China Desk — 27 Aug 2026, 22:11 UTC

Himalayan flood devastates China-Nepal trade route as toll rises

# SCMP 9h

A flash flood and mudslide destroyed infrastructure at the Gyirong crossing, leaving hundreds missing and exposing the environmental risk behind China-Nepal trade plans.

Also: Jiemian, Leiphone, Solidot, Initium, IT Home

The crossing is strategically important to Nepal’s effort to reduce reliance on Indian transit routes, so the disaster affects both humanitarian operations and regional commerce.
Chinese authorities confirmed major casualties and missing people, while preliminary Nepalese figures reported a much higher toll on the other side of the border.
The event followed an ice-rock avalanche and debris flow rather than a conventional earthquake, complicating early-warning and infrastructure design.
Future China-Nepal roads, ports and other projects will need to account for unstable high-altitude terrain, glacial melt and compound flood pathways.
The disaster may temporarily close a trade route, but its longer-term significance is whether both governments change construction standards and hazard planning.
Casualty figures remain preliminary and vary across reports, so the human toll should not be treated as final.

Open at SCMP Ask Claude ChatGPT

Chinese executives may join Xi’s US trip as trade truce extension nears

# SCMP 11m

Chinese coverage presents a possible CEO delegation as part of extending the US-China trade truce, ahead of Xi Jinping’s expected September 24 meeting with Donald Trump.

Officials are discussing whether Chinese business leaders will join Xi’s US delegation, according to people familiar with the talks cited by the source.
The proposed extension concerns a one-year trade agreement reached in Busan last October, suggesting both sides are prioritizing managed continuity over a new comprehensive deal.
The CEO question is strategically important because Trump’s previous Beijing visit featured a large group of American executives, creating pressure for reciprocal business engagement.
Chinese participation would give companies a direct channel into negotiations that otherwise remain dominated by state officials and tariff mechanics.
The source does not identify which Chinese executives might attend, so the political signal is clearer than the commercial substance at this stage.

Open at SCMP Ask Claude ChatGPT

China’s internet giants end the AI office-product race

# TMTPost 11h

Tencent, Alibaba and ByteDance are consolidating AI office teams into fewer centers, revealing three different bets on the future of enterprise agents.

Tencent folded QClaw into WorkBuddy, treating a validated enterprise product as the main line and a lighter consumer assistant as a supporting capability.
Alibaba shut down three parallel product lines and created a Qwen office division, choosing organizational reset over selecting one existing winner.
ByteDance split Feishu between Doubao product leadership and Volcano Engine sales, then moved TRAE and Coze into the Doubao structure.
The common signal is that rising model, distribution and customer-acquisition costs are ending the era of unconstrained internal horse-racing.
The strategic risk differs by company: Alibaba faces migration and boundary problems with DingTalk, while ByteDance must reconcile enterprise customization with a reusable model core.
These are product-organization changes, not proof that each company has unified its underlying models or compute infrastructure.
The next six months should be judged by talent retention, product velocity, customer overlap and whether sales demands contaminate the common agent platform.

Open at TMTPost Ask Claude ChatGPT

PDD’s 40% R&D jump signals a move deeper into supply chains

# InfoQ China 7h

PDD’s second-quarter revenue grew 8% while advertising rose only 3%; R&D jumped 40% to 4.6 billion yuan, roughly $648 million, for supply-chain systems and governance.

The figures show a platform whose familiar advertising engine is slowing even as transaction-service revenue grows faster.
The reported R&D budget is spread across traceability, agricultural standards, factory digitization, rural fulfillment, compliance and overseas logistics rather than a single AI program.
A lychee traceability system illustrates the shift: platform data links farms, batches, inspections, packing, shipping and after-sales responsibility.
The strategic move is from matching demand to helping shape production and fulfillment, which can improve consistency but requires operational coordination outside the platform’s traditional boundary.
PDD reported that one merchant’s lychee repeat purchases rose from about 15% to 30%; that is a local example, not proof of a platform-wide effect.
The story matters because Chinese e-commerce growth increasingly depends on controlling physical bottlenecks, not just buying traffic.

Open at InfoQ China Ask Claude ChatGPT

Xpeng upgrades its driving model from 3D space to 4D time

# New Wisdom 11h

Xpeng says its second-generation VLA now uses 30-second temporal context, six-second prediction and streaming inference to connect perception, prediction and action.

