China Desk

China News, Summarized 30 Aug 2026, 05:13 UTC 34 stories 22/23 sources

Today's brief

The Pentagon's China list is now being answered in court, not by lobbyists

CXMT takes Section 1260H to a Washington judge

CXMT sued the Pentagon on Friday. ChangXin Memory Technologies — the DRAM maker whose first-half revenue of 150.3bn yuan (~$21bn) made it, by Bloomberg's count, China's most valuable listed business — filed in the US District Court for the District of Columbia to get off the Section 1260H list of firms the Defense Department says support China's military. Its chips are civilian, it says. Reuters has the filing.

The procedural history is the argument. Designated in January 2025, CXMT says it spent over a year submitting information; the Pentagon published a notice in February 2026 saying it would be removed, withdrew that notice the same day, then relisted the company in June without explaining the reversal. The complaint calls the decisions "arbitrary." Defense declines to comment on live litigation.

This is now a standard play. Xiaomi sued and got itself delisted in 2021; YMTC filed in December 2025, WuXi AppTec and Alibaba in June. Chinese champions have decided the US Administrative Procedure Act is a usable instrument — and CXMT chose to file three and a half weeks before Xi is expected in Washington on 24 September.

Then Friday's follow button paid off. I noted that CXMT's new Weibo account followed Xiaomi's phone division first. Saturday, the two confirmed that CXMT's LPDDR6 will debut in the Xiaomi 18 Fold next month alongside Xiaomi's in-house 3nm Xring O3. Domestic memory, domestic silicon, one flagship, one lawsuit.

And memory has passed the processor. A hardware leaker relayed by GeekPark puts a 12+256GB memory package at roughly 2,200 yuan (~$310) to procure against about 500 yuan (~$70) a year ago, while the newest flagship SoC went from ~1,500 to 2,000 yuan (~$282). Trade chatter, but the ordering is the point: memory is now the largest line in a phone's bill of materials.

Go deeper on this section: ClaudeChatGPT

Xi lands in Bishkek before he lands in Washington

Xi meets Putin on Monday. The Kremlin says Power of Siberia 2 will come up on the sidelines of the Shanghai Cooperation Organisation summit in the Kyrgyz capital — their second meeting this year. Foreign-policy aide Yury Ushakov said the issue "will be discussed bilaterally," per SCMP. The 1,600-mile line would move up to 50bn cubic metres a year via Mongolia.

The tell is who keeps announcing it. A memorandum exists; a price never has. Moscow puts the pipeline on the public agenda, Beijing declines to say what it will pay, and Chinese readouts of these meetings lead on multipolarity rather than gas. That pattern survived Putin's Beijing visit in May intact.

Treasury kept firing anyway. The US sanctioned a Hong Kong company Friday for allegedly laundering funds for a designated Iranian exchange house, the latest of several Chinese entities hit in days, with Beijing promising retaliation, SCMP reports. Nobody is protecting the atmosphere before 24 September.

Quieter statecraft: AI is now written into the family-doctor contract. A health-commission notice known as document 273 pulls hospital specialists into primary-care contracting and widens the disease list to include infectious disease by end-2027, then tells providers to "deeply advance" AI in those services — analysed in TMTPost. Follow-up volume rises; grassroots headcount does not.

The vendors are already sized for it. iFlytek's medical assistant covers 828 counties and over 77,000 primary institutions, claims above 80% share where AI triage is deployed, and grew that revenue 47.5% in the half. In Tianjin, WeDoctor reports one health manager covering 2,600 patients, about 3.5 times conventional throughput.

Go deeper on this section: ClaudeChatGPT

Alibaba went sparse, Meitu went to the bank

Qwen3.8-Flash shipped Wednesday night, open-weight as Qwen3.8-Flash-Next: a multimodal mixture-of-experts model with 125B main parameters plus 51B of N-gram embedding, only 6B active per token, at 0.8 yuan (~$0.11) per million input tokens and 2.7 (~$0.38) out.

The attention stack is the interesting part. Three of every four layers use gated delta networks compressing history into fixed-size state; the fourth keeps global attention for exact retrieval, with a compressed indexer selecting micro-blocks rather than re-reading context. Alibaba claims up to 7.6x prefill and 4.9x decode kernel speedups at 1M tokens, and in its Qwen office agent, 75% less token consumption per task — InfoQ has both.

Meitu is the one actually banking AI money. The 18-year-old photo-editing company reported first-half revenue of 2.21bn yuan (~$311m), up 22.1%, adjusted net profit 650m yuan (~$92m), up 39.5%, on 18.44m paying subscribers. Its productivity apps carry ARR of about 620m yuan (~$87m), and compute-credit spending roughly doubled across the half.

Set that against ByteDance's Doubao. A Jiemian commentary puts Doubao's daily revenue under 1m yuan (~$141,000) against 200m-plus daily users, and cites trackers showing 18.65m monthly users lost from May to July after a paid tier arrived. Selling a workflow tax beats selling a platform, for now.

Go deeper on this section: ClaudeChatGPT

A robot made 55-step ice cream in Shanghai

Sharpa put a two-armed robot behind a DQ counter. At the chain's Wujiang Road store in Shanghai it takes a metal holder from a steriliser, pulls a paper cup, draws soft serve, adds three scoops of Oreo, blends, and inverts the cup without it falling — the upside-down test the brand has used since 1940. TMTPost watched the cycle.

The specifics make it a product, not a demo. 55 steps, about 90% success, touch feedback in 98% of them, and zero modification to a store that imports every fixture. Chief executive Li Yifan, also a co-founder of lidar maker Hesai, calls purpose-built robot environments "phone booths" and says doing 54 of 55 steps creates a job rather than filling one. Sharpa has raised over 4.5bn yuan (~$634m) at above 20bn yuan (~$2.8bn).

