candidate-company list is largely an industry promotional showcase.

Also: GeekPark, Leiphone, IT Home

The article repeats Unitree's first-day rise of 629.44% at the open and 460.34% at the close, alongside the conference's 300-plus exhibitors.
It emphasizes Deep Robotics' claimed inspection deployments, LimX's motion-control stack and Reeman's hardware-plus-remote-labor model.
RobotEra is presented as a model-and-body company using VLA and world-action models for warehouse and manufacturing tasks.
The useful signal is the evaluation criterion: investors increasingly want repeatable field work, not only dexterity demonstrations.
This is the same Unitree listing event as items 13 and 35; the report is retained because it supplies a different sector framing.

Open at QbitAI Ask Claude ChatGPT

Perfect World says new game has generated $280 million globally

# Culture QbitAI 7h

Perfect World's new game The Perceiver generated more than CNY 2 billion, about $282 million, by August 18, but launch marketing costs left the company in a first-half loss.

First-half revenue was CNY 2.751 billion, about $388 million, while net loss attributable to shareholders was CNY 118 million, about $16.6 million.
The company says cash flow from operations rose 29% to CNY 647 million, about $91 million, and that the game's contribution will appear progressively from the third quarter.
The accounting mismatch is familiar in live-service games: marketing is expensed early while player spending is recognized over the expected player life cycle.
The game launched across more than 180 countries and six device categories, with official PC and Android channels contributing nearly 60% of reported revenue.
The launch is a meaningful test of whether Chinese studios can build global, multi-platform live operations rather than rely on a single domestic release.

Open at QbitAI Ask Claude ChatGPT

A Tencent recommendation contest tests a unified architecture

# AI QbitAI 1d

A Tencent-backed KDD Cup challenge drew 13,913 contestants from 52 countries and found that a unified recommendation block could raise ranking quality while cutting inference cost.

The industrial winning system reportedly improved AUC by 0.0048 while reducing per-sample compute by 18%, a modest-looking change that scales materially at billion-request volumes.
The winning academic design used an intent token to carry user-interest information across layers while allowing profile and item features to steer its updates.
The competition reflects a systems problem distinct from language models: recommendation inference must meet millisecond-scale latency and enormous request volume.
Its 6-million-yuan-plus prize pool and global participation show Chinese internet companies using competitions to expose real, anonymized production problems to international researchers.
The result is not proof of a universal recommendation scaling law, but it offers a concrete architecture and efficiency tradeoff worth testing in production.

Open at QbitAI Ask Claude ChatGPT

Pony.ai plans more than 2,000 Robotaxis across five European cities

# AI Leiphone 6h

Pony.ai and Uber plan to deploy over 2,000 Robotaxis in five European cities, testing whether a Chinese autonomous-driving product can scale through a foreign ride-hailing network.

The partnership will first connect Robotaxis in Zagreb to the Uber app and could expand into Europe and the Middle East.
Pony.ai supplies the L4 driving system, vehicles and operating experience; Uber supplies demand, payments and customer support, while local partners may own and operate fleets.
The company says its seventh-generation hardware bill of materials is 70% below the previous generation and that a 2027 vehicle could cost below CNY 230,000, about $32,400.
Pony.ai reports a Shenzhen unit-economics model at break-even, about 23 daily orders per vehicle and average daily net revenue of CNY 338, about $48.
The scale matters because vehicle utilization, maintenance and remote assistance costs only become visible when a fleet handles real demand rather than a constrained pilot.
The unresolved issue is regulatory and operational transfer: success in Chinese cities does not automatically translate to European weather, curb rules, insurance and labor regimes.

Open at Leiphone Ask Claude ChatGPT

Xiaomi's revenue and operating profit rebound as it funds AI and cars

# Analysis Leiphone 19h

Xiaomi's second-quarter revenue reached CNY 108.9 billion, about $15.3 billion, while operating profit more than doubled sequentially, but the recovery depends on premium phones, cars and sustained R&D spending.

Phone average selling price rose 25.9% to CNY 1,351, about $190, while higher-priced models accounted for 32.1% of mainland sales.
Smart electric-car and AI-related revenue reached CNY 24.9 billion, about $3.5 billion, up 17.1%; vehicle deliveries reached 104,199.
Xiaomi says its connected-device platform reached 1.16 billion devices and that second-quarter R&D spending rose 18.9% to CNY 9.2 billion, about $1.3 billion.
The strategic move is to absorb higher memory costs through product mix and use current cash generation to build cars, chips, operating systems, robotics and AI.
The risk is execution breadth: premiumization can depress unit volume while automotive and AI investments continue to consume capital before their returns are proven.
The reported recovery is operating-profit based; it should not be confused with the company's separate adjusted-net-profit decline discussed in item 25.

