Project Glasswing tackles the growing prowess of frontier AI models while bank charter activity shot up in April and a major rewrite of the OpenAI-Microsoft pact has ended what was effectively OpenAI’s cloud exclusivity.

Anthropic’s Project Glasswing Marks the Start of an AI Cybersecurity Arms Race

The companies behind the world’s most important digital infrastructure are preparing for a future where cyberattacks happen at machine speed. Anthropic unveiled Project Glasswing, a major initiative bringing together Amazon, Apple, Google, Microsoft, Cisco, CrowdStrike, Palo Alto Networks, JPMorganChase, and the Linux Foundation around a shared concern: frontier AI models are becoming exceptionally good at finding and exploiting software vulnerabilities, and the industry may have very little time to prepare for what happens once those capabilities spread more broadly. The project focuses on Claude Mythos Preview, an unreleased AI model that Anthropic describes as sufficiently advanced to identify and exploit vulnerabilities at a level comparable to (and in some cases exceeding) that of top human specialists. The warning from Anthropic feels notably direct. The examples released by the company explain why the industry is taking the warning seriously.

Bank Charters Return to the Fintech Playbook

Charter activity ramped up in late April, with a new application from Mission Lane, conditional approval being granted for Mercury, and a Reuters update on Revolut’s licensing push. Together, these moves suggest fintechs are starting to reevaluate whether partner-bank models are truly enough as lending economics, product control, and valuation resilience are becoming increasingly dependent on regulatory status. Mission Lane’s filing in particular wasn’t just another routine paperwork story. It was the first credit-card bank charter application in roughly 20 years. When a credit-card fintech seeks a Competitive Equality Banking Act (CEBA) charter, it is effectively arguing that direct ownership of a charter is now important enough to justify the capital, compliance, and supervisory costs involved—which can easily run into the tens of millions.

OpenAI’s New Cloud Deal with Microsoft Changes Everything for Amazon, Google, and AI Power

For years, access to OpenAI at enterprise scale came with an unspoken condition that sooner or later, an organization would go through Microsoft. Azure was not simply a preferred cloud provider; in fact, it was the commercial gateway to OpenAI’s most valuable models, giving Microsoft extraordinary leverage in the race to dominate enterprise AI infrastructure. If businesses wanted serious deployment of GPT models, Microsoft sat firmly in the middle of the relationship. But right now, a major rewrite of the OpenAI-Microsoft pact has ended what was effectively OpenAI’s cloud exclusivity, allowing the company to distribute its models across multiple cloud platforms. It sounds like a contractual adjustment, but in practice, it reshapes one of the most important power structures in AI.

Foreign Hackers Are Coming for America’s Infrastructure and They’re Already Inside

Cyberattacks targeting critical infrastructure have surged in recent weeks as tensions in the Middle East continue to escalate, according to a joint advisory from U.S. agencies. When most executives think about cyber threats, they picture phishing emails, stolen passwords, or ransomware locking company files but targets are no longer limited to corporate data or financial systems. Attackers are increasingly going after the operational technology behind essential services: the systems that keep electricity flowing, water running, and local governments functioning. The focus is on operational technology (OT), the industrial systems that monitor and control physical infrastructure. These include devices that regulate water treatment, manage electricity distribution, oversee transportation systems, and support emergency services. Many of these systems were connected to the internet for remote monitoring and convenience, enabling operators to manage infrastructure more quickly and efficiently.

Cross-border Payments Get a Fresh Vote of Confidence from Santander

Spanish banking giant Santander, alongside investment firm Centerbridge Partners and existing shareholders, announced it would back a £550 million funding package for London-based fintech Ebury. The size and mix of investors are noteworthy, because it suggests that there is still room for serious capital in the payments and transactions space, while many fintechs are increasingly focused on diversifying their product portfolios. Santander isn’t simply partnering with Ebury at arm’s length. It’s increasing its stake, having described the company as a strategic cross-border payments platform and source of payments innovation for SMEs.