The model adds an Infini-VLA long-sequence architecture so driving decisions can use recent motion history rather than isolated frames.
Streaming autoregressive inference is claimed to improve end-to-end response speed by 300%, allowing perception and action to proceed continuously rather than in discrete batches.
An X-Foresight world model predicts possible behavior up to six seconds ahead, while a mixture-of-experts design separates driving tasks to reduce interference.
The vehicle-level Master Agent coordinates driving, cabin, chassis and body-control agents, turning spoken intent into multi-system actions such as finding a parking place.
Xpeng says the edge model has 3.5 times more parameters than its predecessor and is being reused across production cars, robotaxis and robots.
These are company claims; the meaningful validation will be safety performance, intervention rates and generalization across markets rather than parameter count.

Open at New Wisdom Ask Claude ChatGPT

iFlytek’s results show AI moving from investment to monetization

# 36Kr 12h

iFlytek reported 11.623 billion yuan, roughly $1.64 billion, in first-half revenue and a second-quarter cash-flow turnaround as platform and enterprise AI gained share.

Revenue rose 6.5%, while net loss narrowed to 204 million yuan, roughly $29 million, against R&D spending of 3.007 billion yuan, roughly $424 million.
Open-platform revenue grew 36% to 3.705 billion yuan, roughly $522 million, becoming the company’s largest business line with a 20.2% gross margin.
API and model-as-a-service revenue rose about 70%, supporting a transition from project delivery toward recurring platform and token usage.
The company says it has more than 11.5 million AI developers, including more than 3.2 million large-model developers, giving it a distribution base for agents and model services.
Education, healthcare and automotive are the important verticals, while hardware supply constraints hurt its AI learning-device business.
The improvement is not yet a clean profitability story: large R&D and sales spending remain necessary, and second-quarter cash flow turning positive is only an early quality signal.

Open at 36Kr Ask Claude ChatGPT

Hands-on tests find GLM-5.3-Flash cheap enough for serious workflows

# AI Zhidx 18h

A Chinese hands-on review says GLM-5.3-Flash can edit video, build websites and prototype games while cutting model prices by up to 97.5%.

Also: Leiphone

The model accepts text, images, video and files and was tested on visual question answering, document recovery, web replication, video editing and code generation.
The reviewer used it to extract a highlight reel from a 48-minute talk, preserving clips, subtitles and intermediate artifacts rather than returning only a summary.
In coding tests it inspected screenshots of a generated game and iterated on behavior, showing the value of multimodal feedback inside an agent loop.
The model is priced at 0.8 yuan per million input tokens, roughly $0.11, and 2.8 yuan per million output tokens, roughly $0.39, before temporary discounts.
The test is useful because it measures workflow completion rather than benchmark scores, but it remains a single publication’s hands-on evaluation.
Its combination of low price, multimodality and domestic-chip serving is the competitive threat to watch, not any one game recreation demo.

GLM-5.3-Flash claims frontier performance on domestic Chinese chips Leiphone

Open at Zhidx Ask Claude ChatGPT

MiniMax shifts from consumer apps toward enterprise model demand

# Signal QbitAI 9h

MiniMax says first-half revenue reached $116.6 million, up 283%, while business revenue rose to 63.4% of the total and token use hit 20 times January’s level.

The company’s annual recurring revenue reportedly exceeded $800 million in August, up from $150 million in February, though the figure is management-disclosed.
Enterprise and developer services generated $73.9 million in the first half, up 703%, and supplied roughly three-quarters of incremental revenue.
Consumer products still grew, so the shift reflects faster business demand rather than a reported collapse in consumer usage.
The token surge suggests agents are increasing workload intensity: one automated task can generate far more inference than a human chat session.
MiniMax says infrastructure efficiency lifted gross margin to 17.9%, while R&D still reached $296.9 million, showing that commercialization has not removed the capital burden.
The strategic question is whether API and enterprise demand remains durable after novelty-driven adoption and open-model experimentation fade.

Open at QbitAI Ask Claude ChatGPT

ByteDance’s Doubao Work makes enterprise context the agent advantage

# AI Zhidx 13h

A hands-on Chinese review finds Doubao Work can research, write documents, build presentations and act across Feishu, using organizational permissions and context as its main edge.