Open at Leiphone Ask Claude ChatGPT

Humanoid robot plays a full autonomous table-tennis match

# AI QbitAI 22h

SuperDynamix demonstrated a humanoid system that perceives, predicts, plans and controls a complete table-tennis exchange, using a world model and 117-degree-of-freedom body.

The SMASH 2.0 system must identify a fast, spinning ball, predict its trajectory and coordinate shoulder, elbow, wrist, torso and legs within millisecond-scale windows.
The company says the same control algorithm can run on different robot bodies, a meaningful claim about policy portability rather than just athletic performance.
Its KAI world model is pretrained on millions of videos and is intended to generate or reconstruct environments for simulation and closed-loop reinforcement learning.
The 37-degree-of-freedom hand and first-person KAI Halo capture system show the company's attempt to control both action hardware and data production.
The demonstration is technically richer than a scripted swing, but the commercial significance depends on whether the stack transfers to manipulation, not whether it wins a rally.

Open at QbitAI Ask Claude ChatGPT

China briefing: youth unemployment, Nvidia chips and the Unitree IPO

# Analysis Sinocism 9h

A China-focused briefing links Wang Yi's diplomacy, Unitree's listing, youth unemployment at 17.9% in July and Beijing's approval of Nvidia H200 purchases.

The 17.9% youth-unemployment figure is a hard social signal, though the excerpt does not provide the methodology or comparison series.
The briefing says Beijing approved purchases of Nvidia H200 chips, an important indication that domestic AI buildout can coexist with selective access to foreign accelerators.
Its structure puts technology, labor-market strain and foreign policy in one policy feed, reflecting how Chinese analysts read AI as part of the broader state economy.
This is a newsletter-style synthesis, so individual claims should be checked against the underlying agencies and companies.

Open at Sinocism Ask Claude ChatGPT

Alipay says agents will take a cut of service distribution

# Analysis GeekPark 5h

Alipay is recasting commerce around intent: its agent can call travel, food and logistics services, shifting merchants from buying visibility to being selected by an AI intermediary.

Alipay says its agent has connected more than 10,000 services across eight categories, including travel, dining, logistics and government services.
The platform's model is closer to delegated execution than search: the agent interprets a request, chooses a provider and completes a transaction.
That changes merchant competition from SEO and ad placement toward structured service interfaces, user-history access and the ability to expose distinctive supply.
The report says the platform already uses user preference and prior service habits before comparing providers for users with no established preference.
The strategic risk is platform power: an agent that controls intent interpretation can become a new gatekeeper, while merchants lose direct visibility into why they were or were not selected.
The commercial model remains unsettled, but the phrase 'intermediary fee' points toward monetizing transaction routing rather than selling an advertising slot.

Open at GeekPark Ask Claude ChatGPT

Chinese neutral-atom quantum startup raises again after reaching 2,310 qubits

# Signal 36Kr 6h

MatriQ says it has raised several hundred million yuan and built a 2,310-physical-qubit neutral-atom system, emphasizing engineering and delivery over laboratory scale alone.

The company says it completed four financing rounds in a year, totaling nearly CNY 1 billion, about $141 million.
Its system automatically identifies randomly loaded atoms, plans optical-tweezer paths and creates defect-free arrays, addressing a core engineering problem in neutral-atom machines.
The reported device uses separate storage and entanglement zones, with single-qubit fidelity above 99.9% and two-qubit fidelity above 99.2%.
MatriQ says it has finished technical definition of a commercial machine and received an order, though the customer and deployment are not identified.
The roadmap calls for 30-plus logical qubits by 2027 and 100-plus by 2028, making error correction and system stability more important than the physical-qubit headline.
Neutral atoms avoid the ultralow temperatures of superconducting systems and offer strong connectivity, but the commercial question remains whether optical control and error correction can be made reliable outside a specialist lab.

Open at 36Kr Ask Claude ChatGPT

Chinese fighter designers warn of hallucinations in military AI

# Analysis SCMP 18h

Engineers associated with China's J-20 and J-36 programs warn that language models can invent aircraft specifications, offering a rare Chinese military critique of AI reliability.