Now read the Beijing robot games two ways. Solidot records Tiangong, from the Beijing Humanoid Robot Innovation Centre, at 8.64 seconds over 100 metres against Bolt's 9.58 — then spends the piece on why it isn't sprinting: nearly a second to react against Bolt's 0.146, more than 50 steps against his 41, no deceleration, and a stop achieved by hitting a crash mat.

AgiBot's version is a medal table. New Wisdom relays 18 golds and 46 medals, including seven of eight dexterous-hand events and six of twelve scenario events — library shelving, hotel rooms, fire response — on production units. Western wires led on the record and used the crash mats as colour; CBS mentioned the stretchers. Correction: I gave Tiangong 8.86 on 25 August. The scenario events are the real scoreboard.

Go deeper on this section: ClaudeChatGPT

A programmer died in a company toilet

Shenzhen refused to call it a workplace injury. A 39-year-old programmer parked in his office building's second basement at 8:56am on 23 April, asked a guard where the second-floor toilet was at 9:00, went in a minute later, and was found unconscious there at 11:28. He was declared dead at 12:50. Solidot has the file.

The stated reason is the story. The city's human-resources bureau ruled on 3 July that his desk was on the 11th floor, that he had clocked in and gone to the second-floor toilet without going to it, and that he was doing nothing work-related. Recognition is worth 1,130,040 yuan (~$159,000) plus funeral and dependants' payments. The family is appealing. Two and a half hours passed before anyone looked.

A second argument is running between two cities. In July, Hong Kong police searched two independent bookstores on suspicion of sedition and arrested five people; one clerk was led away in a T-shirt reading "I am a bookstore clerk." Initium's podcast traces the slogan to Taiwanese bookstores, concert stages and a post by Taiwan's president — then to a fight over whether that was solidarity or consumption.

Go deeper on this section: ClaudeChatGPT

Threads we are pulling

  • MiniMax: the ARR figure I quoted twice is the softest number in the report. TMTPost's read of the interim: recognised first-half revenue was $116.6m against the $800m annualised claim, which Huxiu suspects is one week times 52 and unaudited. Gross margin came in at 17.9% against Morgan Stanley's 29.5%, because enterprise volume was bought with price cuts — M3 doubled API prices on 1 June, the stock fell 15.71% that day, permanent half-price followed within a week. Adjusted net loss $293m, up 111.2%.
  • Gyirong passed 72 hours; Yunnan started flooding. Excavators are reopening the route to the Gyirong border checkpoint with the barrier lake still a breach risk, and CCTV named the officer, Huang Peng, 34, filmed directing people away minutes before the mudslide. Meanwhile Mohan, on the Laos border, took 187.5mm (7.4 inches) of rain in 24 hours; 2,226 people moved, provincial response at level three, 20,500 relief items plus 10,000 from central stores.
  • The discipline machine went back to county work. A day after a legislative committee stripped the army's two most senior officers of their seats, the same apparatus expelled a former vice mayor of Tongliao in Inner Mongolia for, among other things, falsifying shantytown-redevelopment figures — a political-performance charge, not merely a bribery one — plus a former deputy chair of Ulanqab's legislature, with a Chongqing Airport Group executive newly under investigation.
  • Biren's first post-IPO half is 20x revenue and 23% cash. Revenue 1.236bn yuan (~$174m), up 1,997.6% and more than all of 2025; loss narrowed 76% to 380m yuan (~$54m); gross margin 42.7%, down from 53.8% last year as cluster deals drag in third-party hardware. Receivables 1.117bn yuan (~$157m), collected at 23.3%; R&D exceeds gross profit. AIX notes its HK$101.7bn (~$13bn) cap is a third of A-share peers with the same losses.

Go deeper on this section: ClaudeChatGPT

The river · 34

Inference software can change a model's answer even with identical weights

# QbitAI 16h

A long-context test found that attention kernels, KV-cache precision, quantization, and tensor parallelism changed token choices and broke tool calls on the same GPU and weights.

More
  • The experiment captured logits across more than 100,000 tokens from a Qwen model and found that switching attention backends eventually changed the top-ranked token, despite identical prompts, hardware, weights, and forced histories.
  • BF16 KV cache stayed stable, while INT4 produced rapidly rising top-token flips and failed tool calls; INT8 sometimes recovered, showing why short-context spot checks can miss production failures.
  • The causal mechanism is numerical drift from floating-point precision, accumulation order, and CUDA kernel behavior, amplified through autoregressive generation rather than a semantic difference in the model weights.
  • The test found that a community INT8 weight-quantization path preserved activation precision better than some official FP8 and FP4 paths, while an FP4 path approached a 50 percent flip rate near 88,000 tokens in that setup.
  • A striking tensor-parallel result was that one, two, and four GPU configurations did not behave monotonically: single- and four-GPU runs completed a tool call while the two-GPU run failed.
  • The source is a forum experiment, not a peer-reviewed benchmark, but it exposes a real systems boundary: inference reproducibility belongs to the full kernel, cache, quantization, and parallelism stack.
  • For agent operators, long-context regression tests should include exact tool-call trajectories and numerical stability, not just answer quality on short prompts.
Open at QbitAI Ask Claude ChatGPT

A returned scholar becomes a Tsinghua AI class adviser

# QbitAI 16h

Xu Mengdi returned from Stanford to Tsinghua in 2025 and now advises its elite AI cohort, bringing a robotics view that emphasizes adaptation over simply scaling models and data.

More
  • Xu studied vehicle engineering before moving through robotics and machine learning at Johns Hopkins and Carnegie Mellon, then working with Toyota Research, MIT-IBM, Google DeepMind, and Stanford labs.
  • Her RoboTool work lets robots choose tools and compose executable plans from goals, scenes, and capabilities, including unconventional uses such as turning objects into bridges or hooks.
  • Her central research claim is that a robot cannot encounter every future situation in its training set, so generality requires fast adaptation from a small number of demonstrations.
  • Prompting Decision Transformer applies an in-context-learning idea to reinforcement learning by using short task trajectories to infer a new goal without updating model parameters.
  • The appointment is a talent signal: elite Chinese AI education is recruiting researchers who connect foundation-model ideas to physical control, safety, and real-world deployment.
  • It also shows the value of return migration beyond headcount, as overseas research experience is being folded into the training of China's next generation of AI engineers.
Open at QbitAI Ask Claude ChatGPT

China says 261 foreign nationals remain missing after Tibet floods

# SCMP 42m

Chinese authorities say 261 foreign citizens from 23 countries are missing after deadly floods near the Tibet-Nepal border, including 18 Americans and 104 Nepalis.

