Highlights
- LangChain closes $300M Series C at a $3B valuation
- US and 8 allies sign the San Francisco Accord on AI chip supply chains
- Huawei Cloud open-sources embodied world model CloudRobo-WM1.2
- Nokia-led consortium plans Nordic AI gigafactory alliance
- NVIDIA benchmarks Zhipu's GLM-5.3 across dual DGX stations
Today's Headlines
[LangChain raises $300M Series C at a $3 billion valuation]
On November 15, AI agent development platform LangChain closed a $300 million Series C round at a $3 billion post-money valuation. The deal confirms that agent development, observability and governance tooling remains a strong enterprise need, with proceeds earmarked for agent monitoring and permission management. (Source: Aihehuo)
[US and eight allies sign the San Francisco Accord to govern AI hardware supply chains]
On November 15, the United States signed the San Francisco Accord with South Korea, Japan, the UK, Australia, Israel, India, Qatar and the UAE, establishing a multilateral technology-security framework for AI hardware and semiconductor supply chains. The pact is part of Washington's "Pax Silica" initiative; Germany, the Netherlands, Sweden and the EU held off pending compatibility reviews with EU economic-security rules. (Source: AIFOD)
New Products & Tech
[Huawei Cloud CloudRobo-WM1.2: open-source embodied world model tops two benchmarks]
On November 16, Huawei Cloud released and open-sourced its Ascend-native embodied world model CloudRobo-WM1.2. Built on a shared world-model backbone with a native DualAct architecture, it ranks first on both the RoboTwin 2.0 and LiberoPro 2.0 benchmarks. (Source: 36Kr)
[Baidu launches ERNIE 5.5 Flash Preview: coding-focused, low-price API]
On November 15, Baidu released ERNIE 5.5 Flash Preview with an emphasis on coding capability, launching alongside a low-price API that further cuts domestic model inference costs for startups building vertical applications. (Source: Aifaner)
IPO & Fundings
[Xixuan Chuangzhi raises 500M-yuan angel round for embodied edge computing]
On November 12, Xixuan Chuangzhi, co-founded by Moxy (Muxi) and UBTech, announced a 500 million yuan angel round to build an embodied-intelligence edge computing platform. China's chip funding wave is broad: VC/PE investment in semiconductors reached roughly 109.2 billion yuan in Q3, up 91% year over year. (Source: Semi Insights)
[Tushen Zhihe completes consecutive Pre-A and Pre-A+ rounds totaling hundreds of millions of yuan]
AI drug-design startup Tushen Zhihe recently completed two consecutive rounds totaling several hundred million yuan: the Pre-A led by Shanghai Science and Technology Venture Capital Group, and the Pre-A+ co-invested by Jinding Capital, Mega Robotics and Qingke Holdings. Funds will upgrade its ToursynClaw one-stop AI drug-design platform. (Source: 36Kr)
Open Source & Research
[NVIDIA benchmarks Zhipu's 753B GLM-5.3 at 607.8 tokens/s across two DGX stations]
NVIDIA engineers demonstrated a distributed deployment of Zhipu AI's 753-billion-parameter GLM-5.3 NVFP4 model across two DGX Station GB300 workstations linked via ConnectX-8 RDMA networking. Using vLLM tensor parallelism, the setup hit a peak generation throughput of 607.8 tokens per second at concurrency 16, offering a template for local deployment of frontier open models beyond single-node memory limits. (Source: AIFOD)
[Former OpenAI researcher revises forecast: superhuman AI researcher may arrive next year]
On November 11, former OpenAI researcher Daniel Kokotajlo said the AI Futures Project has moved its timeline for a superhuman AI researcher (SAR) to next year, suggesting AI-driven R&D acceleration may run about five months ahead of the scenarios in "AI 2027." (Source: Synced)
Company & People
[OpenAI dissolves standalone robotics team; researchers fold into general model group]
On November 13, Fortune reported exclusively that OpenAI has disbanded its standalone robotics team: research staff moved into the core research group led by chief research officer Mark Chen, while hardware and applied robotics now sit under president Greg Brockman's business unit. The company said robotics remains a long-term priority and the change reflects R&D pacing, not a retreat. (Source: Fortune)
[Giant Network reorganizes its AI lab into the central tech platform]
On November 16, Chinese game publisher Giant Network announced it is merging its AI lab teams into the company's central technology organization to tie AI more tightly to game production, pledging continued investment in its AI strategy. (Source: Jiemian)
Policies & Data
[Fed's semiannual stability report flags AI data-center leverage]
The Federal Reserve's latest semiannual Financial Stability Report devotes a section to leverage in data-center construction: as of October 2026, debt issuance tied to advanced computing infrastructure exceeded $680 billion — roughly five times the full-year 2023 total — with related public debt outstanding above $1.1 trillion. (Source: ETF Evolution)
Powered by AOYII.com