AI kept its foot on the gas this week, and the rest of tech mostly tried not to get run over. From Microsoft leaning on AWS to keep GitHub afloat and OpenAI bulking up Codex to AI export controls, record AI-linked layoffs, mega M&A, and critical security threats, this week showed an industry racing to automate everything while still needing humans to patch the mess.
Top news
AI and cloud power shifts
Microsoft has turned to Amazon Web Services to support GitHub as AI-driven coding demand overwhelms Azure’s capacity. The move follows repeated outages linked to rapid growth from tools like Copilot, and comes as Microsoft faces a shareholder lawsuit over Azure and Copilot adoption claims.
Databricks launched Genie One, an AI “agentic coworker” platform that integrates with company data and apps like Slack and Teams to automate tasks and generate reports. The system’s real-time ontology layer provides context, signaling a shift from conversational chatbots to task-oriented enterprise AI.
OpenAI announced plans to acquire German startup Ona, integrating its secure cloud sandbox technology into Codex. The acquisition aims to let enterprises run coding agents safely within their own environments, bolstering trust in OpenAI’s enterprise tools following its confidential IPO filing.
AI regulation and export controls
The US Commerce Department has ordered Anthropic to disable its Fable 5 and Mythos 5 AI models for foreign users after Amazon flagged a potential jailbreak. This unprecedented move marks the first time export controls have targeted AI models instead of hardware, signaling a new era of AI governance.
Hardware and AR innovation
Snap unveiled its $2,195 Specs augmented reality glasses at the AWE 2026 conference. The new device supports web browsing, gaming, and EyeConnect for shared AR experiences. Despite its standalone design, the high price and limited battery life have raised doubts about mainstream adoption.
Alibaba launched the Qwen Robot Suite, a trio of AI models — RobotNav, RobotManip, and RobotWorld — designed to give robots human-like perception, planning, and motion. The suite enables physical foresight and adaptability for industrial robotics, marking a major step toward autonomous manufacturing.
Social media and AI integration
Meta introduced “AI Mode” in Facebook search, enabling its Muse Spark AI to generate conversational answers from publicly shared posts across Groups, Reels, and Marketplace. Experts warn of misinformation and privacy risks, as users cannot opt out or verify how their data is reused.
Streaming and media consolidation
Fox announced plans to acquire Roku in a $22 billion cash-and-stock deal valued at $160 per share. The merger will combine Fox’s live content and Tubi streaming service with Roku’s platform, giving Fox control of 73% of the new company and access to vast viewer data for ad targeting.
Insider intel
AI-driven layoffs hit a record for the third consecutive month, with 38,579 of the 97,006 US job cuts in May attributed to AI. Analysts report that 99% of CEOs expect further AI-related reductions within two years, though critics argue that many firms are using AI as a convenient justification for routine downsizing.
Security alerts
Critical vulnerabilities and exploits
Fortinet is facing active exploitation of multiple critical vulnerabilities in its FortiSandbox platform, including CVE-2026-25089 and CVE-2026-39813. These flaws allow remote code execution and authentication bypass, prompting urgent patching and log reviews.
Splunk patched CVE-2026-20253, a critical pre-auth remote code execution vulnerability affecting versions 10.0.0–10.0.6 and 10.2.0–10.2.3. Admins are urged to update immediately or disable the PostgreSQL sidecar service as a temporary workaround. Splunk Cloud remains unaffected, but AWS-hosted Enterprise servers are still exposed by default.
GreatXML, a new proof-of-concept exploit, allows attackers with admin rights to access BitLocker-encrypted drives via Windows Recovery. No patch or CVE has been issued yet, and administrators are advised to monitor for suspicious recovery partition activity.
Cybercrime and enforcement
The FBI and Google dismantled Outsider Enterprise, a China-based phishing-as-a-service network that used AI tools like Gemini to send 2.5 million scam texts. Authorities seized servers, crypto wallets, and a Telegram bot, while Google filed a RICO lawsuit to dismantle remaining infrastructure.
The hacker group ShinyHunters claimed responsibility for stealing 297 GB of HR and payroll data from the Council of Europe, including personal records for over 10,000 staff. The breach may be tied to an Oracle PeopleSoft zero-day vulnerability, prompting urgent patching and MFA guidance.
Policy and workforce challenges
CISA has shortened the federal patch window for high-risk vulnerabilities from 15 days to just 72 hours. The directive responds to the accelerating pace of exploit development driven by AI tools and mandates immediate remediation for critical flaws in platforms like Ivanti Sentry and Oracle PeopleSoft.
A joint ISSA and Omdia survey found that 68% of cybersecurity professionals say their jobs are getting harder, with many excluded from key decisions. Rising burnout and declining CISO roles highlight the urgent need for better workforce support and inclusion in security planning.
Industry shakeups
AI and space industry expansion
SpaceX announced a $60 billion all-stock acquisition of AI coding startup Cursor (Anysphere), gaining access to its developer community and data to enhance xAI models like Grok. The deal positions SpaceX as a major AI player while raising questions about integration and developer trust.
SpaceX’s IPO debuted with a 19% surge, pushing its valuation past $2 trillion and making Elon Musk the world’s first trillionaire. Index providers are fast-tracking the stock into major benchmarks, ensuring widespread inclusion in institutional and retirement portfolios.
Corporate finance and M&A
Nvidia raised $25 billion in its first bond sale since 2021, reflecting strong investor demand. The funds will support long-term flexibility as Nvidia continues record-breaking growth and invests in AI ecosystem ventures including OpenAI and Anthropic.
Salesforce will acquire AI customer service startup Fin (formerly Intercom) for $3.6 billion to enhance its Agentforce platform. The deal, expected to close in Q4 FY27, adds Fin’s AI agent technology to Salesforce’s portfolio as it competes with Microsoft and Oracle in enterprise automation.
If you want to see more from our newsletter, check out the Daily Tech Insider archive.