Also: Jiemian

More
  • The nationality breakdown makes the disaster an international consular problem as well as a regional emergency, with people from India, the United States, Britain, Canada, Australia, and several European countries among those missing.
  • The report attributes the figures to Tibetan authorities and does not provide an independent accounting or a confirmed explanation for the missing people.
  • Chinese coverage leads with the official count and the multinational nature of the victims, signaling the scale of a disaster occurring in a tightly controlled border region.
Open at SCMP Ask Claude ChatGPT

A Chinese coastguard war game puts weapons use at 0.3 percent

# SCMP 15h

A Chinese coastguard simulation estimates a very low chance of firing in a disputed-water encounter, revealing how researchers model escalation, ambiguity, and restraint.

More
  • The scenarios involve fishing boats or small fast craft probing a patrol ship, forcing operators to distinguish ordinary activity from coordinated attempts to expose a weakness.
  • The reported 0.3 percent weapons-use probability is a model output, not a forecast or evidence that real-world confrontations are safe.
  • The technical value is in the decision problem: Chinese researchers are treating maritime escalation as a classification and control problem under incomplete information, not simply a question of force superiority.
  • The exercise highlights a preference for calibrated coercion, where coastguard presence and maneuvering achieve pressure while avoiding the political cost of a shot.
  • The number should be read alongside assumptions about actor behavior, rules of engagement, sensor confidence, and repeated encounters; changing those inputs could radically alter the result.
  • The broader signal is that China is institutionalizing gray-zone scenarios and quantifying escalation risk around disputed waters.
Open at SCMP Ask Claude ChatGPT

China's Middle East strategy faces a test as the Iran war drags on

# SCMP 18h

Six months into the US-Israeli war with Iran, Gulf states are building more self-reliant defense networks, challenging China's strategy of regional involvement without firm security commitments.

More
  • The report says Gulf governments increasingly doubt that either Washington or Beijing can protect them from attacks, a judgment that limits China's leverage even as its economic role grows.
  • China's approach has emphasized diplomacy, trade, and political relationships without extending treaty-style military guarantees; the conflict exposes the gap between influence and responsibility.
  • A regional shift toward autonomous defense procurement and partnerships could reduce the value of Chinese diplomatic balancing while increasing pressure on Beijing to clarify its red lines.
  • The issue is not simply whether China can mediate: it is whether Gulf states will continue treating China as a strategic alternative if security still depends on other networks.
  • The story places China's Middle East policy at a crossroads between avoiding entanglement and accepting the costs required to become a credible security actor.
Open at SCMP Ask Claude ChatGPT

Apple study finds Chinese reasoning training nearly matches English

# New Wisdom 5h

A study across nine base models and 11 languages found only a 1.1-point Chinese penalty for native-language reasoning training, challenging the assumption that serious reasoning must happen in English.

More
  • The experiments compared reinforcement-learning reasoning rewards delivered in English with rewards delivered in the questioner's language, using more than 200 controlled runs across models from 1B to 8B parameters.
  • Chinese, French, German, and Japanese stayed within roughly two points of English, while lower-resource languages suffered larger average gaps, tying the result to pretraining data depth.
  • Cross-language transfer was broad: one language's training often improved others, and mixed-language training came within 0.3 points of an ideal scheme that separately optimized every target language.
  • The engineering implication is meaningful for Chinese model builders: native-language reinforcement learning can preserve control over reasoning language without automatically sacrificing much capability.
  • The caveats are substantial: the study used open models below 8B parameters, synthetic math and logic tasks, and single-seed experiments without per-configuration hyperparameter tuning.
  • Some model-language combinations suffered severe regressions, including drops above 19 points on unseen English-containing tasks and larger losses on difficult math, so multilingual training needs compatibility testing rather than a universal recipe.
Open at New Wisdom Ask Claude ChatGPT

Unitree and UBTECH are close in sales, but not in profit

# TMTPost 3h

UBTECH's humanoid-robot revenue reached 1.27 billion yuan (about $1.8 billion) in the first half, roughly matching Unitree, but it still lost money while Unitree reported profit.

More
  • UBTECH's total revenue was 1.269 billion yuan (about $179 million), versus Unitree's 1.152 billion yuan (roughly $162 million); excluding a newly consolidated industrial business, UBTECH's robot revenue was slightly lower.
  • UBTECH's full-size humanoid revenue jumped 1,445 percent to 590 million yuan (about $83 million), with 921 units sold and a reported 66.8 percent gross margin.
  • Unitree reported 274 million yuan (about $39 million) of net profit, but its 244 million yuan (roughly $34 million) adjusted profit is the cleaner comparison because a prior employee-equity charge distorted the base period.
  • The companies represent different commercialization strategies: UBTECH emphasizes industrial scenarios, customer delivery, and a planned 10,000-unit factory, while Unitree historically used hardware efficiency to scale and is now increasing model and brand spending.
  • UBTECH remains loss-making, but its adjusted EBITDA loss narrowed 45.9 percent; Unitree's adjusted profit fell 19.3 percent, suggesting neither headline number alone settles the competitive question.
  • The durable test is whether robots work repeatedly in customer workflows and whether either company can turn high gross margins into cash after research, sales, and deployment costs.
Open at TMTPost Ask Claude ChatGPT

China flies an aluminum-ice rocket to a record 5.3 kilometers

# IT Home 5m

A Beijing university and private space firm flew Asia's first aluminum-ice sounding rocket, more than ten times the previous public record, advancing a propulsion route tied to using lunar materials.