The reviewer says it produced a roughly 25,000-word research document with 47 traceable sources and a 14-slide presentation, though these are test results from one publication.
The agent can inspect Feishu chats, documents and tables, then turn project status into a stored weekly report and context-aware reminders.
Its skill system packages search, document access, table queries and messaging into reusable enterprise capabilities rather than one-off prompts.
The architecture inherits Feishu permissions and audit boundaries, addressing the central enterprise-agent problem of what the system may read and change.
The product’s advantage is not necessarily a smarter base model; it is privileged access to organizational memory and structured workflows.
The test does not establish reliability across arbitrary companies, and the review is closely tied to the product launch.

Open at Zhidx Ask Claude ChatGPT

Chinese analysts link tax enforcement to local-government funding pressure

# Analysis Sinocism 12h

A China-focused roundup connects tax crackdowns, a surprise discipline meeting, economic messaging and Unitree’s stock decline to Beijing’s effort to manage fiscal and market stress.

The roundup frames tax enforcement as a way for the central government to cover local-government shortfalls, rather than as a narrow compliance campaign.
It also flags a surprise plenary session of the Central Commission for Discipline Inspection and the life sentence of Evergrande founder Hui Ka Yan.
The discussion of Unitree’s falling stock price asks what the Party wants from markets: financing and confidence, but without speculative excess or loss of control.
The source is a newsletter and podcast summary, so its bundled claims are analysis and agenda-setting rather than a single reported government decision.
Its value is the juxtaposition of fiscal extraction, political discipline, property-sector punishment and market management as parts of one governance problem.

Open at Sinocism Ask Claude ChatGPT

Qwen Office says its new Flash model doubles speed and cuts tokens 75%

# AI QbitAI 20h

Alibaba’s Qwen Office launched a model-tuned Qwen3.8-Flash mode that it says doubles task speed and cuts average token consumption by 75%.

Also: Leiphone

The model was tuned jointly for multi-step planning, tool selection and context compression rather than exposed as a generic base-model upgrade.
Alibaba says 95% of routine office tasks can use the lower-cost standard mode, reserving the advanced mode for harder work.
This is an application-level optimization claim: harness design, model routing and task-specific training matter as much as raw benchmark intelligence.
The source is authorized promotional material and supplies no independent latency test, token accounting or quality tradeoff.
If the gains generalize, office agents can become economically viable by reducing wasted reasoning and context rather than simply lowering model prices.

Qwen Office launches Qwen3.8-Flash as its new standard mode Leiphone

Open at QbitAI Ask Claude ChatGPT

China’s border disaster dominates a policy roundup before Xi’s SCO trip

# Analysis Sinocism 23h

A China-focused newsletter links the Gyirong disaster, Xi Jinping’s upcoming SCO travel, India relations and a new domestic AI model in one political snapshot.

The roundup says Xi is due to attend the Shanghai Cooperation Organisation summit and visit Egypt, while an expected Politburo meeting may address the economy.
It places the border disaster alongside China-India and China-Nepal issues, showing how a local catastrophe quickly enters a wider frontier and infrastructure agenda.
The inclusion of a new Chinese AI model beside politics and disaster response reflects how technology has become a standing component of China policy coverage.
This is a newsletter summary, not a single event report; its value lies in what a China-focused analyst groups together as the day’s governing priorities.
The economic section reportedly finds no clear sign of a major new stimulus package, an absence that may matter as much as any announced measure.

Open at Sinocism Ask Claude ChatGPT

Kunlun’s embodied world model ranks second in WorldArena

# AI Leiphone 12h

Chinese robotics startup Kunlun says its GeWu world model won the image-quality category and ranked second overall in the WorldArena 2.0 evaluation.

The model scored 65.91 overall and 69.58 on image quality among 77 entries, according to the company’s account of the benchmark.
Its architecture separates intent, intervention and consequence into three parameter-independent transformer towers.
A one-way causal-attention pattern prevents the action tower from seeing future visual consequences while allowing the consequence tower to condition on intent and action.
The company reports that removing these pathways caused the model to imitate familiar video patterns instead of predicting physically consistent outcomes.
The result is technically interesting because it treats a world model as a simulator of action consequences rather than merely a video renderer.
The ranking is reported by the vendor and should be checked against the benchmark’s published protocol, task mix and reproducibility before being treated as a field-wide lead.

