Meta Debuts $299 AI Glasses; Wearables Go Always‑On

AI & Machine Learning

Seltz raised $12.5M in seed funding to rebuild web search specifically for autonomous AI agents, backing a product focused on search primitives and indexing optimized for agentic retrieval and long‑horizon workflows; investors said the raise reflects continuing appetite for infrastructure that supports agent-native applications. The startup’s approach aims to reduce latency and improve relevance for chains of queries and memory retrievals that multi-step agents require, positioning search as a distinct layer in agent stacks rather than a repackaged consumer search product. This could accelerate development of specialized retrieval pipelines and prompt engineering patterns used by tool-using agents, and it signals renewed VC interest in foundational web indexing for AI. Expect competitors and cloud providers to watch for integrations that fold agent-optimized search into larger AI platforms. Source: Fortune Verified: True

Consumer Hardware

Meta launched a new lower-cost line of smart glasses in partnership with EssilorLuxottica starting at $299 and positioned the devices as an “all-day AI assistant,” pushing always-on wearable AI into mainstream consumer hardware. The glasses combine on-device sensors, lightweight form factors and assistant integrations that aim to surface context-aware assistance without a phone, while Meta emphasized fashion-forward styles to broaden appeal. By lowering the entry price and leaning on EssilorLuxottica’s eyewear ecosystem, Meta hopes to accelerate adoption and developer interest in wearable-first experiences. The move also raises privacy and battery-life questions as always-on sensors and local AI inference become standard in consumer wearables. Source: Meta Blog Verified: True

Cybersecurity

A new local‑privilege‑escalation vulnerability dubbed “DirtyClone” (CVE‑2026‑43503) was disclosed, with JFrog researchers showing an exploit path in the XFRM/IPsec packet‑cloning code that can allow unprivileged users to overwrite page‑cache memory and modify in‑memory binaries without leaving normal audit traces. The flaw affects Linux kernel packet-cloning mechanics and could be weaponized to persist by patching running suids in memory, making detection and forensic analysis harder for defenders. A mainline fix landed in the kernel in late May and distributions are now issuing backports and advisories, but the window between patch and widespread deployment creates an active risk for exposed systems. Organizations running IPsec/XFRM or multitenant Linux hosts should prioritize kernel updates and monitor for anomalous memory tampering. Source: Cybersecurity News Verified: True

IBM, Red Hat and Palo Alto Networks expanded Project Lightwell to accelerate coordinated detection and response to software vulnerabilities across open‑source and packaged components used in enterprise environments. The collaboration extends cross‑vendor scanning, remediation playbooks and protections designed to close the gap between discovery and patching, aiming to reduce time-to-fix for complex supply-chain and dependency issues. By standardizing playbooks and sharing telemetry across vendors, Lightwell intends to streamline automated remediation and decrease the burden on security teams dealing with mixed stacks. The initiative highlights growing industry interest in cooperative, platform-agnostic approaches to reduce systemic risk from widespread vulnerabilities. Source: IBM Newsroom Verified: True

Enterprise Infrastructure

Nebius released AI Cloud 3.6 with updates focused on developer experience, governance, and tighter integrations for regulated and enterprise customers running production model operations. The release adds governance controls for model lifecycle and access, improves the UX for deployment and inference pipelines, and includes features aimed at easier auditing and policy enforcement for compliance-bound workloads. These changes reflect enterprise demand for managed inference platforms that balance velocity with control, particularly for customers in regulated industries. The update positions Nebius to compete more directly with larger cloud vendors by addressing operational gaps enterprises encounter when moving models into production. Source: HPCwire Verified: True

Amazon’s planned hyperscale data‑center campus in Montgomery County, Missouri has drawn scrutiny from residents and local officials over projected water use and environmental impacts, with reporting estimating the campus could use tens of millions of gallons of water per year and prompt infrastructure strain. Community meetings and planned state/local hearings are now expected as stakeholders debate permitting, water rights, and upgrades to municipal systems to accommodate the buildout. The story underscores recurring tensions between hyperscale operators and local communities where water scarcity or aging infrastructure raise costs and political pushback. How regulators and Amazon negotiate mitigation measures will influence future site selections and community relations for large data-center projects. Source: KSDK Verified: True

RunPod announced a $100M raise to expand its AI compute platform, aiming to scale GPU cloud offerings, managed training and inference services, and developer tooling for teams that need elastic, AI‑optimized infrastructure. The financing will help RunPod grow capacity and integrations that target startups and ML engineers seeking lower-cost alternatives to hyperscaler compute or more flexible billing. As demand for GPU-backed training and inference continues, RunPod’s push reflects competition in the AI infrastructure market between specialist clouds and major providers. If executed well, the funding could accelerate time-to-scale for smaller teams and influence pricing dynamics for on-demand GPU usage. Source: Technical.ly Verified: True

Nearfield closed a $380M financing to accelerate development of 3D‑metrology and process‑control tools used in advanced semiconductor production, citing rising demand from AI-accelerator manufacturers for tighter process control at advanced nodes and packaging steps. The capital will fund product expansion and increased production capacity for metrology equipment critical to yield and performance tuning in AI chip fabs. Investors pointed to surging need for precision measurement as chips push smaller geometries and complex multi-die assemblies, making metrology a strategic bottleneck and investment area. This funding round signals growing investor focus on the tooling layer of the semiconductor supply chain that underpins AI hardware scaling. Source: Metrology News Verified: True

Policy & Regulation

No major stories this sector today.