More
  • The aluminum-ice propellant combines aluminum powder and water in a frozen mixture; both materials are potentially available on the Moon, making this an in-situ resource utilization technology rather than just a new terrestrial fuel.
  • The joint team reported a 5,276-meter measured altitude on August 24, exceeding the roughly 400-meter public record from a 2009 US university test supported by NASA.
  • The result is an engineering demonstration, not evidence that lunar fuel production or an operational launch system is close; the next bottlenecks are sustained propulsion, storage, feed systems, and repeatable manufacturing.
  • Chinese coverage presents the flight as a milestone in the national lunar program and links it to the strategic race created by both Chinese and US crewed-lunar efforts.
  • The team says it will pursue engineering validation around a lunar-polar resource-use program, so the meaningful follow-up is whether the concept survives larger motors and longer burns.
Open at IT Home Ask Claude ChatGPT

Chinese-backed humanoid robot factory opens in Serbia

# IT Home 1h

A factory built by Chinese auto supplier Minth has begun assembling humanoid robots and robot dogs in Serbia, with planned output of up to 20,000 units a year for European and global markets.

More
  • The first phase costs 20 million euros and covers assembly, motion and vision calibration, and performance testing; a planned industrial park would take total investment to about 200 million euros.
  • The project is not merely a plant relocation: Minth says Serbian staff trained at its Chinese headquarters, exporting supply-chain know-how, production methods, and workforce development.
  • The factory is presented as Europe's first humanoid-robot mass-production base, but the source does not provide customer orders, automation rates, or evidence of sustained volume production.
  • The project shows how Chinese manufacturers are using overseas plants to internationalize emerging hardware supply chains while embedding local labor and training into the expansion.
  • The strategic test is whether Serbia becomes a meaningful European production node or mainly a symbolic assembly and market-access platform.
Open at IT Home Ask Claude ChatGPT

Anthropic's automated researcher improves AI safety at machine speed

# AI QbitAI 16h

Anthropic reports that Claude searched papers, designed training, and improved ten alignment tasks at roughly $4 an hour, while acknowledging that humans still define the goals and tests.

Also: New Wisdom

More
  • The automated alignment researcher closes a loop from literature search to hypothesis, data generation, fine-tuning, and evaluation, with each 30-minute experiment informing the next.
  • On a deception task, the best automated method reportedly closed 82 percent of the safety gap and averaged 85 percent across runs, compared with 20 percent for six human researchers given one attempt each.
  • The comparison is not fully fair: Claude could run repeated experiments and learn from outcomes, while human experts submitted a single method, so the result measures iterative search as much as research insight.
  • A weaker Sonnet model reportedly improved an early Opus model, closing about 65 percent of the selected gap with more than 2,000 examples and approaching the production model's 72 percent.
  • The reported API cost was about $4 per hour versus $150 per hour for human researchers, but the cheap loop depends on access to compute, a target model, evaluation infrastructure, and human-defined boundaries.
  • Anthropic's own caveat is the important one: the system optimizes what the benchmark specifies, and a more compliant model can still fail on risks or capabilities outside the test suite.
Open at QbitAI Ask Claude ChatGPT

Tencent releases a 770-billion-parameter Hunyuan preview model

# AI GeekPark 23h

Tencent open-sourced Hunyuan Hy4 preview with 770 billion total parameters, a 1-million-token context window, and low API prices aimed at office, coding, and research workloads.

More
  • Hy4 expands from Hy3's 295 billion total parameters and 21 billion activated parameters to 770 billion and 49 billion, while quadrupling the advertised context window from 256K to 1M tokens.
  • Tencent prices input at 6 yuan per million tokens (about $0.85), output at 18 yuan (roughly $2.54), and cached input at 0.3 yuan (about $0.04), positioning it as a high-volume enterprise utility.
  • The company says 163 experts blind-tested 203 engineering tasks and gave Hy4 an average 2.99 out of 4, narrowly ahead of two other Chinese models; these are internal evaluation results, not an open benchmark.
  • Tencent also says the model contributed to optimizing its own training methods, data strategy, evaluation, and low-level operators, improving inference throughput 31.8 percent against a baseline.
  • The real product question is long-horizon execution: reading many documents, calling tools, recovering from errors, and delivering usable software rather than producing attractive one-shot text.
  • Hy4's deployment across Tencent's office, coding, content, and assistant products shows a vertically integrated Chinese model strategy in which distribution can matter as much as benchmark rank.
Open at GeekPark Ask Claude ChatGPT

OpenClaw's fad is giving way to agent harnesses

# Analysis QbitAI 17h

A Chinese analysis argues that OpenClaw has not died but has been absorbed into vendor products, as developers move from assembling personal agents to using integrated runtimes and harnesses.

More
  • The article says OpenClaw reached more than 380,000 GitHub stars and spawned over 30 domestic derivatives, including products from major Chinese internet companies.
  • Its original appeal was permission-rich local autonomy: connecting email, browsers, terminals, calendars, and chat tools to a model that could run continuously without supervision.
  • The newer competition packages context management, file operations, command execution, sandboxing, long tasks, and tool orchestration into first-party agent runtimes such as Claude Code and Codex.
  • This is a useful product-cycle observation: open-source projects can define an interface and market before incumbents absorb the underlying workflow into more reliable, managed systems.
  • Chinese coverage treats the shift as a move from novelty and installation services toward infrastructure and harness quality, where stateful context, permissions, and recovery determine daily usefulness.
  • The question is whether OpenClaw retains a community advantage in extensibility, or whether vendors win by making the same capabilities safer and less labor-intensive.
Open at QbitAI Ask Claude ChatGPT

Alibaba's Qoder moves coding agents from IDEs to desktop workflows

# AI QbitAI 17h

Alibaba says Qoder can turn natural-language requests into tested applications and modify existing repositories, using a plan-code-run-verify loop rather than merely generating snippets.