The engineers say large language models can rapidly mine intelligence and analyze weapons performance, but may fabricate technical specifications.
The warning is especially important in defense because a plausible but false aircraft parameter can contaminate threat assessment, simulation or procurement decisions.
The paper's authors work at the Chengdu aircraft design institute associated with the J-20 and next-generation J-36 programs, according to the report.
The Chinese framing is cautious rather than triumphalist: AI is useful for accelerating analysis, but human verification remains mandatory in high-consequence domains.
The report does not describe a deployed military system or a specific incident caused by a hallucination.

Open at SCMP Ask Claude ChatGPT

Wang Yi urges Washington to change its North Korea policy

# Policy SCMP 7h

China's foreign minister used talks in Seoul to endorse peaceful coexistence on the Korean peninsula and call Washington's policy toward Pyongyang hostile.

The statement came during Wang Yi's meeting with South Korean foreign minister Cho Hyun, according to China's foreign ministry.
Beijing's framing presents China as a constructive stabilizer while shifting responsibility for tension toward Washington's approach to Pyongyang.
The timing, ahead of the anniversary of China-South Korea diplomatic ties, combines peninsula diplomacy with an effort to strengthen bilateral trust.
The report does not show that Seoul accepted Beijing's characterization of US policy or that a new negotiating mechanism was agreed.

Open at SCMP Ask Claude ChatGPT

China's commercial space sector moves toward reusable operations

# Policy 36Kr 1d

A Chinese industry report says commercial space is shifting from technology validation to scaled operations, driven by reusable rockets, flexible satellite lines and state-backed capital.

The report estimates China's commercial-space market at CNY 2.83 trillion in 2025, about $399 billion, and forecasts more than CNY 3.5 trillion, about $493 billion, in 2026.
It highlights a sea-net recovery test by Long March 10B and a land recovery test by LandSpace's Zhuque-3 as signs that reusable launch is entering deeper engineering validation.
The report says first-half 2026 financing exceeded CNY 15 billion, about $2.1 billion, while state-owned capital's share rose to 59.8%.
The sector combines national launch infrastructure, private rocket firms, satellite manufacturing and downstream remote-sensing services rather than operating as a pure private market.
The largest constraints remain launch cadence, component localization, orbital-frequency competition and the ability to turn demonstrations into repeatable service revenue.
The report is industry research, with some market-size figures that may use broad definitions and should not be compared directly with narrower launch-market estimates.

Open at 36Kr Ask Claude ChatGPT

Black Myth: Wukong studio shows 15 minutes of its next game

# Culture 36Kr 5h

Game Science released the first substantial gameplay footage of Black Myth: Zhong Kui, showing China’s breakout AAA studio building a larger mythology franchise without using generative AI in production.

The 15-minute footage includes combat and story material from an early development build, with a 21:9 presentation format.
The studio says the game centers on life and death, uses a sword rather than Wukong's staff and keeps the realistic visual style of the original.
The first Black Myth sold more than 30 million copies and generated over $1 billion according to the report, giving the sequel unusual commercial room to expand.
Lead creator Yang Qi has said the team is not currently using AI and intends to keep making the game by hand, a deliberate cultural and production stance.
The release also shows a more disciplined communication cadence: the sequel's first substantial playable footage arrives roughly a year after its initial cinematic teaser.

Open at 36Kr Ask Claude ChatGPT

Zhu Rongji's four rules for state television

# Policy Pekingnology 14h

A translated 1998 speech by former premier Zhu Rongji reveals how China's leadership once used investigative television to expose local abuses while keeping media inside a Party-directed system.

Zhu's remarks to the Focus Report team praised reporting that exposed corruption, bureaucratic abuse and local wrongdoing.
The speech illustrates a controlled-accountability model: media can criticize lower-level officials when that serves central governance, but not challenge the political system's authority.
Publishing the historical text after Zhu's death invites comparison with today's tighter cultural and media environment, where investigative space is more constrained.
The source is a translation of archival remarks, not a current policy announcement.

Open at Pekingnology Ask Claude ChatGPT

Seres reports $81 billion in first-half revenue as Aito deliveries rise

# Signal Leiphone 19h

Seres reported CNY 57.5 billion, about $8.1 billion, in first-half revenue and 10.2% growth in Aito deliveries, while pouring CNY 7.0 billion into R&D.

The company reported cash reserves above CNY 73.15 billion, about $10.3 billion, and says it has completed more than CNY 587 million, about $83 million, in share repurchases.
Aito M9 deliveries exceeded 300,000 cumulatively, while the M6 passed 40,000 deliveries within 97 days of launch.
Seres says first-half R&D spending rose 34.8% and is supporting its platform, range-extender systems, safety stack and AI applications.
The company is also testing humanoid robots in manufacturing and reception, but the report provides no revenue or productivity evidence for that business.
The figures show a Chinese automaker using Huawei-linked premium vehicles to move beyond contract manufacturing, with brand value and software integration becoming part of the moat.