Open at Leiphone Ask Claude ChatGPT

Feihe puts 436 agents into dairy-factory operations

# AI Leiphone 15h

Chinese dairy maker Feihe says it has built 436 agents across maintenance, process control and workforce management, using a common platform and factory data.

The company reports more than 240,000 active users over three months and nearly 23 billion token consumption across more than ten factories.
Its equipment-management agent cut average fault-recovery time from 104 hours to 47 and repeated-fault rates from 34% to 8%, according to Feihe.
A membrane-cleaning workflow reportedly saves more than 500,000 yuan, roughly $70,000, in annual purchasing and 144,000 tons of water.
A workforce agent reduced process nodes from 15 to five and reportedly saved more than 2 million yuan, roughly $282,000, in development and maintenance costs.
The operating model matters more than the agent count: frontline workers propose use cases while a central platform handles models, skills, permissions and governance.
These are company-reported results, but they show a credible path from agents as demonstrations to agents embedded in industrial control and administrative loops.

Open at Leiphone Ask Claude ChatGPT

Why Kimi and GLM converge on the same linear-attention setting

# AI Leiphone 11h

Chinese researchers note that Kimi K3 and GLM-5.3-Flash both use a linear-attention forgetting bound of minus five, illustrating a narrow stability regime for long context.

Linear attention compresses history into a fixed-size state instead of retaining a full KV cache, reducing memory growth for very long sequences.
Its weakness is recurrent numerical instability: repeated state updates can underflow, explode or produce NaNs in half-precision training.
The reported minus-five gate bound preserves a small residual memory signal while avoiding the overload associated with retaining too much history.
Kimi appears to use the setting as part of a native hybrid architecture, while GLM combines related ideas with sparse attention and other components.
Similar public configuration values do not prove copying; independent teams can converge on the same narrow engineering optimum.
The broader signal is that Chinese model competition is moving below headline benchmarks into kernels, state representations and memory economics.

Open at Leiphone Ask Claude ChatGPT

Leju’s industrial VLA uses a middle-training layer for robots

# AI Leiphone 14h

Chinese robotics maker Leju says its KUAVO VLA model used more than 600 hours of same-platform robot data and doubled efficiency in selected industrial tasks.

The model is positioned between a general VLA foundation model and task-specific post-training, storing common industrial capabilities before individual skills are learned.
Training on one robot configuration reduces variation in action space, mechanics and control behavior, making later adaptation less expensive on that platform.
The company reports a 48.27% task-success rate and 74.51% process score across 25 tasks, ahead of the comparison models listed in the article.
The design emphasizes process quality, smoothness and repeatability rather than only whether the final task succeeded.
This is a strong argument for a middle-training layer in embodied AI, but the reported benchmark is company-run and uses a limited task set.
The commercial payoff depends on whether skills transfer across factories and workflows without recreating the same data-collection burden.

Open at Leiphone Ask Claude ChatGPT

Extreme weather is unlikely to derail China’s nuclear expansion

# Policy SCMP 8h

Chinese nuclear expansion is less exposed to river-water constraints than some European fleets, according to an industry specialist, though heat remains a design risk.

The analysis contrasts China’s continued build-out with European reactors forced to reduce output as rivers become warmer and shallower.
China’s newer nuclear sites and broader cooling choices may reduce exposure to the specific river conditions affecting France and Romania.
That does not remove climate risk: heat waves, drought, storms and water competition still affect plant siting, cooling design and grid operations.
The strategic implication is that climate adaptation may reinforce China’s nuclear push by making firm low-carbon generation more valuable, rather than stopping it.
The source offers an industry view rather than a plant-by-plant safety assessment, so it does not establish that China’s fleet is immune to extreme weather.

Open at SCMP Ask Claude ChatGPT

OceanBase leads China’s distributed database market with 20.2%

# Signal Leiphone 15h

A new market study puts China’s distributed transactional database market at about 5.95 billion yuan, roughly $838 million, with OceanBase leading at 20.2%.