More
  • The agent reads an existing repository, infers its architecture, produces a change plan, edits multiple files, runs the project in a real environment, and revises when the result fails.
  • Its computer-use and browser-use layers matter more than the desktop mascot: they let the model verify generated software against actual interfaces instead of stopping at syntactically plausible code.
  • The company says Qoder has served 6 million users and 100,000 businesses, and supports more than 40 connectors, 70 plugins, and 20,000 skills; these are company-reported adoption figures.
  • The product reflects China's push to make agentic coding a general business tool, not just an IDE assistant, by connecting repositories to project management, cloud services, and internal systems.
  • The unresolved risk is governance: automatic approval and cross-system execution can turn a small coding error into a production change unless permissions, sandboxes, and review gates are explicit.
Open at QbitAI Ask Claude ChatGPT

Xi and Putin may revive talks on Power of Siberia 2

# Policy SCMP 16h

The Kremlin says Xi Jinping and Vladimir Putin are expected to discuss the Power of Siberia 2 gas pipeline in Bishkek, underscoring the strategic depth of the China-Russia relationship.

More
  • The meeting would be the leaders' second face-to-face encounter of the year and comes after Putin's visit to Beijing following a US presidential visit there.
  • Power of Siberia 2 would create another major gas link from Russia toward China, but the report confirms discussion rather than a signed construction or pricing agreement.
  • For Beijing, the project could diversify pipeline supply and deepen continental energy interdependence; for Moscow, it offers an alternative market as relations with the West remain strained.
  • The timing shows how Central Asian diplomacy, energy security, and great-power signaling are being negotiated together rather than as separate commercial files.
  • The meaningful follow-up is whether the two sides resolve route, financing, capacity, and pricing questions that have kept the project politically attractive but commercially unsettled.
Open at SCMP Ask Claude ChatGPT

China and Russia revisit an Arctic route through the Irtysh

# Policy SCMP 7h

Xi Jinping and Vladimir Putin are expected to discuss a river-to-Arctic cargo route linking China through Kazakhstan and Russia, a concept that could reduce exposure to contested sea lanes.

More
  • The proposed corridor would use the 4,248-kilometer Irtysh River, which begins in China's Xinjiang region, crosses Kazakhstan, and joins Russia's Ob River before reaching the Arctic Ocean.
  • Its attraction is strategic geography: an inland waterway could complement maritime routes that are vulnerable to conflict, chokepoints, sanctions, and naval pressure.
  • The route would face major engineering, seasonal, environmental, sovereignty, and commercial constraints; the report describes talks, not an approved infrastructure project.
  • The political value may arrive before the logistics: China and Russia can use the corridor to signal deeper continental integration even if cargo economics remain uncertain.
  • The story belongs in the context of China's search for resilient trade and energy corridors across Central Asia and the Arctic.
Open at SCMP Ask Claude ChatGPT

A Shenzhen programmer's death tests the limits of Chinese workplace protection

# Culture Solidot 19h

Authorities rejected a family's claim that a programmer who died in his office building's restroom suffered a work injury, arguing he had not reached his desk; the family is appealing.

More
  • The 39-year-old reportedly entered the building, went to a second-floor restroom, and was found unconscious more than two hours later before being pronounced dead.
  • The disputed issue is whether a workplace restroom is a reasonable extension of the job site or whether the worker must first reach the assigned workstation, a narrow rule with large consequences for families.
  • Recognition could unlock about 1.13 million yuan (roughly $159,000) plus funeral and survivor benefits, giving the legal classification direct economic weight.
  • The case exposes a recurring gap in China's labor system: formal workplace boundaries can lag behind how employees actually move through employer-controlled premises.
  • The source includes a lawyer's interpretation but no court ruling, so the agency's decision is not the final legal position.
Open at Solidot Ask Claude ChatGPT

Huawei pitches hybrid-cloud data infrastructure built for agents

# Policy Leiphone 21h

Huawei says its hybrid-cloud stack now links governance, multimodal data lakes, trusted data spaces, and agent access, reflecting China's effort to make enterprise data usable without moving it all to a central cloud.

More
  • The architecture separates reliable management, efficient supply, intelligent use, and trusted circulation, connecting cloud, data-center, factory, and office data under a hybrid-cloud control plane.
  • Huawei describes cross-region disaster recovery, billion-scale metadata management, multiple data engines, and data access without relocating every source, aimed at enterprises that cannot surrender operational or industrial data.
  • The platform also uses privacy computing, APIs, sandboxes, and blockchain to create auditable data-sharing environments across organizations, a direct response to data sovereignty and provenance concerns.
  • Chinese industry examples include a power utility's multidimensional labeling system, a regional data chain covering nine provinces and 40 cities, and city-level trusted data spaces.
  • The policy and business signal is that Chinese AI deployment is being framed as a data-infrastructure problem before it is a model problem: agents need governed, current, machine-readable enterprise context.
  • The claims come from a Huawei-led conference and partner case studies, so independent measures of latency, accuracy, cost, and real production dependence are still missing.
Open at Leiphone Ask Claude ChatGPT

China spends a decade building a national map of its languages

# Culture SCMP 20h

A new Hunan museum hall showcases dialects, minority languages, and oral culture collected across more than 1,800 sites over ten years, turning linguistic preservation into a technology project.

More
  • Visitors can speak into a recognition booth that identifies their likely hometown dialect, making a large archival effort legible through an interactive public interface.
  • The collection covers dialects, minority languages, and oral traditions, preserving material that is vulnerable to migration, schooling, media standardization, and generational change.
  • The project shows China's state capacity at its most cultural: museums, field researchers, language technology, and national data collection are being combined to document diversity while the country also promotes a common standard language.
  • The source does not specify the corpus size, licensing terms, or whether the data will be available for commercial model training, which will determine its value to speech and language AI.
  • The strategic importance is both cultural and technical: language archives can support education and preservation, but they also become infrastructure for recognition systems and future generative models.
Open at SCMP Ask Claude ChatGPT

Huawei previews a more protective wing-fold tri-fold phone

# AI Leiphone 17h

Huawei's next luxury tri-fold phone encloses its inner display when closed and uses a centered hinge layout, extending the company's lead in China's ultra-premium foldable market.

