AI News, 17 November 2026

Tuesday 17 November

Commerce Faces Scrutiny for Leaving Megaspeed Off Entity List as $787M in Nvidia Chips Followed

Commerce Faces Scrutiny for Leaving Megaspeed Off Entity List as $787M in Nvidia Chips Followed

AI-generated summary The US Commerce Department faces scrutiny for omitting Megaspeed from its Entity List following reports that billions in Nvidia chips reached Chinese buyers. Chris McGuire stated that internal company restrictions failed to halt the supply of Nvidia hardware to the Singapore-based firm and its subsidiaries.

Earlier:

China Uses US AI Outputs to Train Rivals: What Reuters Found

China US AI Model Distillation: Reuters 300-Document Review Explained

AI-generated summary According to a November 17, 2026 Reuters investigation, Chinese teams are using American frontier model outputs to train domestic rivals. Reuters reports that US Commerce Department officials told Chinese regulators that collecting frontier model outputs could cross into intellectual-property theft.

Also covered byWccftechPivot NewsThe Decoder
Earlier:

Subsea data centers: pilots end, proposals meet backlash

Subsea data centers shift from pilot projects to contested proposals

AI-generated summary Rising opposition to onshore data center construction is pushing companies to test the ocean as a siting option for AI facilities. Microsoft discontinued its Project Natick subsea data center trial, while other firms continue to pursue underwater proposals in the United States and China.

Earlier:

llama.cpp and ggml Reach Version 1.0.0, Stabilizing Local AI Development

llama.cpp and ggml Reach Version 1.0.0, Stabilizing Local AI Development

AI-generated summary The open-source project llama.cpp has released llama.cpp v1.0.0 and ggml v1.0.0 after three and a half years of development. The new versions provide a stable public C API and ABI to support local AI development and commercial product integration.

Earlier:

Notre Dame Researchers Study AI Speech Recognition To Improve Healthcare For Patients With Language Impairments

AI-generated summary Researchers at the University of Notre Dame published a study in Nature Communications examining how automatic speech recognition and large language models interpret speech from adults and children with communication challenges. The team received a $1.1 million grant from the National Institutes of Health to develop digital infrastructure supporting Alzheimer's disease and related dementias care.

Alibaba adds Kimi, GLM, and DeepSeek models to QwenWork office agent

SI Agent News. SI is the new AI. Two AI podcast hosts on a newspaper front page

AI-generated summary Alibaba updated its QwenWork office agent on November 16 to include third-party AI models from Kimi, GLM, and DeepSeek alongside its own Qwen lineup. Alibaba states that the new options allow personal users and enterprise standard editions to choose specific tools based on speed, cost, and reasoning needs.

Earlier:

Mubadala CEO warns of global AI compute shortage and calls for a reset

Source-provided image accompanying Mubadala CEO warns of global AI compute shortage and calls for a reset

AI-generated summary Mubadala CEO Khaldoon Al Mubarak warned Semafor that the rapid expansion of artificial intelligence is hitting a compute shortage that could force a sector slowdown. Al Mubarak noted that Mubadala holds stakes in OpenAI and Anthropic through its MGX vehicle and maintains large investments in AI infrastructure and data-center developers.

Also covered bysemafor.comBPDATA AI News