🟡 OpenAI and Anthropic warn Chinese models gaining ground

Good Morning,

Hugging Face said an autonomous AI agent framework drove a breach of internal databases and service credentials, while Anthropic sued cybersecurity firm Abnormal AI over a logo it says is too close to its own branding.

U.S. concern over Chinese AI is rising again. OpenAI and Anthropic warned that low-cost Chinese open-weight models are gaining ground with American businesses, Moonshot AI paused new Kimi K3 subscriptions after demand overwhelmed compute capacity, and Alibaba unveiled Qwen3.8 Max as investors see China narrowing the gap with top U.S. labs.

Washington is starting to respond, with the Trump Administration exploring restrictions on advanced Chinese AI models and Senator Mark Warner introducing an AI framework that would require government testing of frontier models and new data-center disclosures.

Here’s your daily update.

The Nasdaq slipped 0.05% as a rebound in AI and semiconductor stocks was offset by rising oil prices and Treasury yields. SanDisk, Broadcom, Micron and AMD gained between 1.6% and 2.7%, while Nvidia edged up 0.2% as investors awaited earnings from major technology companies.

Hugging Face says an autonomous AI agent framework drove a breach of internal databases and service credentials. (Axios)

Anthropic sues AI cybersecurity firm Abnormal AI over a logo it says is too similar to its own branding. (Axios)

AMD rises after Microsoft expands its Azure AI infrastructure partnership around AMD’s Helios platform and EPYC processors. (Barron’s)

OpenAI and Anthropic warn that low-cost Chinese open-weight models are gaining ground with U.S. businesses. (The Wall Street Journal)

Moonshot AI pauses new Kimi K3 subscriptions after demand overwhelms compute capacity. (New York Post)

Alibaba unveils Qwen3.8 Max, an open-weight AI model that investors see narrowing the gap with Anthropic and OpenAI. (Barron’s)

Google is readying a new AI accelerator, “Frozen v2,” with Gemini capabilities built into the chip. (Investor’s Business Daily)

Harmonic launches a new mathematician-built AI benchmark with the American Institute of Mathematics. (Axios)

Trump Administration explores restrictions on advanced Chinese AI models as Kimi K3 renews national-security concerns. (Axios)

U.S. Senate sees Mark Warner introduce an AI framework requiring government testing of frontier models and new data-center disclosures. (Axios)

Sign up today