AI News

Follow the latest AI and automation news

15 Sept 20264 Rabiʻ II 1448 AH

Bill Gates Warns Governments Lag in Addressing AI Challenges

Bill Gates warned that governments worldwide are lagging in their preparedness to face the changes brought by artificial intelligence. In an interview with Reuters, he stated that no government has sufficiently studied the implications of AI, emphasizing the need for swift action to address its risks. Gates also announced that his foundation plans to invest at least one billion dollars to expand access to AI tools, focusing on education, healthcare, and agriculture.

Business

15 Sept 20264 Rabiʻ II 1448 AH

Samsung Supports Euclyd with €200 Million Funding Round

Samsung has supported the Dutch startup Euclyd in a funding round of €200 million, approximately $230 million, as investments in alternatives to Nvidia chips for AI applications increase. Samsung led the round alongside Somerset Capital Partners and the Scaleup Europe Fund, with Euclyd announcing that the funding will accelerate the development of its infrastructure for core AI models. Founded in 2024, the company is developing an AI chip system based on a different architecture than GPUs.

Business

15 Sept 20264 Rabiʻ II 1448 AH

AI start-up Mal carries Middle East FinTech in H1 with mega-funding round

FinTech funding in the Middle East and North Africa fell by 9% year-on-year in the first half of 2026 to $617 million, according to Dubai-based data firm Magnitt. Abu Dhabi-based start-up Mal, launching the world's first AI-native Islamic digital bank, accounted for 37% of total FinTech funding in the region, securing $230 million.

💡

[AINews] AEF-1 standard emerges for Third Party Evaluators, as Xai, OpenAI, and Anthropic all cosign

How can safety standards in AI be effectively enhanced?

Models

Axera Debuts 5nm M9 ADAS SoC Series With Up to About 720 TOPS

15 Sept 20264 Rabiʻ II 1448 AH

Models

15 Sept 20264 Rabiʻ II 1448 AH

Axera Debuts 5nm M9 ADAS SoC Series With Up to About 720 TOPS

Axera has announced the launch of its M9 series of advanced driver-assistance SoCs, utilizing a 5 nm process, led by the M97 and M95 models. The peak equivalent AI compute reaches approximately 720 TOPS on the M97 and 360 TOPS on the M95, featuring a shared architecture that allows software to transition between the two without a complete rewrite.

Models

Tsinghua AIR Zeva Lifts Embodied Manipulation Success From 26% to 73% Without Retraining

15 Sept 20264 Rabiʻ II 1448 AH

Models

15 Sept 20264 Rabiʻ II 1448 AH

Tsinghua AIR Zeva Lifts Embodied Manipulation Success From 26% to 73% Without Retraining

Tsinghua University, in collaboration with Domain Transform, has introduced Zeva, an embodied manipulation method. Zeva employs in-context causal memory, resulting in a success rate increase from 26% to 73% across various benchmarks. This approach allows performance improvement without the traditional data collection and fine-tuning cycle. Zeva utilizes the Cosmos3 backbone and operates on the principle of separating short-term and long-term memory, preventing noisy failures from impacting long-term memory. The process includes in-context policy injection, making site adaptation a contextual engineering challenge rather than traditional training.

Models

Tsingway and BAAI Open-Source Open3D-PIMC for 3D Compute Chips Under FlagOS

15 Sept 20264 Rabiʻ II 1448 AH

Models

15 Sept 20264 Rabiʻ II 1448 AH

Tsingway and BAAI Open-Source Open3D-PIMC for 3D Compute Chips Under FlagOS

Tsingway and the Beijing Academy of Artificial Intelligence have launched Open3D-PIMC, a programming model and software framework aimed at enhancing 3D compute chips. The project was introduced at the China Computing Power Conference and is positioned as a FlagOS pathway for next-generation 3D accelerators, with code made available to public communities. FlagOS, led by BAAI with university and chip vendor partners, targets write-once, run-across-many-chips workflows, including GPU and NPU. The initiative aims to reduce data movement and improve energy efficiency by stacking memory closer to processing units.

Models

Unisound Launches U2-Flash MoE as Post-Training RSI Flagship Flash Model

15 Sept 20264 Rabiʻ II 1448 AH

Models

15 Sept 20264 Rabiʻ II 1448 AH

Unisound Launches U2-Flash MoE as Post-Training RSI Flagship Flash Model

Unisound has announced the launch of its U2-Flash model, featuring a mixed-expert design with 266 billion parameters, making it a fully capable model. The model targets programming, mathematical reasoning, and instruction following, offering high-quality task performance at a low cost. It emphasizes post-training quality improvement using over a decade of vertical scenario data. New features include a training loop that allows the model to generate tasks and analyze trajectories, enhancing its self-improvement capabilities. The model has also achieved high scores in performance tests, reflecting its superiority over previous models.

Governance

Is Big Tech’s AI slowdown a safety pact or a cartel?

14 Sept 20263 Rabiʻ II 1448 AH

Governance

14 Sept 20263 Rabiʻ II 1448 AH

Is Big Tech’s AI slowdown a safety pact or a cartel?

Sam Altman, CEO of OpenAI, Dario Amodei of Anthropic, Demis Hassabis of Google DeepMind, and Elon Musk of SpaceX have informally agreed to slow down AI development. They stated that the aim is to "pace the frontier," partially committing to embedding third-party auditors and regulating domestic labs. However, skeptics argue that this agreement is a tactic to hinder potential competitors and stifle the open-source movement rather than to implement genuine legal safeguards. Some have even labeled it an outright "cartel" aimed at protecting their interests.

Business

Lagarde: Europe Must Become a Producer of AI Technology

14 Sept 20263 Rabiʻ II 1448 AH

Business

14 Sept 20263 Rabiʻ II 1448 AH

Lagarde: Europe Must Become a Producer of AI Technology

Christine Lagarde, President of the European Central Bank, stated that Europe must become a producer of AI technology to ensure its independence and enhance efficiency. She highlighted that the reliance on imported technology, especially from the U.S., poses risks to the European economy. Lagarde explained that AI will increasingly play a role across various sectors, from inspecting goods to monitoring patients. She warned that losing access to this technology would impact all sectors simultaneously, granting significant leverage to countries controlling it.