Open at Leiphone Ask Claude ChatGPT

China needs a more persuasive security offer to win middle powers

# Policy SCMP 10h

Chinese foreign-policy analysis argues that Beijing must offer middle powers more than criticism of Washington, as US allies are being pressed to align more tightly on defense.

The article responds to a US call for regional partners to spend more on defense and act as partners rather than protectorates.
Its central claim is that China cannot rely on dissatisfaction with US pressure; it must make its own security and economic offer credible to states that want room to maneuver.
The timing matters because a possible Xi-Trump meeting raises the cost of vague positioning and makes regional reassurance part of the summit backdrop.
This is analysis rather than a reported policy decision, but it exposes an internal Chinese debate about influence and strategic competition.

Open at SCMP Ask Claude ChatGPT

ChinaJoy shows AI moving from game demo to production pipeline

# Culture 36Kr 7h

At ChinaJoy, Chinese game companies presented AI both as a player-facing feature and as infrastructure for design, art, rendering and distribution, while still treating cultural content as the product.

Tencent separated consumer experiences from production tools, including natural-language game prototyping, asset generation and rendering support.
The broader industry discussion has shifted from whether AI can make games to how studios can use it, distribute it and comply with content and classification rules.
Sony's booth provided a useful counterexample by emphasizing playable games and Chinese-developed titles without making AI the central pitch.
The report's strongest signal is organizational: AI is being absorbed into planning, art, development, testing and publishing rather than marketed only as a standalone feature.
That integration could lower production costs, but it may also increase the importance of platform rules and content review as more material is generated procedurally.

Open at 36Kr Ask Claude ChatGPT

A Chinese animated film becomes a viral hit after viewers mocked it

# Culture SCMP 7h

The Chinese film The Arrival of the Ox jumped from CNY 3,420, about $480, on opening day to more than CNY 11 million, about $1.55 million, as a backlash-driven audience turned it into an event.

The film opened on August 5 with almost no audience, then reached CNY 717,000 in one day before passing CNY 11 million about ten days later.
The unusual trajectory suggests viewers were not simply discovering a hidden quality product; social circulation and ironic participation became part of the movie's demand.
This is a useful consumer-mood signal in China's highly networked entertainment market, where ridicule can provide distribution that paid promotion cannot buy.
The report does not establish whether the surge will create durable word-of-mouth or fade once the joke is exhausted.

Open at SCMP Ask Claude ChatGPT

Chinese researchers warn that defense contractors can capture policy

# Policy SCMP 7h

A Shanghai Jiao Tong University study warns Beijing about the influence of defense contractors, using the US military-industrial complex as a cautionary example.

The study describes an 'iron triangle' linking the US defense department, private contractors and congressional committees.
Its China-specific value is institutional: researchers are warning that commercial incentives can shape national-security priorities and spending decisions.
The argument fits a broader Chinese policy preference for keeping strategic industries subordinate to state goals rather than allowing contractors to define them.
The report is a study's analysis, not evidence that Beijing has adopted a new defense-procurement rule.

Open at SCMP Ask Claude ChatGPT

Chinese green tea becomes a social ritual in West Africa

# Culture SCMP 4h

Chinese tea dominates several West African markets, where a three-round, highly social brewing ritual turns an imported commodity into a shared daily institution.

The ataya ritual uses strong, sweet green tea served in three rounds, creating time for conversation among family, friends and colleagues.
The report places the practice in Mauritania, Senegal, Mali and Niger, from government offices to roadside stalls.
The Chinese angle is cultural and commercial rather than political: China supplies a product that local communities have adapted into their own social code.
The story is a useful reminder that China's overseas influence can travel through ordinary consumption and ritual, not only infrastructure or diplomacy.

Open at SCMP Ask Claude ChatGPT

Online retail reached 28.7% of China's above-quota goods sales

# Signal Jiemian 39m

China's Commerce Ministry says internet channels accounted for 28.7% of above-quota goods retail in the first seven months, while online services and rural commerce kept expanding.

Online goods and services retail grew 4.8% year over year from January through July, and the internet's share of above-quota goods retail rose 2.5 percentage points.
Online travel sales rose 24.9% and online lifestyle services 16.3% in the ministry's monitored data.
Online agricultural and rural goods sales rose 12.8%, while agricultural and rural commodity sales rose 5%.
The figures show e-commerce shifting from a consumer-distribution story toward an operating layer for rural supply, industrial clusters and services.
The source is a ministry briefing, so definitions and coverage differ from a single national retail series.

