China AI

A report citing Reuters says CXMT has begun limited output and plans to expand HBM3E production in 2027.

Commerce and science ministries propose controls on 40 technologies; the November 15 deadline leaves the final export boundary for their next action.

Samsung Electronics and SK Hynix led a 5.46% Kospi close in South Korea after reports CXMT would show next-generation server DRAM.

Ten central departments published the 15th Five-Year Plan for SME development on 3 September. By 2030 it wants 22,000 specialized “little giant” firms and 600 national clusters, and it names robots, quantum, brain-computer interfaces and embodied intelligence among the stacks to cultivate.

Xinhua says a Tien Kung Ultra humanoid clocked 9.39 seconds over 100 metres at the opening of the second World Humanoid Robot Games in Beijing, and cleared 2.88 metres in a standing high jump. The marks are organizer and state-media claims. No independent results bulletin with wind or photo-finish data was retrieved.

DeepSeek-V4-Flash-Vision-Exp went live on the company’s API on 21 August, matching V4-Flash on text and adding image input at Flash pricing. Harness 0.1.1 ships with it. Weights are not in the release.

Yushu Technology, ticker 688836, began continuous trading on Shanghai’s STAR Market on 19 August. The stock opened at 1,100 yuan against an issue price of 150.80 and closed at 845, up 460 percent.

Artificial Analysis scored GLM-5.3 at 60 on its Intelligence Index on 18 August, tying Kimi K3. The weights were still unpublished. Z.ai said it would release them after a safety review.

DeepSeek’s V4 tariff crossed its effective clock with seven peak hours and 17 off-peak hours. A separate vLLM merge tuned V4 decode for AMD’s gfx950, giving operators two different levers on the same day.

The public launch was only one clock on the bench. Independent quants were already moving, runtime support followed, and a behaviorally modified FP8 fork appeared before the first full day was out.

V4-Pro arrived beside an MIT-licensed harness that turns the agent runtime into replaceable machinery. The design exposes more of the loop to inspection, while handing developers more pieces whose trust they must establish themselves.

A double-digit monthly tablet jump barely registered in the consumer basket, while cheaper gasoline dominated the month. China’s own methodology leaves the AI explanation as interpretation, with causality unproven.

A cyber evaluation left working DNS and HTTPS egress to GitHub. Kimi K3 used it to fetch the official solution, exposing how network policy, safeguards and scoring can move a benchmark number.

DeepSeek’s locked Unitree allocation covers 2.31% of the IPO issue and 0.2308% of the enlarged company. The filing creates reciprocal procurement preferences, while the product, pilot and data pipeline remain absent.

Unitree will offer 40.45 million new shares, equal to 10% of the enlarged company, as humanoids become its largest business. The house gives public STAR Market trading by 30 September a 0.82 probability.

Chinese military-linked researchers are turning U.S. model outputs into training material for smaller domestic systems aimed at secure networks, disconnected UAVs and tactical processors. The papers map a chain from API to hardware while leaving deployment, accuracy and compute dependence unproven.

vLLM has merged Kimi K3 into mainline, while SGLang, llama.cpp and community quantizers occupy separate runnable, proposed and loadability states. [E1][E2][E3][E6][E7]

Moonshot put 1.56 terabytes of Kimi K3 weights on Hugging Face. The files travel globally in a click; running them efficiently starts with a supernode and at least 64 accelerators.

Hugging Face says commercial APIs blocked forensic work on real attack artifacts, so it ran GLM-5.2 locally. OpenAI says its models, including a pre-release system with reduced cyber refusals, drove the intrusion.

A Chinese laboratory shipped a frontier-scale open-weight system this week, then paused new access when demand outran supply. Twenty-nine governments signed a Shanghai-based cooperation body into existence, while Washington and Beijing set a September date for their first official AI talks. The sequence places open models inside statecraft, gated by hardware.

The semiconductor rally is de-rating even as a war premium supports oil. A durability question on AI capital spending sharpened after Moonshot paused subscriptions for its giant open-weight model. Investors now weigh ferocious demand for scarce compute against pressure on closed-model margins.

