IT News Roundup: AI Infrastructure Race, Critical Vulnerabilities, and Open Source Security - June 26, 2026

The technology landscape this week reflects an industry at a crossroads: artificial intelligence infrastructure is scaling at unprecedented speed while the security challenges it creates grow in parallel. From custom AI chips to record-breaking vulnerability disclosures, from ransomware campaigns targeting household names to new initiatives aimed at hardening open-source software, the stories below paint a picture of an ecosystem racing to build faster than it can secure.

AI Infrastructure Enters the Hardware Wars

The race to build AI infrastructure has moved beyond model development into full-stack silicon design. OpenAI revealed its first custom inference chip, signaling that major AI labs are no longer content to rely on third-party GPU suppliers for their compute needs. At the same time, Micron delivered next-generation HBM4 memory modules designed specifically for AI training and inference workloads.

Alphabet announced an 0 billion funding plan dedicated to extending its AI compute buildout, while Anthropic moved closer to a public listing after submitting a confidential S-1 filing. The infrastructure arms race is no longer about who has the best model — it is about who controls the entire stack from silicon to deployment.

Source: Tech Startups

Microsoft Sets Record with 208 Vulnerabilities in June Patch Tuesday

Microsoft released what may be its largest-ever Patch Tuesday update, addressing 208 software vulnerabilities across its product portfolio. The sheer volume of fixes underscores the expanding attack surface of modern enterprise environments and places significant pressure on IT teams to prioritize and deploy patches within tight windows.

Among the most critical were flaws in Windows Netlogon (CVE-2026-41089), a stack-based buffer overflow being actively exploited against domain controllers, and multiple remote code execution vulnerabilities in Exchange Server. Security teams are advised to apply these updates immediately, particularly for systems exposed to untrusted networks.

Source: CRN

Oracle PeopleSoft Vulnerability Actively Exploited by ShinyHunters Ransomware Gang

A critical vulnerability in Oracle PeopleSoft (CVE-2026-35273) was added to the CISA Known Exploited Vulnerabilities catalog after confirmed use in ransomware campaigns. The ShinyHunters threat group has been actively exploiting this flaw, with the CISA remediation deadline now overdue.

The same gang has run multiple parallel extortion tracks throughout June 2026, targeting high-profile victims including Eastman Kodak and Madison Square Garden Sports. Kodak confirmed that an unauthorized party briefly accessed a limited amount of company data after ShinyHunters threatened to publish over 2.2 million records of customer PII and internal corporate files. The group also released a 45 GB archive of Madison Square Garden data after the company reportedly declined to pay.

Sources: SWK Technologies, Threat Modeling

Google Patches Fifth Chrome Zero-Day of 2026 — Actively Exploited in the Wild

Google has patched its fifth actively exploited zero-day vulnerability in Chrome for 2026. The flaw, tracked as CVE-2026-11645, is an out-of-bounds memory access bug in Chrome’s V8 JavaScript engine that earned its discoverer a 5,000 bounty through the Zero-Day Initiative program.

Google confirmed the vulnerability was being exploited in targeted attacks before the patch was released. With five zero-days already this year, browser vendors face mounting pressure to improve their security models as attackers increasingly use compromised browsers as initial access vectors for more sophisticated campaigns.

Source: Cyber Security Review

Major Hack Campaign Compromises Fortinet Firewalls at Prominent Organizations

Security researchers disclosed a significant campaign targeting Fortinet firewall devices that resulted in the compromise of multiple prominent organizations. The attack leveraged vulnerabilities in FortiGate firewalls to establish persistent access, turning network security appliances into entry points for lateral movement.

The campaign was attributed to an initial access broker operation, highlighting how threat actors are increasingly targeting perimeter security infrastructure rather than attempting traditional phishing or social engineering approaches. Organizations running Fortinet equipment should verify their firmware versions and review firewall logs for signs of compromise.

Source: Reuters

OpenAI Launches Daybreak Cybersecurity Initiative with GPT-5.5-Cyber

OpenAI announced Daybreak, a new cybersecurity initiative designed to help organizations move beyond vulnerability discovery into automated patch deployment. The program includes the full release of GPT-5.5-Cyber, an AI model specifically trained for security analysis and remediation workflows.

A key component is “Patch the Planet,” a sub-initiative aimed at supporting open-source maintainers by using AI to identify and help fix vulnerabilities before they reach production. The program already demonstrated results during safety evaluations that caught CVE-2026-8390, which Mozilla patched two days before Pwn2Own Berlin — prompting five of six registered Firefox exploit entries to withdraw.

Sources: OpenAI, Engadget

Ransomware Surges 48% in May — Education Sector Most Targeted

Check Point Research reported that while overall cyberattacks eased slightly in May 2026, ransomware incidents surged by 48%. The education sector was targeted more than any other industry, with schools and universities facing increasingly sophisticated attacks aimed at disrupting operations during peak academic periods.

The rise coincides with a broader trend of AI-assisted cybercrime, where threat actors use machine learning tools to automate reconnaissance, craft convincing phishing campaigns, and adapt their tactics faster than traditional defenses can respond. Organizations in education and other critical sectors are being urged to review backup strategies and incident response plans.

Source: World Economic Forum