Open at Jiemian Ask Claude ChatGPT

Nokia plans to sharply shrink its China operations

# Analysis Solidot 1d

Nokia reportedly plans to cut most China staff and close factories by year-end while retaining service operations, signaling a deeper retreat from a market once used for R&D and global delivery.

Nokia had about 7,200 employees across mainland China and Hong Kong at the end of 2025, according to its annual report.
The company previously described China as hosting research, marketing, service, global delivery and supply operations, including a Beijing R&D center founded in 1995.
The reported plan would leave mainly after-sales support, though the number of affected workers has not been disclosed.
The move illustrates how foreign technology firms are separating China sales and service from globally integrated engineering and manufacturing footprints.
It is an adjacent China story: the subject is a foreign company, but the China operation itself is the substantive event.

Open at Solidot Ask Claude ChatGPT

Bambu Lab starts a consumer UV-printer project

# Analysis 36Kr 1h

Chinese 3D-printer company Bambu Lab is hiring for printhead, ink-path, FPGA and software roles as it moves toward consumer UV printing and a broader desktop-manufacturing platform.

The company has opened roles requiring experience with Epson, Ricoh or Xaar printheads and commercial inkjet production.
UV printing is more demanding than home printing because ink chemistry, nozzle control, curing and material compatibility are tightly coupled.
Bambu's move follows expansion into laser cutting and other fabrication modes, suggesting an ambition to make several professional manufacturing processes approachable at desktop scale.
The project is still at team-building stage; no product shape, price or release date has been disclosed.

Open at 36Kr Ask Claude ChatGPT

Baidu and Xiaomi face different kinds of profit pressure

# Analysis Jiemian 36m

Chinese analysis says Baidu's profit decline was largely an investment-marking effect, while Xiaomi's came from storage costs, subsidies and loss-making cars, making their recovery stories fundamentally different.

The article says Baidu's reported profit fell 68%, with about CNY 4.6 billion, roughly $650 million, attributed to investment fair-value and foreign-exchange effects.
It says Xiaomi's adjusted profit fell 42.6% because storage costs hurt phone margins, IoT subsidies eased and its car business remained loss-making.
Xiaomi's average selling price rose 25.9% to CNY 1,351, about $190, while smartphone shipments fell 26.5% to 31.2 million units.
The contrast is strategically useful: Baidu's problem is partly accounting volatility, whereas Xiaomi is paying a real operating cost to move upmarket and fund cars and AI.
The piece is commentary rather than a filing, so its causal decomposition should be checked against the companies' disclosures.

Open at Jiemian Ask Claude ChatGPT

Initium readers debate privacy, disasters and life at China's margins

# Culture Initium 9h

An Initium reader-response roundup links concerns about data collection, old dams, news avoidance and marginal urban identity, showing how Hong Kong readers connect technology to institutional trust.

Reader responses discuss controversy over a navigation app in Taiwan, where precise traffic data raised concerns about personal-data collection and national security.
Other comments connect extreme flooding to aging reservoirs and climate stress, and describe news avoidance as a response to political fatigue and social exhaustion.
The discussion is opinion, not reporting, but it reveals how readers interpret Chinese technology and governance through privacy, safety and institutional accountability.
The roundup also shows the role of reader comments as a site for processing local identity and belonging in and around Beijing and Hong Kong.

Open at Initium Ask Claude ChatGPT

ChinAI	403	0/0		HTTPError: 403 Client Error: Forbidden for url: https://chinai.substack.com/feed
ChinaTalk	200	1/20
Pekingnology	200	1/20
Ginger River Review	200	0/20
Sinocism	200	1/20
QbitAI	200	10/10
36Kr	200	9/30
Solidot	200	13/20
Leiphone	200	15/20
GeekPark	200	4/30
IT Home	200	12/60	+40 over cap
SCMP	200	10/50	+5 over cap
Initium	200	3/15
Jiemian	200	8/30	+3 over cap
China Digital Times	200	1/15
The Wire China	200	0/50
Techmeme	200	3/15
Hacker News	200	0/20
V2EX	200	8/50	+38 over cap

Unitree took $48 billion out of a falling market, and its earliest backer says most of this sector dies

Unitree up 460%, Baidu down 12%

LandSpace landed a steel booster on legs

Nobody knows who runs the AI talks

DeepSeek's harness now ships whole agents

A Qianmen travelogue got deleted anyway

Threads we are pulling