Anthropic just dropped a bomb on the cybersecurity world. The AI company has trained a model so good at finding software bugs that it can out-hack all but the most elite human coders. And instead of keeping it secret, it’s giving the keys to a coalition of tech giants—including Amazon, Apple, Google, and Microsoft—in a project it’s calling Project Glasswing.
The model, Claude Mythos 2 Preview, isn’t just another code assistant. Anthropic says it has already uncovered thousands of high-severity vulnerabilities, including critical flaws in every major operating system and every major web browser. The company is blunt about the stakes: these capabilities will soon be weaponized by malicious actors, and the fallout for economies, public safety, and national security could be devastating.
A Coalition of Heavy Hitters
Project Glasswing isn’t a solo mission. Anthropic has assembled a who’s-who of infrastructure and security players—AWS, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks—to put Mythos to work on defense. These partners will use the model to hunt bugs in their own systems, and Anthropic will share everything it learns publicly.
The company is putting serious money behind the effort. It’s committing up to $100 million in usage credits for Mythos Preview, plus $4 million in direct donations to open-source security organizations. More than 40 additional organizations that build or maintain critical software infrastructure have also been granted access to scan and secure both proprietary code and open-source dependencies.
Why This Matters for Space Security
For the space industry, this is a wake-up call. Satellite software, ground station firmware, and orbital communication protocols are notoriously hard to patch once deployed. A model that can autonomously find zero-day exploits in operating systems and browsers can do the same in space-grade code—where a single vulnerability could mean lost payloads, hijacked satellites, or compromised national security.
The same AI arms race playing out in terrestrial cybersecurity is coming to orbit. Project Glasswing is an urgent attempt to get ahead of it, but the timeline is tight. As Anthropic warns, the proliferation of these capabilities is inevitable. The question isn’t whether they’ll be used, but who gets there first—and whether defense can outpace offense before the next major incident.
—
*Originally reported by [SatelliteToday Cyber](https://www.satellitetoday.com/cybersecurity/2026/04/12/anthropic-launches-project-glasswing-for-cybersecurity/). Adapted and republished with editorial context for SpaceSecurityNews.*
