China AI & robotics developments, rewritten weekly-context-free for US readers. Published daily at 07:00 Beijing time.
Thu, Oct 22, 2026
Nvidia's Jensen Huang heads to Berlin as Europe seeks AI footingNvidia CEO Jensen Huang is set to appear at the company's GTC Berlin event, arriving as European policymakers and chip investors look to close a widening gap with the US and China in AI infrastructure.
The appearance was flagged by Chinese tech outlet Geekpark, which noted the Philadelphia Semiconductor Index had a rough stretch in the two weeks before GTC Berlin opened. The item offered no further details on Huang's agenda or announcements.
Europe has leaned on Nvidia hardware to build out sovereign AI capacity, while Brussels weighs funding and regulatory measures to support domestic model developers and chip firms.
Nvidia's GTC events typically serve as venues for product and partnership news; the company has not confirmed specifics for the Berlin stop.
Intel CTO says leading-edge wafer output can at most doubleIntel chief technology officer Pushkar Ranade said on the Dwarkesh podcast that global leading-edge wafer production cannot scale tenfold even with ample funding, with roughly a doubling the realistic ceiling — a supply constraint US AI chip buyers should factor into capacity planning.
Ranade made the comments in an interview published around Oct. 23, according to the Chinese tech outlet ITHome, which credited a reader tip. He framed the limit as physical and industrial rather than financial.
The remark lands as US hyperscalers and chip designers press foundries for more advanced-node capacity. Intel is itself building out leading-edge fabs under its foundry strategy, alongside TSMC and Samsung.
ITHome did not give further detail on Ranade's reasoning or Intel's own capacity targets. No independent confirmation of the figures was provided.
AI firms pledge public donations as political backlash growsAnthropic, OpenAI and Elon Musk's xAI are each proposing different mechanisms to direct money toward public causes, according to Bloomberg, as the AI industry faces mounting political scrutiny over its economic and social impact.
The three organizations are pitching distinct approaches to philanthropy rather than a shared framework, Bloomberg reported on Oct. 22. The pledges arrive as US policymakers and the public debate AI's effects on jobs, content and concentration of wealth.
The report does not detail donation amounts, recipients or timelines. Bloomberg framed the moves as a response to criticism rather than a coordinated industry effort.
Anthropic, OpenAI and Musk's xAI are direct competitors in frontier model development, and each has faced separate questions about safety, governance and public accountability.
Canva launches in-house AI image model, taking on OpenAI and GoogleCanva has released its first proprietary AI image generation model, moving beyond third-party integrations to compete directly with offerings from OpenAI, Microsoft, and Google, according to The Verge. The launch marks the Australian design platform's push to control its own generative AI stack as image tools become table stakes for design software.
Canva, the Sydney-based design platform founded in 2013 by Melanie Perkins, Cliff Obrecht, and Cameron Adams, has built its business on making graphic design accessible to non-designers. The company has previously relied on external AI models for features like Magic Media.
The new model is described as planning designs before generating images, a departure from prompt-to-image approaches used by rivals. Canva did not disclose pricing or a release date in the report.
The move follows similar in-house model efforts by Adobe and others seeking to differentiate on workflow rather than raw image quality.
Claix pitches five-minute retrieval setup for AI agent buildersClaix, a developer-tools startup, is marketing a retrieval layer that it says AI agent creators can configure in about five minutes, positioning the product as a way to offload database reading from the agent loop. The pitch targets a crowded field of agent infrastructure vendors competing to simplify how models query enterprise data.
According to a listing on Hacker News, the company frames the offering around a division of labor: builders handle the agent loop, tools and interface, while Claix manages how the agent reads a database.
The post offers few technical specifics, and Claix's funding, founding team and prior milestones were not disclosed in the material. The claim that setup takes five minutes is the company's own and was not independently verified.
Retrieval configuration remains a common bottleneck for agent developers, who typically stitch together vector stores, embedding pipelines and permission controls. Claix is not alone in targeting that pain point.
Wed, Oct 21, 2026
CATL shares plunge 12% as battery margins narrowCATL, the world's largest EV battery maker, saw its Shenzhen-listed A-shares fall 12.03% to 227.90 yuan on Oct. 21, 2026, a day after its third-quarter report showed shrinking gross margins despite rising sales and market share.
The selloff reflects investor concern over profitability rather than demand: CATL continues to sell more batteries and gain share, but keeps less gross profit per 100 yuan of product sold, according to the analysis.
The company, formally Contemporary Amperex Technology Co. Limited, is headquartered in Ningde, Fujian, and supplies major global automakers. Its scale has long made it a bellwether for China's EV supply chain.
The article, published by Huxiu and written by Wang Zhiyuan, attributes the drop to eroding earnings power. No company comment was cited.