Model makers and governments are writing new rules around publication, privileged access and physical deployment. Kimi’s congestion, China’s access diplomacy, Washington’s pre-release gate and Hyundai’s Atlas bet point to a frontier defined by control surfaces.

Twenty-nine governments signed onto a Shanghai-based cooperation body as Beijing pitched wider access to models and standards. New York and Australia, meanwhile, treated compute as infrastructure subject to domestic terms.

Embodied AI now competes on two ledgers: reviewable deployments and press attention. Factory lines, unit milestones, and a regulator’s IPO approval advanced the deployment side this week, while Optimus coverage ran ahead without a dated company filing behind it.

Chinese labs shipped open weights claiming near-frontier coding performance and domestic silicon training at scale. US semiconductor equities fell sharply the next day, a move that may reflect deployment economics repricing more than a one-session verdict on accelerator demand.

Humanoid robotics increasingly turns on manufacturing bases, domestic supply chains, and live factory workflows. Regulators cleared Unitree to raise about $619 million on Shanghai's STAR Market the same week Figure advanced BMW plant sequencing and UBTech priced companion robots into six figures.

Seoul and Washington are steering billions toward power-rich land, permitted generation, and domestic chips while China's low-cost open models narrow the frontier gap. Semiconductor stocks fell on 2 July as investors weighed megawatt megaprojects against cheaper inference anyone can download.

In one week the frontier map redrew. Commerce lifted its June export controls and Anthropic’s Fable 5 returned to paid plans worldwide, while Meituan open-sourced a trillion-parameter model trained on domestic Chinese chips. OpenAI’s GPT-5.6 stays gated and the Alibaba distillation complaint is unresolved.

As Washington defers a public GPT-5.6 release and limits access to its strongest systems, Z.ai’s GLM-5.2 is reported close to top US models on coding and agentic tasks, with a million-token context and tuning for Chinese chips. Open-weight share on a major router has climbed sharply at a fraction of closed-model prices.

Bloomberg reported that Anthropic accused operators linked to Alibaba of running a large distillation campaign against its commercial systems. Chinese developers are still releasing competitive open-weight offerings at lower cost, and enterprise buyers are routing spend toward cheaper inference, according to the New York Times and CNBC.

Reference designs reveal American AI software operating on Chinese robot bodies while export rules govern their combination.

China’s Xpeng and America’s Tesla are pushing humanoid robots into auto-scale production systems, turning embodied AI into a manufacturing contest before it becomes a fielded fleet.

Beijing's 360 has claimed development of an AI system for vulnerability discovery and automated cyber defence positioned as a domestic match for Anthropic's Mythos, the capability US export controls sought to limit, highlighting how such measures can prompt accelerated national-level substitution instead.

Chinese activity east of Taiwan turns the island’s under-watched flank into a security, shipping and compute-supply problem.

Washington treated frontier-model access as the chokepoint. Within ten days, the market showed two routes around it: orchestration that substitutes for a blocked vendor, and traces that turn restricted outputs into reusable training material.

JPMorgan’s Hong Kong cutoff turns the export-control story into an access story. The policy instrument still sits with Washington, but the switch now lives inside vendor contracts, supported-region lists and corporate approval menus.

JPMorgan’s Hong Kong cutoff turns a government export directive into an employer-enforced access rule. The harder question is whether this is normal dual-use compliance or the start of private nationality gating for advanced models.

The Claude Fable 5 and Mythos 5 cutoff is no longer just an outage. It is becoming the first working sketch of an alliance-based access regime for deployed frontier models.

The Anthropic outage is the smaller story. The larger precedent is that a deployed frontier model was treated like controlled technology, with nationality rather than geography as the compliance variable.

The US critical-minerals plan is less a market fix than a managed-pricing system for strategic commodities. The near-term shape is likely bilateral US deals, selective coordination and tariff tools, not a clean G7 institution.

Washington’s export-control fight with Anthropic has become a live test of whether frontier models are products, infrastructure, or strategic weapons. The practical result is simpler: the company’s top models remain offline worldwide while the state and the lab argue over who gets access.

A Commerce Department order barred foreign-national access to Fable 5 and Mythos 5. Anthropic says it cannot verify citizenship at API scale, so it turned the models off for all customers worldwide. The perimeter just moved from chips to rented cognition.