AI Upgrades, Security Flaws, and SpaceX’s Record IPO Define the Week in Tech

AI kept moving closer to the device, the workflow, and the boardroom this week, with OpenAI, Google, Microsoft, Anthropic, Nvidia, and Zoom pushing smarter assistants and on-device tools while security researchers kept finding new ways for those same systems to go sideways. Add in major Android and web server flaws, AI-enabled scams, and SpaceX’s record IPO plans, and the message was clear: tech’s next era is arriving fast, carrying both productivity gains and a suspiciously large number of blinking warning lights.

Top news

AI assistants and on-device intelligence take center stage

OpenAI introduced Dreaming V3, a new memory system for ChatGPT that integrates past chat details into user profiles. The update doubles storage for Plus and Pro users and adds privacy-focused features like Temporary Chats. While the system enhances personalization, it also raises new privacy concerns as ChatGPT surpasses one billion monthly users.

Google DeepMind launched Gemma 4 12B, a 12-billion-parameter multimodal AI model that runs entirely on devices with 16 GB of RAM. Capable of processing text, images, audio, and video offline, Gemma 4 12B supports a 256K context window and reduces reliance on cloud GPUs, offering privacy and cost benefits for developers.

Microsoft unveiled Scout, an always-on AI assistant embedded across Teams, Outlook, and Windows. Built on the OpenClaw framework, Scout automates scheduling, communication, and task management as part of Microsoft’s shift toward agent-first workflows and a consumption-based AI model.

Furthering its AI ecosystem, Microsoft also revealed Project Solara, an Android-based OS designed for AI agent-driven devices. The platform supports adaptive interfaces and enterprise management tools, with pilot programs already underway in retail and healthcare sectors.

AI hardware and cloud partnerships expand

Nvidia and Microsoft jointly introduced the RTX Spark, an Arm-based superchip combining a 20-core Grace CPU with a Blackwell RTX GPU and up to 128 GB of unified memory. The chip powers the new Surface Laptop Ultra and promises one petaflop of on-device AI performance, with Spark-powered PCs expected this fall.

Pinterest signed a $4 billion cloud and AI deal with Amazon Web Services through 2031. The agreement leverages AWS Trainium and Graviton chips and includes a migration to Kubernetes for improved scalability and efficiency, boosting both companies’ shares.

Insider intel

AI governance and security in focus

At Build 2026, Microsoft introduced Microsoft Execution Containers (MXC), an OS-level sandbox that restricts AI agents’ access to files, networks, and UI elements. The feature enhances enterprise trust by enforcing real-time containment and governance through Microsoft’s security suite.

Anthropic launched Claude Opus 4.8, an AI model emphasizing honesty and transparency. The update introduces dynamic workflows supporting up to 1,000 subagents, a faster and cheaper mode, and improved code flaw detection. The company’s valuation now exceeds OpenAI’s as it pursues more trustworthy enterprise AI.

AI-free search gains momentum

DuckDuckGo saw a 70% surge in US iPhone installs and record search traffic following backlash to Google’s AI-powered search results. The privacy-focused engine also launched Chrome and Firefox extensions to preserve AI-free search, while competitors like Brave and Bing experienced parallel growth from AI-wary users.

AI productivity tools expand

Zoom introduced ZoomMate, an AI workspace that transforms meeting discussions into actionable tasks and documents across platforms like Salesforce, Jira, and Slack. The new AI Productivity Suite allows users to generate documents and slides directly within Zoom, integrating seamlessly with competing collaboration tools.

Security alerts

Critical vulnerabilities and exploits

A zero-day flaw in GitHub’s browser-based VS Code allowed attackers to steal OAuth tokens with a single click, exposing private repositories. Researcher Ammar Askar disclosed the issue, prompting Microsoft to issue mitigations and advise users to clear cookies and review access tokens.

Researchers uncovered a new HTTP/2 “Bomb” exploit capable of crashing major web servers — including NGINX, Apache, IIS, Envoy, and Cloudflare Pingora — by abusing HPACK compression and flow control. Patches are available for NGINX, Apache, and Envoy, while IIS and Pingora remain vulnerable.

Google released its June 2026 Android update, patching an actively exploited privilege escalation flaw and over 120 other vulnerabilities, including a remote zero-click exploit. Pixel devices received the patch first, with other vendors to follow.

AI and platform security risks

A vulnerability dubbed Fake Context Alignment allowed crafted notifications from apps like WhatsApp and Slack to manipulate Google Gemini on Android devices. Google patched the issue server-side last November and recommends revoking Gemini’s notification access for added safety.

ChatGPhish, a new prompt injection exploit, embeds malicious instructions into web pages. When ChatGPT summarizes such pages, it can display phishing links, fake alerts, and tracking pixels. OpenAI has been notified but has not yet issued a fix.

A flaw in Meta’s AI support bot allowed attackers to reset Instagram passwords and hijack accounts, bypassing two-factor authentication. The exploit affected several high-profile users before Meta deployed a fix.

Browser and device protections

Google Chrome introduced Device Bound Session Credentials (DBSC), which tie session cookies to a device’s hardware security module. The feature prevents cookie theft attacks and is enabled by default on Windows and macOS, with admin enforcement options available.

Android is adding deepfake call detection to combat AI voice scams. The feature, rolling out first to Pixel devices, uses RCS handshakes to verify callers and warn users of potential spoofing or AI-generated voices.

Data breaches and fraud

Charter Communications confirmed a breach by the ShinyHunters group that exposed data from up to 13 million Spectrum customer accounts. The attack, triggered by a voice-phishing scam, compromised both customer and employee information.

The FBI reported nearly $900 million in AI-related scam losses last year, with voice-cloning cons leading the surge. Criminals used short audio clips to impersonate victims’ relatives or colleagues, prompting fraudulent money transfers. Experts urge users to verify calls and limit public audio sharing.

Industry shakeups

Major market moves and IPOs

SpaceX announced plans for a record-breaking $75 billion IPO, setting a $135 share price for its June 12 Nasdaq debut. The offering targets a $1.77 trillion valuation — surpassing Saudi Aramco’s record — and grants Elon Musk over 82% voting power, fueling speculation about a potential Tesla merger.

Policy and regulation

President Trump signed an executive order establishing a voluntary AI vetting process, allowing the NSA to review frontier AI models for 30 days before launch. The move aims to identify cyber risks but has drawn mixed reactions from industry leaders and safety advocates. OpenAI has called for mandatory federal oversight instead.

Corporate acquisitions and product shifts

Asana acquired no-code agent builder StackAI for $75 million to strengthen its AI-driven workflow automation. The acquisition enhances Asana’s execution tools and positions it more competitively against monday.com and ServiceNow.

Meta is developing an AI pendant and new smart glasses as part of a plan to sell 10 million wearable devices by late 2026. The pendant records conversations for Meta AI processing, raising privacy concerns, while Apple’s competing AI glasses are reportedly delayed until 2027.

A leaked roadmap revealed that Disney plans to phase out the Hulu app by year’s end as part of Project Gemini, consolidating streaming operations under Disney+ to streamline content delivery and improve ad targeting.

If you want to see more from our newsletter, check out the Daily Tech Insider archive.