Nvidia-backed Pigment raises $250M to challenge MicrosoftPigment, a business planning software maker backed by Nvidia, has raised $250 million in new funding as it seeks to compete with Microsoft in the corporate planning market, according to Bloomberg Tech.
The financing, reported by Bloomberg Tech on Oct. 21, values the company at an undisclosed amount. Pigment develops business planning and forecasting software used by large enterprises.
The raise adds to a crowded field of planning-tool vendors, where Microsoft offers overlapping capabilities through its Excel and Power BI products. Nvidia's backing gives Pigment a high-profile investor with ties to the AI computing supply chain.
Bloomberg did not disclose the round's lead investor or the company's post-money valuation. Pigment has not publicly confirmed the funding.
OpenAI models head to Google Cloud in unusual rival tie-upOpenAI's newest AI models will be made available through Google Cloud, according to Ars Technica, a notable arrangement between two companies that compete directly in frontier model development and cloud AI services.
The report, published by Ars Technica, says the models are also coming to Antigravity, a product named in the article. No further details on pricing, availability dates or which specific models are covered were included in the source material.
The arrangement is notable because Google Cloud is a direct competitor to Microsoft Azure, OpenAI's primary cloud partner, and Google develops the Gemini model family that competes with OpenAI's GPT line.
Details remain limited; the source provided no additional figures, quotes or timelines.
SpaceX, Anthropic and Nvidia lift tech stocksTech shares rebounded from a recent slump as SpaceX, Anthropic and Nvidia drew investor attention, according to WSJ Markets. The move signals renewed appetite for AI and space-related names after a soft stretch.
The report did not specify the size of the gains or which individual stocks moved most. SpaceX is the private rocket and satellite company led by Elon Musk; Anthropic is the AI developer behind the Claude models; Nvidia is the chipmaker whose GPUs underpin most large-model training.
The three names span launch services, frontier AI and semiconductors, the sectors that have driven much of the recent tech rally and its volatility. No figures or closing levels were provided in the source summary.
Tesla, IBM and Meta report earnings as AI takes center stageTesla, IBM and Meta reported quarterly earnings, with artificial intelligence emerging as a central theme across the results, according to Bloomberg Tech. The reports offer US investors a read on how AI spending and strategy are shaping the financial performance of three of the largest technology companies.
The three companies span distinct parts of the technology sector: Tesla in electric vehicles and autonomy, IBM in enterprise software and services, and Meta in social media and AI infrastructure.
The earnings arrive as AI investment remains a dominant theme for US technology investors, though the source material did not detail specific figures or guidance from any of the companies.
Additional specifics on revenue, margins and AI-related commentary were not provided in the available summary.
Anthropic launches Claude Duet AI agent, challenging Meta and OpenAIAnthropic announced Claude Duet, a new AI agent aimed at professionals and small businesses, positioning it against Meta's Muse and OpenAI's Dots in an increasingly crowded agent market.
Anthropic announced Claude Duet, an AI agent it says is designed for professionals and small businesses, according to The Verge. The launch places the company in direct competition with Meta's Muse and OpenAI's Dots.
The report did not detail pricing, availability, or the underlying model. Anthropic, founded in 2021 by former OpenAI researchers, has built its business around the Claude family of models and enterprise deployments.
Agentic AI, which lets software complete multi-step tasks with limited human input, has become a focal point for major labs this year.
Anthropic opens US waitlist for Claude Duet agentAnthropic has begun letting US subscribers join a waitlist for Claude Duet, a new agent the company is positioning against rival offerings from Meta and OpenAI, according to Engadget. The move signals intensifying competition among major AI labs to ship autonomous agent products to consumers.
The report, citing Engadget, frames Claude Duet as Anthropic's response to Meta Muse and OpenAI Dots, two competing agent products from rival labs. Anthropic did not disclose pricing, availability dates, or technical details in the summary provided.
Anthropic, the San Francisco AI developer behind the Claude chatbot, has focused its recent releases on enterprise and developer tooling. The waitlist approach is common among US AI labs seeking to manage compute demand before wider rollout.
Details on capabilities, model version, and general availability remain unconfirmed.
Microsoft links China-tied group to AI-driven attack frameworkMicrosoft says Storm-0940, a China-state-aligned actor, used an agentic AI penetration-testing framework called ARTEX to automate attacks, according to a company security blog post dated Oct. 21. The report is an early public example of AI agents handling intrusion tasks at scale.
The blog, titled for a ConnectWise supply-chain compromise, describes activity beginning as early as September 2026. Microsoft attributes Storm-0940 to Chinese state interests, a designation the company has used for other China-linked clusters.
Microsoft said ARTEX is an agentic AI penetration-testing framework, meaning software agents plan and execute steps with limited human direction. The post ties the tooling to a campaign Microsoft characterizes as large-scale and nation-state driven.
