Tech News Roundup: Anthropic, Quantinuum, and Kubernetes Updates

This week’s technology digest highlights major developments across the industry, from a landmark legal settlement in AI that could reshape how models are trained, to massive funding in quantum computing, and crucial updates for cloud-native infrastructure and open-source software.

Anthropic’s $1.5B Settlement Over AI Training Data Sets Industry Precedent

AI company Anthropic has agreed to a $1.5 billion settlement in a major lawsuit concerning the use of pirated books to train its AI models. The lawsuit, filed by authors, alleged that the company used their copyrighted works without permission to develop its chatbots. This settlement represents a significant moment for the AI industry and could establish new standards regarding the use of copyrighted materials for training large language models (LLMs).

Sources:

WeRide Launches 24/7 Driverless Robotaxi Service in Guangzhou

Autonomous driving technology company WeRide has launched what it claims is the world’s first 24-hour, fully driverless robotaxi service in the Huangpu District of Guangzhou, China. Accessible via the WeRide Go app, the service allows users to hail a Robotaxi GXR for point-to-point trips at any time. The Robotaxi GXR is a Level 4 autonomous vehicle equipped with WeRide’s Sensor Suite 5.6, engineered to operate in diverse conditions, including fog, low light, and rain. The vehicle’s robust redundancy system—spanning its architecture, steer-by-wire chassis, braking, and computing units—is designed to prevent single-point failures and ensure passenger safety.

OpenAI Research Targets the Root Causes of AI Hallucinations

OpenAI has published new research investigating why language models sometimes produce inaccurate or fabricated information, a phenomenon known as hallucination. The research aims to provide a deeper understanding of the internal mechanics of these models and the conditions that lead to such outputs. This work is a key part of OpenAI’s broader effort to improve the safety and reliability of its AI systems. By understanding the root causes of hallucinations, researchers hope to develop methods to mitigate them and build more trustworthy AI.

Sources:

Kubernetes v1.34 Released with Stable Dynamic Resource Allocation

The latest version of the container orchestration platform, Kubernetes v1.34, has been released. A key feature of this update is the promotion of Dynamic Resource Allocation (DRA) to stable.

Sources:

Chainguard Extends Zero-CVE Promise to Virtual Machine Security

Six months after launching container host virtual machines (VMs), Chainguard is now offering application and base VM images with a promise of zero common vulnerabilities and exposures (CVEs). While renowned for its container security solutions, the company is expanding to meet customer demand for securing VM-based infrastructure. The new application VMs will come pre-loaded with popular software like Jenkins and Nginx, while the base VMs provide a secure foundation for custom applications. This strategic move signals a broader effort to provide trusted, secure artifacts for all infrastructure components.

Sources:

Kong Acquires OpenMeter to Integrate API Monetization

API management leader Kong Inc. has acquired OpenMeter, an open-source usage-based metering and billing platform. This acquisition will enable Kong to integrate OpenMeter’s capabilities directly into its Kong Konnect platform, allowing customers to implement API monetization strategies, especially for AI and large language model (LLM) APIs. The full integration is expected in early 2026, but OpenMeter will remain available as an open-source project and a SaaS offering. This move addresses the growing need for organizations to track and bill for API consumption in an AI-driven market.

Sources:

Quantinuum Hits $10B Valuation with $600M Funding for Quantum Computing

Quantinuum, a full-stack quantum computing company, has raised approximately $600 million in an equity capital round, achieving a pre-money $10 billion valuation. The funding round attracted new investors like NVentures (NVIDIA’s venture capital arm), Quanta Computer, and QED Investors, with continued support from existing shareholders including Honeywell and JPMorgan Chase. This capital will accelerate Quantinuum’s development of scalable quantum computers and support the launch of its next-generation machine, Helios. The company remains focused on achieving universal fault-tolerant quantum computing and will collaborate with NVIDIA at its Accelerated Quantum Research Center.

Sources:

China Refocuses Hong Kong Research Labs on Quantum Technology

China is restructuring Hong Kong’s state-backed research laboratories to align with national technology priorities, placing a significant emphasis on quantum science. The overhaul involves closing underperforming labs and establishing new State Key Laboratories, two of which will be dedicated to quantum information and optical quantum materials. This strategic shift aims to concentrate resources on critical fields to bolster China’s technological competitiveness against the United States. The revamped labs are tasked with advancing foundational research in quantum communication, sensing, and computing as part of a nationwide initiative to create more efficient, mission-oriented innovation hubs.

Sources:

Tilde AI Launches TildeOpen LLM for European Languages

Latvian AI company Tilde AI has launched TildeOpen LLM, an open-source large language model designed specifically for European languages, particularly those underrepresented in major AI models. The 30-billion-parameter model was released on Hugging Face under a permissive CC-BY-4.0 license. Developed with support from the European Commission’s Large AI Grand Challenge and leveraging the LUMI and JUPITER supercomputers, this initiative aims to promote linguistic equity and digital sovereignty in the EU. By providing a powerful tool that can be hosted locally, it ensures data privacy and compliance with regulations like GDPR.

Sources:

Microsoft Open-Sources Historic 6502 BASIC Code on GitHub

Microsoft has released the source code for its historic 6502 BASIC, a foundational piece of software from the dawn of personal computing. The code, dating back to 1976, was instrumental for early computers like the Commodore 64 and Apple II. It is now available on GitHub under the MIT license. This release of version 1.1 includes garbage collector fixes implemented by Commodore engineer John Feagans and Bill Gates in 1978. The 6,955 lines of assembly code offer a valuable resource for retro-computing enthusiasts, historians, and developers, preserving a key part of technological heritage.

Sources:

Linux 6.17 to Ship Major Stability Fixes for NVIDIA’s Nouveau Driver

The upcoming Linux 6.17 kernel will merge critical stability fixes for the Nouveau open-source driver for NVIDIA graphics cards. These patches resolve long-standing bugs that caused applications and games to freeze or crash, often resulting in a lost desktop session. The fixes, which address issues with GPU fault handling and fencing, are also planned for back-porting to earlier stable kernel releases. This development is expected to make the Nouveau driver a significantly more reliable option for daily use and gaming on Linux systems with NVIDIA GPUs.

Sources: