Trilogue negotiations for reforming the EU AI Act under the Digital Omnibus for AI were delayed after the Parliament and Council disagreed on the act's overlap with sectoral regulations, with talks resuming next month ahead of a looming compliance deadline in August. A delay in the trilogue negotiations could mean that high-risk systems under Annex III would still need to comply by the AI Act's original 2 Aug. 2026 compliance deadline, creating regulatory uncertainty.

Meta acquired humanoid robotics startup Assured Robot Intelligence for an undisclosed sum, with the company described as at the frontier of robotic intelligence designed to enable robots to understand, predict, and adapt to human behaviors in complex and dynamic environments. The acquisition integrates ARI's team into Meta's Superintelligence Labs and reflects the industry's conviction that embodied AI training in the physical world is critical for AGI development.

Wetour Robotics is expanding into Physical AI infrastructure, developing Orchestra—a next-generation operating system that coordinates human intent with intelligent physical devices including wearable robotics, with sensory modules including VisionLink (computer vision) and Conductor (sEMG-based neural gesture recognition), both processed in real time on a portable edge AI hub. The launch represents emerging infrastructure for coordinating human-machine interaction at scale.

DeepSeek unveiled the V4 Flash and V4 Pro series, touting top-tier performance in coding benchmarks and big advancements in reasoning and agentic tasks with architecture upgrades and optimization improvements. The model pushes cost-efficiency boundaries while maintaining frontier performance, with Hybrid Attention Architecture improving the ability of an AI platform to remember queries across long conversations.

OpenAI announced GPT-5.5, its latest AI model that is better at coding, using computers and pursuing deeper research capabilities, with the model excelling at analyzing data, writing and debugging code, operating software, researching online and creating documents and spreadsheets. The release demonstrates convergence at the frontier: three models (GPT-5.4, Gemini 3.1 Pro, Claude Opus 4.6) now compete within 4 percentage points on composite benchmarks.

NVIDIA unveiled Nemotron 3 Nano Omni, an open multimodal model bringing vision, audio, and language together into one system, enabling agents to deliver faster, smarter responses with advanced reasoning across video, audio, image and text, providing enterprises and developers a production path for more efficient and accurate multimodal AI agents. The release establishes NVIDIA's leadership in physical AI infrastructure for robotics and edge deployment.