The study defines the category by horizontal scalability, cross-node transaction consistency, strong consistency, high availability and zero recovery-point loss.
Alibaba Cloud, Tencent Cloud and Huawei Cloud followed OceanBase with 17.9%, 16.5% and 14.3% shares respectively.
The market grew from about 1.69 billion yuan in 2021 to 5.52 billion yuan in 2025, a reported 34.5% compound annual growth rate.
Adoption is moving from internet and payment back ends into financial cores, communications and other infrastructure sectors where migration risk matters more than benchmark speed.
OceanBase is also extending toward HTAP, multimodal data and agent-facing services, pushing the category toward an AI data-platform role.
The figures come from a consulting report and should be read as a market-definition claim, not an audited industry census.

Open at Leiphone Ask Claude ChatGPT

Taiwan prosecutes nine people over alleged AI-server sales to China

# Policy Initium 23h

Chinese coverage highlights a Wall Street Journal report that Taiwan charged nine people accused of reselling AI servers to China, underscoring enforcement around compute controls.

The alleged transactions concern servers whose final destination was China, placing the case in the broader effort to restrict advanced AI-compute access.
Taiwan’s prosecution shows that export-control enforcement depends on intermediaries, resellers and end-user verification, not only on rules imposed by Washington.
For Chinese AI developers, the case signals that alternative procurement channels are becoming legal and operational liabilities as scrutiny spreads across the supply chain.
The source is a translated headline and page fragment, so it does not provide the charges, hardware specifications or evidence in sufficient detail to judge the case.
The strategic implication is that compute access is being regulated as a network of transactions rather than a single bilateral export decision.

Open at Initium Ask Claude ChatGPT

US lawmakers condemn calls to intern Chinese Americans

# Culture SCMP 9h

Chinese coverage foregrounds US lawmakers’ warning that internment rhetoric against Chinese Americans would fuel hate during a hypothetical war with Beijing.

The statement came from Asian American members of Congress after a right-wing podcast advocated forced internment in a conflict scenario.
The issue is not a current government policy, but it exposes how geopolitical confrontation can turn into domestic suspicion of an ethnic community.
Chinese media attention to the episode reflects sensitivity to the historical link between national-security language and discrimination against people of Chinese descent.
The lawmakers’ response is a political repudiation, not evidence that such proposals have institutional support across Washington.
Its significance lies in the social cost of US-China rivalry: even hypothetical conflict planning can reshape how Chinese Americans are treated at home.

Open at SCMP Ask Claude ChatGPT

Bill Gates pitches global AI cooperation before a possible Xi meeting

# Policy SCMP 13h

Chinese coverage presents Bill Gates’s possible meeting with Xi as part of a push for international cooperation on AI risks and governance.

Gates may meet Xi later in the year, according to a Reuters report cited by the source, after last seeing him in 2023.
He also published a long essay calling for cooperation on AI risks, giving the possible meeting a policy rather than purely ceremonial context.
For Beijing, a discussion with a prominent US technology figure offers a channel for engaging global AI governance without relying solely on formal US government talks.
For Gates, China is both a major technology ecosystem and a necessary participant in any claim to global AI-risk coordination.
The meeting is not confirmed and the article provides no indication that it would produce an agreement or change bilateral restrictions.

Open at SCMP Ask Claude ChatGPT

Greentown’s older inventory finally falls below half its stock

# Signal 36Kr 19h

Greentown China says pre-2022 inventory fell below half of remaining unsold value, but first-half revenue dropped 26% as property margins and asset values weakened.

Revenue was 39.481 billion yuan, roughly $5.56 billion, while attributable profit fell 60.8% to 81.7 million yuan, roughly $11.5 million.
The company recorded 1.908 billion yuan, roughly $269 million, in impairment and fair-value losses, including 1.19 billion yuan of inventory impairment.
Older inventory still represented about 1.06 trillion yuan, roughly $149 billion, of remaining unsold value, making disposal speed central to future earnings.
Greentown says it is tying group-level resources and management incentives to difficult stock, including parking spaces, tail-end units and undeveloped land.
Cash and leverage remain relatively strong, which gives the company time to manage inventory, but management cannot forecast how much further impairment may be needed.
The result captures China property’s current tradeoff: balance-sheet safety can improve even while the profit engine remains under pressure.