Also: IT Home

More
  • The new wing-fold form is intended to reduce abrasion and impact exposure compared with Huawei's earlier outward-facing Z-fold arrangement.
  • The company is signaling that the next competitive frontier is reliability and ergonomics: hinge precision, weight balance, crease control, and resistance to drops matter as much as screen area.
  • Huawei's prior tri-fold reportedly attracted more than six million reservations and sold nearly 500,000 units, according to the article, but those figures are company and media estimates rather than audited demand data.
  • The product's high-end materials and distinctive styling show how Chinese phone makers use mechanical novelty and luxury branding to defend premium pricing.
Open at Leiphone Ask Claude ChatGPT

Sunwoda posts record revenue but much weaker underlying profit

# Signal TMTPost 2h

Sunwoda's first-half revenue reached 38.18 billion yuan (about $5.4 billion), while reported net profit fell 30 percent; HEV batteries and data-center storage are the strategic bright spots.

More
  • Net profit was 603 million yuan (roughly $85 million), while adjusted net profit fell to 97 million yuan (about $14 million), showing how accounting and hedging effects obscure operating performance.
  • Second-quarter net profit reached 489 million yuan (about $69 million), up 328 percent sequentially, but the report warns that foreign-exchange and hedge accounting explain part of the rebound.
  • The company reportedly became the world's top HEV lithium-battery supplier by installations in the second quarter, entering a segment long dominated by Japanese and Korean manufacturers.
  • Its 684Ah stacked cell has passed 2 million units of cumulative production, and a 588Ah wound cell entered volume production, positioning Sunwoda for AI data-center storage demand.
  • The strategic pattern is a Chinese second-tier battery maker buying growth through specialized niches rather than trying to match the scale economics of CATL or BYD.
  • Negative operating cash flow, rising leverage, and expanding receivables are the constraints; sustained shipments and adjusted margin matter more than headline revenue.
Open at TMTPost Ask Claude ChatGPT

Meitu is finally monetizing AI, but its moat is still narrow

# AI TMTPost 3h

Meitu's first-half revenue rose 22 percent to 2.21 billion yuan (about $311 million), with adjusted profit up 40 percent as AI tools push users from subscriptions toward usage-based spending.

More
  • Paid subscribers exceeded 18.44 million, while image and design products generated 1.77 billion yuan (roughly $249 million), showing that the legacy consumer base still funds the transition.
  • Productivity apps had 2.35 million paying users, 330 million monthly active users, and about 620 million yuan (roughly $87 million) in annual recurring revenue; their users consume more AI compute and pay more than lifestyle users.
  • Meitu is adding an AI compute-credit layer to subscriptions, allowing revenue to scale with completed work rather than only with seats, a closer analogue to a workflow toll than a model subscription.
  • The company benefits from an existing aesthetic workflow and users already accustomed to paying for image tools, unlike general-purpose assistants that must create willingness to pay from scratch.
  • The risks are low subscription penetration, a much lower price ceiling than Adobe, and pressure from Chinese general models and global design platforms that can replicate image features.
  • Meitu's overseas growth suggests a differentiated export path based on taste and workflow rather than recommendation algorithms, but sustained retention will decide whether this is a business or a favorable product cycle.
Open at TMTPost Ask Claude ChatGPT

Chinese scholars warn that AI can make Beijing sound more threatening

# Policy SCMP 23h

A state-affiliated Chinese academic paper urges human review of AI translations because literal renderings of political language may distort Beijing's intent and worsen foreign perceptions.

More
  • The recommendation is for stronger human translation capacity and oversight as AI takes over more diplomatic and policy translation work.
  • The concern is not just wording quality: political language often relies on institutional context, calibrated ambiguity, and audience-specific conventions that a literal model may flatten into threats.
  • The story reveals an official Chinese anxiety about losing control of external signaling even while adopting AI internally, especially when machine output can harden language for foreign readers.
  • Human translators are being positioned as strategic interpreters rather than a temporary legacy layer, suggesting that high-stakes state communication will remain hybrid.
  • The missing evidence is any measured comparison of AI and human translations or a concrete policy mandate, so this is a warning about governance rather than a completed rule.
Open at SCMP Ask Claude ChatGPT

Shanghai becomes a major Chinese fusion-finance hub

# Signal Jiemian 35m

Chinese fusion startups raised about 8 billion yuan (roughly $1.1 billion) in first-half primary financing, while Shanghai's cumulative sector funding approaches 18 billion yuan (about $2.5 billion).

More
  • The reported first-half total covers roughly 30 primary-market events, up about eightfold in count and more than 233-fold in value year on year, though the comparison starts from a very small base.
  • Capital is moving beyond grants for laboratory research toward private companies and engineering validation, with routes including high-temperature superconducting and spherical tokamaks.
  • One startup reportedly accumulated 2.38 billion yuan (about $335 million) this year, while another reached a post-money valuation of 10.6 billion yuan (about $1.5 billion); these are financing signals, not proof of commercial fusion.
  • The concentration of investors, talent, and varied technical approaches is exactly what Shanghai's industrial-policy model is designed to create: a dense regional platform rather than a single state champion.
  • The strategic question is whether this money buys durable hardware and supply chains or simply creates a crowded field before the physics, materials, and maintenance economics are settled.
Open at Jiemian Ask Claude ChatGPT

A yogurt challenger built a 1 billion yuan business on a blurry label

# Culture TMTPost 2h

Jiaoban grew a Chinese grain-yogurt brand to reported annual sales near 1 billion yuan (about $141 million), but investigators say its flagship drink is legally a milk beverage, not yogurt.

