The United States is preparing to crack down on alleged industrial-scale theft of artificial intelligence intellectual property by Chinese entities. Specific accusations concern the “distillation” technique for Large Language Models, with OpenAI, Google, and Anthropic reporting attempts to clone their models. The dispute highlights growing geopolitical tensions and challenges for data sovereignty and AI asset security.
Opera has expanded deployment options for its Opera GX browser on Linux, now making it available via Flatpak and Snap containerized application formats. This move, following the initial launch with RPM and Debian packages, offers greater flexibility to users and introduces benefits in terms of security and dependency management, crucial aspects for infrastructure architects and DevOps teams.
The White House's OSTP has formally accused China of "industrial-scale" distillation of US AI models. In response, it committed to sharing intelligence with leading US AI companies like OpenAI, Anthropic, and Google, while also exploring accountability measures. Previous accusations from OpenAI against DeepSeek and Anthropic against DeepSeek, MiniMax, and Moonshot AI for fraudulent activities underscore these concerns.
Virginia-based startup Rilian has announced $17.5 million in funding to develop agentic AI systems. Its Caspian platform is designed to operate as a command layer over existing security stacks, deploying pre-trained AI agents into air-gapped and compliance-restricted environments, targeting the defense and national security sector.
Noscroll introduces an AI-powered bot designed to automate the reading and filtering of online content, aiming to combat the phenomenon of "doomscrolling." This solution seeks to offer users a more curated information experience by delegating the task of scanning the web and selecting relevant information to AI.
Configuring Codex is crucial for optimizing task execution and personalizing workflows. Through settings like personalization, detail level, and permissions, users can fine-tune system behavior. This approach is fundamental for ensuring efficiency and control, aspects particularly relevant in on-premise AI deployment environments, where precise resource and data management is a priority.
Claim Clarity advocates for the application of specialized artificial intelligence to improve decision precision in the workers' compensation sector. This domain, while part of the broader healthcare ecosystem, features unique regulatory and operational dynamics, necessitating a targeted rather than generic approach to address its complexities and optimize efficiency.
Kostiantyn Gitko, leader of Devox, built his career in complex enterprise IT environments where daily reliability was a core requirement. This foundational experience, progressing from software engineer to leadership roles in infrastructure and operations management, profoundly influences his approach to building new tech businesses. His vision emphasizes structure, scale, and stability, principles critical for modern LLM deployments.
Canonical has released Ubuntu 26.04 LTS, along with Ubuntu Server 26.04 LTS and its various flavors, powered by the Linux 7.0 kernel. This Long Term Support version represents a significant update for enterprises managing AI workloads, offering a stable and reliable platform for on-premise deployments. Its availability is crucial for CTOs and infrastructure architects seeking control, data sovereignty, and TCO optimization.
Microsoft is expanding the capabilities of its Copilot assistant, which will now be able to not only suggest edits but also apply them directly in Word documents. This evolution, introducing an "agentic Copilot" in Excel and PowerPoint, raises questions for companies evaluating data control and the autonomy of AI solutions, especially in contexts where sovereignty and on-premise management are priorities.
The Pentagon's ANPI program has selected three companies for the deployment of nuclear microreactors at Buckley Space Force Base and Malmstrom Air Force Base by 2030. The initiative, which began with eight vendors, aims to strengthen the energy resilience and independence of critical military installations, leveraging commercially owned nuclear technology.
Datadog has expanded its observability platform capabilities by integrating GPU monitoring. This new feature offers organizations managing AI workloads essential tools to analyze the efficiency and utilization of their most expensive hardware, thereby addressing the rising operational costs associated with artificial intelligence.
A sophisticated cryptocurrency scam leveraged the Strait of Hormuz crisis, with fake 'Iranian authorities' extorting payments from oil tankers awaiting cargo. Despite making payments, two vessels were reportedly fired upon. The incident highlights the increasing complexity of digital threats and the need for robust security strategies, a crucial aspect also for Large Language Models (LLM) deployments in on-premise environments.
Eleven data center campuses in the US, powered by new natural gas projects, could generate more greenhouse gases than Morocco's total emissions in 2024. These facilities, serving AI giants like OpenAI and Microsoft, bypass traditional grids for energy autonomy, highlighting the sector's growing environmental impact and the infrastructure challenges for on-premise deployments.
AI agent training startup Context AI has disclosed a security incident. TechCrunch confirmed that Delve, a compliance company already facing scrutiny, had performed Context AI's security certifications. The incident raises questions about the robustness of security audits and the implications for data sovereignty and security within the industry, highlighting the need for rigorous due diligence.
Canonical and AMD collaborated to integrate ROCm libraries into the Ubuntu 26.04 LTS archive, promising simplified installation for AI/ML and HPC workloads. However, at the time of release, the versions available via `sudo apt install rocm` are several months out-of-date. This situation presents a trade-off for on-premise deployments, balancing ease of setup with the need for up-to-date software to maximize performance and compatibility.
Corsair's configurator for its 4000 Series PC cases offers extensive customization options, including distinctive materials. While aimed at the consumer market, this approach to hardware modularity reflects a growing need in the enterprise sector. For those managing on-premise AI infrastructures, the ability to configure and adapt hardware is crucial for optimizing TCO, performance, and data sovereignty, highlighting the importance of flexibility in system design.
NVIDIA and Google Cloud have unveiled a joint hardware and software roadmap to optimize large-scale AI Inference. The new A5X bare-metal instances, powered by NVIDIA Vera Rubin NVL72 systems, promise significant cost reductions and increased Throughput. The initiative also addresses data sovereignty challenges with solutions like Google Distributed Cloud and NVIDIA Confidential Computing, extending AI capabilities to regulated sectors and complex industrial applications.
A Sony project has achieved a significant breakthrough, demonstrating artificial intelligence's ability to surpass human performance in tasks requiring extreme speed and accuracy, such as ping-pong. This success is part of a broader trend of AI advancement, which has already seen machines prevail in complex games like chess and Go, highlighting the growing potential for enterprise automation.
Berlin-based climatetech startup VREY has closed a €3.3 million seed funding round. The company offers a platform enabling property owners to deploy solar systems in multi-family buildings and supply energy directly to tenants, bypassing regulatory complexities. Leveraging a new German framework, VREY aims to unlock the potential of shared-roof solar, also supporting batteries and EV charging infrastructure.