Open at 36Kr Ask Claude ChatGPT

AliExpress Brand Plus grows 97% in the EU despite tax changes

# Policy Leiphone 18h

AliExpress says its Brand Plus business grew 97% during an EU promotion as taxes and regulation reshape China’s cross-border e-commerce competition.

The reported growth is concentrated in branded, higher-value goods rather than the low-ticket parcels that face the strongest regulatory pressure.
A French price-comparison service cited by the source says Temu’s orders fell about half, while AliExpress and Shein were less affected.
The shift suggests customs and tax changes can accelerate segmentation: platforms with brands, local inventory or higher average order values may absorb compliance costs better.
AliExpress is trying to move from a lightweight marketplace toward a more controlled brand and fulfillment position in Europe.
The growth and competitor comparisons are not independently audited in the supplied report, so the strategic direction is more reliable than the precise percentage.

Open at Leiphone Ask Claude ChatGPT

Hong Kong fandom culture turns on the men chasing a boy-band company

# Culture Initium 23h

An Initium feature examines hostile online dynamics around Times Fengjun Entertainment, including incel men and women in the fandom facing harassment and exclusion.

The story treats fandom as a social power structure, not merely entertainment consumption, with gendered expectations shaping who is allowed to participate.
Its focus on Hong Kong and Chinese-language fan culture gives the conflict a local social context that a general celebrity report would miss.
The title signals competing identities: men seeking access to the company’s orbit and female fans portrayed as objects of suspicion or attack.
The supplied text is largely a headline and page fragment, so the underlying incidents, participants and scale cannot be independently assessed here.
The item is still useful as a signal of how online communities around Chinese pop companies can reproduce broader anxieties about masculinity, intimacy and status.

Open at Initium Ask Claude ChatGPT

Brazil will not loosen China’s rare-earth magnet grip this decade

# Policy SCMP 5h

Chinese coverage highlights research that Brazil may not produce rare-earth magnets before 2032, leaving China’s manufacturing advantage intact for years.

The source quotes the coordinator of Brazil’s main magnet pilot plant, who expects initial production only by the end of the decade’s first half.
The relevant bottleneck is not mining alone: converting separated rare-earth material into high-performance permanent magnets requires specialized processing and manufacturing.
For China, the significance is strategic rather than merely commercial because magnet production sits downstream of mining and is critical to electric motors, wind turbines and defense systems.
The forecast suggests that new mines outside China will not quickly translate into independent magnet supply chains.
The estimate comes from a Brazilian industry representative and describes a current timetable, not a guarantee that China’s position will remain unchanged.

Open at SCMP Ask Claude ChatGPT

Li Auto’s profit recovery now depends on its fourth quarter

# Signal TMTPost 8h

Li Auto delivered 193,500 vehicles in the first half but lost 3.98 billion yuan, roughly $561 million, as vehicle gross margin fell from 19.6% to 7.8%.

Revenue fell 13.4% to 48.65 billion yuan, roughly $6.9 billion, while deliveries declined only 5.1%, showing that product mix and pricing hurt more than volume.
The transition combines new extended-range models, rising pure-electric exposure, inventory clearance and continued spending on batteries, chips and charging.
Second-quarter vehicle margin recovered to 9.4% from 6.1% in the first quarter, but remained far below the prior year.
Management says extended-range and pure-electric orders are approaching an even split, marking a real shift away from Li Auto’s original product identity.
The company still held 87.5 billion yuan, roughly $12.3 billion, in cash, but first-half operating cash flow was negative by about 6.1 billion yuan.
The recovery case depends on premium models and the i9 improving mix; a volume rebound without margin recovery would leave the strategic transition unfinished.

Open at TMTPost Ask Claude ChatGPT

A Chinese tech roundup mixes AI growth, layoffs and corporate restraint

# Analysis Leiphone 21h

A Chinese roundup juxtaposes DeepSeek’s revenue and infrastructure spending, Tencent’s AI caution, cloud layoffs and labor disputes to capture a more anxious tech mood.