Details on targets, victims and technical indicators were not included in the summary provided. Microsoft's full post is the primary source.
GM bets on compute and AI to counter Chinese automakersGeneral Motors unveiled a fourth-generation electric vehicle platform on Wednesday, positioning onboard compute and AI as its answer to Chinese automakers' cost and software advantages. The automaker left key details undisclosed, according to TechCrunch.
GM announced the new EV platform on Wednesday, framing compute and AI as central to competing with Chinese automakers. TechCrunch reported that what GM withheld was nearly as notable as what it disclosed.
The input does not specify which details were omitted, nor the platform's timeline, pricing, or vehicle lineup. GM has previously faced pressure from lower-cost Chinese EV makers in global markets.
No figures or named executives were provided in the source material.
Top Democrat calls chip curbs key lever for US AI edgeRepresentative Jim Himes, the ranking Democrat on the House Intelligence Committee, said export restrictions on advanced semiconductors are a central tool for keeping the US ahead of China in artificial intelligence, according to SCMP Tech.
Himes argued that slowing the spread of advanced chips is essential to preserving American advantages in AI, according to the report. He framed the controls as a "key lever" in US policy toward China.
The comments come as Washington weighs further tightening of chip export rules. The ranking member's remarks signal continued bipartisan support in Congress for restricting China's access to top-tier processors.
SCMP Tech reported the remarks; no new policy measures were announced.
Anthropic launches Claude Duet agent in preview, taking on Meta, OpenAIAnthropic released Claude Duet on Oct. 21 as a research preview, an app it describes as a proactive AI agent that works continuously on a user's behalf. The move places the company in a race with Meta's Muse and OpenAI's Dots for always-on assistant products.
Anthropic, the San Francisco AI developer behind the Claude chatbot, said the app is available in a limited research preview. The company did not detail pricing or a general release date, according to the announcement.
Meta's Muse and OpenAI's Dots are also positioned as proactive assistants that act without waiting for prompts, a category drawing investment as vendors seek recurring consumer use.
Chinese tech outlet ITHome reported the launch on Oct. 22.
Tue, Oct 20, 2026
Eric Schmidt's early Anthropic stake set for large payoutFormer Google CEO Eric Schmidt holds roughly 1% of Anthropic, a position reportedly built through an early investment that is now poised to deliver a substantial windfall as the AI developer's valuation climbs.
The stake was reported by Bloomberg Tech on Oct. 20, 2026. Bloomberg did not disclose the size of Schmidt's original investment or the current paper value of the holding.
Anthropic, based in San Francisco, develops the Claude family of AI models and competes with OpenAI, Google and Meta in frontier model development.
Schmidt led Google from 2001 to 2011 and has since invested across AI and defense technology through his family office and philanthropic vehicles.
Eric Schmidt reportedly holds about 1% of AnthropicFormer Google CEO Eric Schmidt is reported to own roughly 1% of Anthropic, about half the stake attributed to CEO Dario Amodei, after investing tens of millions of dollars in the AI developer. The disclosure, reported by TechCrunch, adds detail to how early backers are positioned in one of the most valuable private AI labs.
Anthropic, founded in 2021 by former OpenAI researchers including siblings Dario and Daniela Amodei, develops the Claude family of AI models and competes with OpenAI and Google DeepMind.
The report does not specify Schmidt's investment date or terms. TechCrunch attributed the figure to reporting on his holdings.
Schmidt, who led Google from 2001 to 2011, has backed multiple AI startups through his family office and personal investments.
Google opens TorchTPU, taking aim at Nvidia's developer lock-inGoogle said on Tuesday that TorchTPU, its long-awaited PyTorch implementation for its tensor processing units, is now available as an open-source project, putting it alongside Amazon in challenging Nvidia's grip on AI application developers.
PyTorch, originally developed at Meta, is the dominant framework for building and training AI models, and most of that work has run on Nvidia GPUs. A TPU-native version lowers the cost of moving existing code onto Google's chips.
Amazon has pursued a similar strategy with its own silicon and tooling. Google did not disclose performance figures or which TPU generations TorchTPU supports, according to the announcement.
Nvidia-backed Pigment raises $250M to challenge MicrosoftPigment, a Paris-based business planning software maker backed by Nvidia, has raised $250 million in new funding as it seeks to take market share from Microsoft in corporate planning tools, according to Bloomberg Tech.
Pigment builds AI-powered software for corporate planning and forecasting, a category where Microsoft competes with its Excel and Power BI products. The company is co-founded and run by CEO Eléonore Crespo.
The round's investors and valuation were not disclosed in the source summary. Pigment has previously drawn backing from Nvidia, whose involvement reflects the chipmaker's expanding bets on enterprise AI software.
The raise adds to a run of large rounds for European enterprise software startups positioning themselves against US incumbents.