Europe Pushes AI Sovereignty as Databricks Ships Genie One

AI & Machine Learning

Databricks this week launched “Genie One,” an agentic coworker product positioned to automate and orchestrate workflows for business teams by grounding agents in enterprise data inside the Databricks Lakehouse. The announcement emphasizes governed agents, built‑in connectors to common enterprise systems, and controls that Databricks says will help organizations run long‑running, business‑critical agent tasks with auditability. Given Databricks’ role in enterprise ML stacks, Genie One signals a move to productize agent architectures for non‑engineering teams and compete on data governance and integration rather than raw model quality alone. Observers will watch adoption in pilot customers and how Databricks balances automation capabilities with governance and compliance requirements. Source: Databricks Verified: True

Consumer Hardware

New usage data shows ChatGPT’s share of the global AI assistant market has fallen below 50% for the first time as competitors like Google’s Gemini and Anthropic’s Claude gain user traction. The report attributes the shift to rising multi‑assistant usage, newer feature sets from rivals, and intensifying product competition that could pressure ChatGPT’s monetization and retention strategies. For consumers, a more fragmented assistant market may drive faster feature innovation but also raises questions about cross‑assistant data portability and experience consistency. The market movement also underscores how quickly leadership in consumer AI can change as major platform players expand distribution and integration. Source: TechCrunch Verified: True

Google officially launched Android 17 this week alongside Pixel Drop updates and Wear OS enhancements, delivering developer‑facing changes, privacy tweaks, and improved on‑device model support that target multimodal and local AI scenarios. Initial rollouts are hitting Google Pixel devices first, with carriers and OEMs expected to follow in staged timelines; Google highlighted battery and UX optimizations as part of the release notes. The update signals Google doubling down on enabling on‑device AI experiences and giving developers more tools to leverage local models, which could shift some workloads off cloud inference. Android 17’s adoption pace will be a key metric for how quickly these on‑device AI capabilities propagate across the Android ecosystem. Source: Mashable Verified: True

Genesis AI, a French startup, unveiled its humanoid concept “Eno,” emphasizing capability‑first design for mobility and manipulation rather than human likeness in appearance. The company presented Eno as a general‑purpose manipulator with teleoperation and autonomy features aimed at service and logistics scenarios, and framed safety and operator control as core engineering priorities. Genesis AI’s announcement places Eno among a crowded set of humanoid and general‑purpose robots that are targeting industrial and commercial tasks, and it highlights continued investor interest in hardware startups attempting to bridge research and product readiness. The coming months will reveal how prototype demonstrations match up to real‑world robustness and whether customers see Eno as a practical alternative to more task‑specific robotic solutions. Source: The Verge Verified: True

Cybersecurity

Microsoft acknowledged an authenticated elevation‑of‑privilege zero‑day in Microsoft Defender tracked as CVE‑2026‑50656 and dubbed “RoguePlanet” after proof‑of‑concept exploit code began circulating publicly. Security researchers reported the flaw can be chained to escalate to SYSTEM on affected Windows installations, prompting Microsoft to prepare a patch and urging administrators to apply mitigations immediately. The public availability of exploit code has raised incident‑response urgency, with defenders advised to monitor telemetry, apply vendor workarounds where available, and prioritize vulnerable systems for rapid remediation. The episode underscores continued risks from widely used endpoint products and the operational strain produced when exploit code spreads ahead of patches. Source: Help Net Security Verified: True

Threat intelligence firms reported active exploitation campaigns chaining multiple Fortinet FortiSandbox vulnerabilities to gain footholds and execute post‑exploit activity against exposed appliances. Vendors and national CERTs have published indicators of compromise and mitigation guidance while urging urgent patching, network segmentation of management interfaces, and enhanced monitoring for affected customers. The attacks highlight how attackers are rapidly weaponizing newly disclosed vulnerabilities in widely deployed security appliances to pivot into enterprise networks. Organizations using FortiSandbox are recommended to prioritize updates, audit access controls, and apply vendor recommended hardening to reduce attack surface and lateral movement potential. Source: The Hacker News Verified: True

Enterprise Infrastructure

Nebius completed its acquisition of Eigen AI to bolster its inference and managed model hosting capabilities, saying the deal will accelerate the provider’s inference stack optimizations and tooling for enterprise customers. The acquisition reflects ongoing consolidation as cloud and AI‑infrastructure providers seek specialized engineering teams to improve latency, cost, and deployment workflows for production models. For enterprise customers, tighter integration between Nebius’ platform and Eigen’s inference optimizations could simplify managed hosting and reduce operational overhead for scaling inference workloads. The move also signals competitive pressure on other AI cloud vendors to shore up inference performance through acquisitions or in‑house engineering. Source: Nebius Verified: True

Arcade announced a $60 million financing round to expand its authorization, governance, and policy controls for agentic AI deployments in the enterprise, positioning itself as a control plane for long‑running agents. The company emphasized features like RBAC, audit trails, policy enforcement, and integrations designed to meet Fortune‑500 security and compliance requirements as agents move into production. The raise signals investor confidence in infrastructure startups that provide governance primitives rather than models themselves, reflecting demand from enterprises worried about control, safety, and accountability for autonomous agent behavior. Wider adoption of such control layers could become a prerequisite for large organizations deploying agentic systems at scale. Source: HPCwire Verified: True

SpaceX launched three Block‑2 BlueBird satellites for AST SpaceMobile in a mission designed to expand space‑based cellular coverage and operator partnerships, aiming to extend mobile connectivity into remote regions. AST says the Block‑2 satellites will increase capacity for direct‑to‑cell services while the launch underscores ongoing commercial experimentation with space‑based telephony and the regulatory complexities that accompany it. The mission carries implications for remote connectivity strategies, potential edge compute offload, and how hyperscalers and telcos might integrate satellite capacity into broader network offerings. Observers will track operator certification, spectrum coordination, and commercial agreements that determine whether space‑cell architectures scale profitably. Source: Spaceflight Now Verified: True

Policy & Regulation

The UAE’s media office announced a minimum social‑media age of 15 and said platforms must implement age‑verification measures and additional child safety protections, presenting the rule as the first of its kind in the Arab world. The policy forces platforms to consider stronger identity and age controls and aligns with a global trend of national measures restricting younger teens’ access to social platforms. Practically, the rule will accelerate deployment of age‑checks and may raise questions about privacy, verification methods, and enforcement across international platforms operating in the UAE. The policy also serves as a regional test case for balancing child safety objectives with technical feasibility and civil‑liberties concerns. Source: Reuters Verified: True

European policymakers and tech leaders meeting around the G7 and VivaTech gatherings warned about U.S. dominance in AI and pressed for stronger European cloud and AI sovereignty, including certification, procurement levers, and in‑region data governance. The reporting frames this moment as a test of Europe’s ability to translate concern into industrial strategy and regulatory measures that keep strategic workloads and data within European borders. Discussions highlighted political appetite for state support, standards development, and procurement tools to favor sovereign providers, but also noted the practical hurdles of matching scale and talent of U.S. hyperscalers. The outcome of these debates will shape whether Europe can carve a credible alternative path for strategic AI deployments or will continue to rely on non‑EU cloud providers. Source: Reuters Verified: True