The roundup says DeepSeek earned 475 million yuan, roughly $67 million, in the first seven months while losing 715 million yuan, roughly $101 million, and spending about 11 billion yuan on AI infrastructure.
It pairs that capital intensity with Tencent’s argument that lasting AI execution matters more than being first, a framing of patience over panic.
Unverified employee accounts of layoffs and pay changes at China Telecom’s cloud unit show how AI investment is colliding with weaker or uncertain enterprise economics.
The item also cites a comparison between Unitree’s R&D spending and that of a large pig-farming company, using it to question whether China’s robotics boom is underfunded.
Because this is a headline roundup mixing reported, attributed and unverified claims, it is best read as a snapshot of Chinese tech discourse rather than a clean fact set.
The common thread is that model capability is rising alongside pressure to prove monetization, control compute costs and protect jobs.

Open at Leiphone Ask Claude ChatGPT

Shenzhen’s proposed edge in the AI-for-science race

# Analysis SCMP 13h

Chinese analysis argues Shenzhen can compete in AI for science by connecting models to experiments, simulations, instruments and industrial data rather than chasing model size alone.

The emerging competition is framed as shortening the loop from hypothesis to simulation, experiment, measurement and updated theory.
The analysis compares this with the US Department of Energy’s Genesis Mission, implying that integrated research infrastructure matters as much as raw compute.
Shenzhen’s advantage would come from dense industrial, manufacturing and hardware networks that can supply real problems and experimental feedback.
The city would still need stronger links between universities, laboratories, equipment and data governance; the article offers a strategy argument rather than evidence of a completed platform.
The useful takeaway for technology leaders is that AI-for-science moats may be built from instruments, workflows and proprietary experimental data, not only from foundation models.

Open at SCMP Ask Claude ChatGPT

Sword of Justice creator Guo Weiwei leaves Xishanju after 22 years

# Culture 36Kr 12h

Guo Weiwei, the longtime creator and public face of Xishanju’s Sword of Justice franchise, is leaving after 22 years as the studio faces strategic uncertainty.

Guo left the CEO role in late 2025 and shifted to chief producer, making this departure the final step in a management transition rather than a sudden executive removal.
His career is intertwined with the franchise’s PC, mobile, seasonal and fan-community operations, so the risk is cultural as much as managerial.
Xishanju’s newer mech shooter reportedly underperformed against expectations, leaving the studio exposed on both its legacy MMO and newer-product fronts.
The story illustrates how Chinese game franchises can become identified with individual producers whose taste and community relationships are part of the product.
The next test is whether Xishanju can preserve the franchise’s operating rhythm while reallocating resources toward newer formats.

Open at 36Kr Ask Claude ChatGPT

Former TV leader Konka seeks voluntary delisting after another loss

# Signal IT Home 2h

Konka, once China’s top television maker, plans to leave Shenzhen’s exchange after negative net assets and a first-half loss of 173 million yuan, roughly $24 million.

Konka says its 2025 audited year-end net assets were negative, triggering delisting-risk status under Shenzhen exchange rules.
The company reported first-half revenue of 3.852 billion yuan, roughly $540 million, down 26.6%, and a net loss of 173 million yuan, roughly $24 million.
Its semiconductor business remains in an early industrialization phase and is still loss-making, while consumer electronics margins were squeezed by supply volatility and higher costs.
The proposed voluntary withdrawal covers both its A and B shares and would move trading to the over-the-counter delisting board.
Konka’s trajectory illustrates how former state-linked consumer-electronics champions are trying to reposition into semiconductors without yet achieving the scale needed to offset the decline of legacy hardware.
China’s exchange rules increasingly make negative net assets an existential corporate condition, not merely a warning to investors.

Open at IT Home Ask Claude ChatGPT

Why Southeast Asia is resisting the US offer of flexible realism

# Analysis SCMP 41m

Chinese analysis argues that Southeast Asian governments want strategic room rather than a US-defined realism, amid pressure to choose sides in the wider US-China competition.

The article centers on Elbridge Colby’s Southeast Asian tour and his effort to package US policy as pragmatic, nationalist and economically focused.
Its core argument is that regional governments may welcome dialogue with Washington while resisting a framework that implicitly narrows their options toward China or the United States.
This is a reminder that Southeast Asia’s hedging is not simple neutrality: governments seek security ties, trade and infrastructure access from competing powers at the same time.
The analysis is commentary rather than evidence of a coordinated regional decision, so its value lies in the framing of strategic distrust rather than a new policy announcement.