More
  • The brand's growth came from pairing shelf-stable yogurt-like drinks with grain toppings, live commerce, and nationwide distribution before major dairy groups copied the format.
  • Its reported 2023 revenue was 1 billion yuan (roughly $141 million), while internal sales targets cited in the article reached 1.8 billion yuan (about $254 million) for 2023 and 5 billion yuan (about $704 million) for 2024.
  • The key regulatory distinction is between a milk beverage and fermented milk: the former can contain added water and stabilizers, while the latter must meet the fermented-dairy standard.
  • Consumers found some products tasted more like sweetened drinks, and investigators found them displayed alongside genuine yogurt, creating a risk that branding outran labeling.
  • The case shows how China's fast-moving consumer market rewards packaging and channel innovation, while category boundaries and retailer practices lag behind what shoppers infer from shelf placement.
  • The follow-up question is whether large dairy companies can capture the category while the challenger retains trust, or whether the labeling issue becomes a durable liability.
Open at TMTPost Ask Claude ChatGPT

Huawei exposes on-device semantic image search and upscaling

# AI InfoQ China 15h

HarmonyOS 7 adds beta APIs for natural-language image retrieval and super-resolution, giving small Chinese app teams a packaged alternative to building their own vision stack.

More
  • Semantic search maps natural-language descriptions to indexed images, while super-resolution reconstructs detail after a user finds a low-resolution asset; together they cover discovery and reuse rather than one generic AI feature.
  • The platform abstracts embedding, retrieval, and inference infrastructure, potentially saving small teams from operating vector indexes, model-serving systems, and device-compatibility layers.
  • The article stresses that filename search, tags, and OCR remain better for exact identifiers and text, so semantic retrieval should complement rather than replace conventional indexing.
  • The APIs are beta, and the practical risks are device support, build compatibility, runtime quality, privacy consent, and maintenance when the platform changes.
  • Huawei says personal images are not retained in its cloud service, but applications remain responsible for explaining data handling and obtaining user consent.
  • The strategic implication is platform-level AI distribution: Chinese operating-system vendors can make capabilities available before app teams have the resources to build them independently.
Open at InfoQ China Ask Claude ChatGPT

Wuhan foundry unveils a decade-long plan to quadruple capacity

# Policy IT Home 26m

Wuhan Xinxin is betting on 3D integration, optical-electrical fusion and specialty memory, targeting four times today's capacity and a large domestic semiconductor ecosystem over ten years.

More
  • The plan centers on through-silicon vias, hybrid bonding, multi-wafer stacking and 2.5D silicon interposers, with the company saying these capabilities are already in volume production.
  • Its other positions are 12-inch RF-SOI and silicon-photonics platforms, back-side-illuminated manufacturing, and mainland China's largest NOR Flash operation, according to the report.
  • The proposed expansion is less about competing in leading-edge logic than about building a domestically controlled platform for sensors, connectivity, automotive electronics, optical computing, and industrial systems.
  • The target is four times current capacity, more than 1 trillion yuan (roughly $141 billion) of associated supply-chain activity, and over 10,000 semiconductor specialists across the ecosystem.
  • The scale of the commitment and the presence of major industry and academic figures show how Chinese policy is consolidating specialty-process capacity alongside the better-known advanced-node push; execution and customer demand remain unproven.
Open at IT Home Ask Claude ChatGPT

China explores commercial pricing for AI tokens

# Policy Jiemian 56m

China's data regulator wants token-based pricing for AI services and will launch data-labeling pilots in 32 cities, linking data infrastructure to commercial AI use.

More
  • The proposed models include subscriptions tied to token usage and payment based on delivered results, signaling a move from generic AI access toward metered enterprise consumption.
  • The regulator also plans data-labeling pilots across manufacturing, health, and transport, where the bottleneck is often production-grade annotation rather than model availability.
  • This is an early policy signal rather than a finalized tariff regime; the important question is whether token accounting becomes a common contract and settlement layer for Chinese AI services.
  • The combination of pricing experiments, annotation hubs, and graduate-level data-skills programs shows the state treating data as an economic factor to be standardized and commercialized.
Open at Jiemian Ask Claude ChatGPT

Panda park conservation is increasing conflict with farmers

# Culture SCMP 3h

A study says human-wildlife conflict in China's Giant Panda National Park rose sharply from 2015 to 2024 as conservation expanded, making farmer losses a leading management problem.

More
  • Farmers reported damage to crops and property as well as attacks involving people and livestock, showing that habitat recovery can redistribute costs onto communities at the park boundary.
  • The park covers about 22,000 square kilometers, so the challenge is governance across a large inhabited landscape rather than simply protecting animals inside a fenced reserve.
  • The study's significance is institutional: successful conservation creates a compensation, prevention, and local-legitimacy problem that China's administrators must solve to sustain public support.
  • The source excerpt does not quantify compensation or compare species, leaving unclear whether the rise reflects more animals, better reporting, changed land use, or all three.
Open at SCMP Ask Claude ChatGPT

OpenAI ends direct model access for Cursor after its acquisition

# AI New Wisdom 5h

Chinese coverage frames OpenAI's November 12 cutoff as a warning that an AI application without its own model has no durable control after a change in ownership.

Also: GeekPark

More
  • OpenAI says it will stop supplying Cursor's direct model access after the editor was acquired by SpaceX, citing trust and alleged service-term violations involving related companies.
  • The reported transition preserves user-supplied API keys and a first-party extension but removes the bundled channel through which Cursor sold model access as part of its subscription.
  • The dispute exposes a structural dependency: application-layer companies can build distribution and revenue without owning weights, yet a change of control can still revoke their most important supplier relationship.
  • Chinese coverage is notably strategic and adversarial, presenting the episode as part of a recurring cycle of acquisitions, model lockouts, and platform conflict rather than a narrow contract matter.
  • For enterprise software buyers, the practical lesson is to price model portability, multiple providers, and contractual continuity as core architecture rather than optional resilience.
Open at New Wisdom Ask Claude ChatGPT

China-linked accounts are accused of amplifying US data-center backlash

# Policy Solidot 13h

X and SpaceX say about 200 China-linked accounts promoted opposition to US data centers, while a former Twitter executive argues the backlash reflects genuine local concerns, not simply foreign influence.