Open at SCMP Ask Claude ChatGPT

Jitai’s 134-fold revenue jump is mostly a one-off licensing payment

# Signal TMTPost 8h

Chinese drug-delivery AI firm Jitai reported 154.3 million yuan, roughly $21.7 million, in first-half revenue, but 89% came from a single US licensing payment.

Revenue rose 134-fold from a tiny base after Jitai booked a $20 million upfront payment for the global license of MTS-128.
The company reported a 153.2 million yuan, roughly $21.6 million, attributable loss under IFRS, while adjusted losses narrowed to 51.1 million yuan, roughly $7.2 million.
The accounting improvement is less dramatic than the headline because higher share-based compensation and other add-backs explain part of the adjusted result.
Research spending reached 172.4 million yuan, roughly $24.3 million, and remained above revenue, consistent with a platform still funding clinical and preclinical programs.
The licensing deal validates external interest in Jitai’s assets, but a single upfront payment is not recurring commercial traction.
Most pipeline programs remain preclinical or early-stage, leaving regulatory execution as the key test of whether the valuation story can mature.

Open at TMTPost Ask Claude ChatGPT

Tianchen Medical enters a difficult transition between product cycles

# Signal TMTPost 8h

Tianchen Medical’s first-half revenue fell 27% and core profit fell 89% to 5.1 million yuan, roughly $0.7 million, as old products weakened before new ones scaled.

Reported net profit was 13.7 million yuan, roughly $1.9 million, but more than 60% came from investment and other nonrecurring gains.
Passive stapler revenue still represented 69% of sales and was pressured by Chinese volume-based procurement and lower winning prices.
Active products carry a 72.1% gross margin but remain too small to offset the decline of the legacy business.
R&D rose 32.3% to 25 million yuan, roughly $3.5 million, funding surgical platforms, powered staplers and minimally invasive tools.
The company is betting on intelligent and robotic surgery while its cash generation weakens, creating a timing problem between investment needs and commercial proof.

Open at TMTPost Ask Claude ChatGPT

Sinodaan plans a 2.4 billion yuan raise amid a graphite price war

# Signal TMTPost 9h

Sinodaan’s revenue rose 44% but profit fell 46%; it plans a 2.4 billion yuan, roughly $338 million, raise for a 300,000-ton Sichuan anode project.

Lithium-battery anode revenue grew with shipments, but gross margin fell to 10.2% as raw materials and outsourced processing stayed expensive.
Operating cash flow was negative by 905 million yuan, roughly $127 million, because inventory purchases and receivables absorbed cash.
The financing concentrates resources in a vertically integrated Sichuan project while the company cancels more than 100 billion yuan of other planned capacity.
That is a deliberate scale-and-cost bet in an oversupplied market, not evidence that demand alone justifies expansion.
Hard-carbon anodes are already in mass production and silicon-carbon materials are approaching introduction, but both face intense competition.
The project’s value depends on whether integration creates a durable cost advantage before the industry’s excess capacity is cleared.

Open at TMTPost Ask Claude ChatGPT

ChinAI	403	0/0		HTTPError: 403 Client Error: Forbidden for url: https://chinai.substack.com/feed
ChinaTalk	200	0/20
Pekingnology	200	0/20
Ginger River Review	200	0/20
Sinocism	200	2/20
QbitAI	200	4/10
36Kr	200	10/30
Solidot	200	11/20
Leiphone	200	19/20
GeekPark	200	3/30
IT Home	200	12/60	+41 over cap
SCMP	200	10/50	+11 over cap
Initium	200	5/15
Jiemian	200	8/30	+16 over cap
China Digital Times	200	0/15
The Wire China	200	0/50
Techmeme	200	0/15
Hacker News	200	0/20
V2EX	200	8/50	+23 over cap
Zhidx	200	5/20
InfoQ China	200	8/20	+6 over cap
New Wisdom	200	8/15	+2 over cap
TMTPost	200	8/16	+8 over cap

The flood that erased Gyirong has a sequel, and Beijing is trying to compute when it arrives

Gyirong is gone and a lake is filling upstream

Xi lands in Washington on 24 September

Four ministries come for "pseudo-innovation"

GLM-5.3-Flash served 62 trillion tokens on Chinese chips

Shawshank sells out a thousand-seat hall

Threads we are pulling