More
  • The alleged posts focused on grid strain, higher electricity prices, and data-center operators profiting at public expense, themes that also appear in domestic US opposition.
  • The report provides an accusation from a company tied to the dispute and a rebuttal from a former executive, but no independent verification of the accounts' coordination or Chinese control.
  • The Chinese dimension matters because Beijing is being invoked as a convenient explanation for a conflict that spans energy infrastructure, local land use, and technology-company legitimacy.
  • The episode shows how attribution itself has become part of the data-center fight: labeling criticism as influence operations can deflect attention from real distributional costs.
  • For technical leaders, the signal is that AI infrastructure projects now carry an information-war and community-relations burden alongside power and permitting risk.
Open at Solidot Ask Claude ChatGPT

Beijing's HICOOL turns a startup contest into industrial infrastructure

# Policy New Wisdom 5h

HICOOL drew projects from 141 countries and regions and awarded more than 110 million yuan (about $15.5 million), while Beijing uses the event to connect capital, talent, policy, and industrial deployment.

More
  • The event attracted 10,209 projects and 13,472 entrepreneurs, with 200 winners and 12 top prizes of 2 million yuan (roughly $282,000) each.
  • The winning projects span embodied AI, chips, quantum information, biomedicine, and physical-world systems, showing a preference for companies with products, customers, teams, and funding rather than slideware.
  • A notable technical theme is embodied learning through touch and real-world trial, contrasting with the industry's emphasis on larger datasets, vision-language-action models, and human labeling.
  • The event says seven years of accumulated winners have produced five listed companies, 18 unicorns, and more than 719 billion yuan (about $101 billion) in post-event financing, figures that are program claims rather than an audited causal estimate.
  • Its deeper function is ecosystem recycling: former winners return as founders, investors, and mentors, allowing Beijing to compound technical networks and policy access across startup generations.
  • The model illustrates China's industrial-policy preference for a managed pipeline where competitions, local incentives, universities, capital, and factory scenarios reinforce one another.
Open at New Wisdom Ask Claude ChatGPT

A Chinese humanoid robot beats Bolt's 100-meter time, then hits the pad

# Signal Solidot 18h

Beijing's Tiangong humanoid reportedly ran 100 meters in 8.64 seconds, faster than Usain Bolt, but its inability to brake or turn exposes the gap between peak speed and useful autonomy.

More
  • The robot averaged nearly 42 kilometers per hour, using more than 50 short steps compared with Bolt's 41 longer strides and taking almost a second to start moving.
  • Its gait is optimized around motor capability rather than human biomechanics, while the lack of deceleration forced it to stop by crashing into a safety mat.
  • The result is a sharp systems lesson: locomotion benchmarks can reward actuator power while hiding weak perception, planning, balance recovery, and terminal-state control.
  • The source quotes experts saying robots that can turn and make autonomous decisions will matter more than another increment in straight-line speed.
Open at Solidot Ask Claude ChatGPT

A celebrity lawsuit exposes China's machinery of fame and reputation

# Culture TMTPost 3h

A Chinese actor and crypto entrepreneur are fighting over more than 30 million yuan (about $4.2 million), while online debate turns a civil case into a referendum on women, money, and status.

More
  • The entrepreneur's lawyer confirmed a civil suit against the actor and her parents, while both sides have issued conflicting public statements and the case remains procedural rather than adjudicated.
  • The article's deeper subject is celebrity framing: the actor's old labels, relationships, advertising penalty, and career trajectory are being repackaged by online audiences into moral judgments about dependence and social mobility.
  • Chinese entertainment coverage routinely collapses private disputes into public narratives about class, gender, and access to resources; the speed of this case shows how social platforms can become an informal courtroom.
  • The reported claim exceeds 30 million yuan (roughly $4.2 million), but the source does not establish which allegations will survive legal review.
  • The story is useful less as a verdict on either party than as a view into how reputational capital, romantic relationships, and commercial opportunity interact in China's celebrity economy.
Open at TMTPost Ask Claude ChatGPT

Aibo discusses the next mRNA battlefield: personalized cancer vaccines

# Analysis TMTPost 2h

The CEO of Chinese mRNA developer Aibo argues that Moderna's phase-three cancer-vaccine result validates a second mRNA market, while delivery and manufacturing remain the hard problems.

More
  • Personalized cancer vaccines sequence an individual's tumor, select targets, and use mRNA to train an immune response against those targets; the commercial model is closer to one patient, one product than mass vaccination.
  • The interview frames the field's breakthrough as a convergence of platform maturity, lipid-nanoparticle delivery, and clinical execution rather than a sudden discovery of the cancer-vaccine concept.
  • Aibo's founder previously worked at Moderna and returned to China in 2019, choosing mRNA because he saw fewer domestic companies building the platform than in small interfering RNA or gene editing.
  • The Chinese angle is strategic: domestic developers are trying to build local expertise in delivery, chemistry, manufacturing, and controls while competing with Moderna and BioNTech on personalized oncology.
  • The reported US stock-market reaction is not evidence of clinical success in China; the relevant follow-up is whether Chinese programs can reproduce efficacy, manufacturing speed, and reimbursement economics.
Open at TMTPost Ask Claude ChatGPT
Source ledger
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 0/20
QbitAI 200 7/10
36Kr 200 1/30
Solidot 200 7/20
Leiphone 200 2/20
GeekPark 200 2/30
IT Home 200 12/60 +39 over cap
SCMP 200 10/50 +2 over cap
Initium 200 1/15
Jiemian 200 8/30 +20 over cap
China Digital Times 200 0/15
The Wire China 200 0/50
Techmeme 200 2/15
Hacker News 200 1/20
V2EX 200 8/50 +11 over cap
Zhidx 200 0/20
InfoQ China 200 2/20
New Wisdom 200 8/15 +1 over cap
TMTPost 200 8/18 +10 over cap