Anthropic has launched "Project Glasswing," a new initiative aimed at securing the world's most critical software, in collaboration with major partners including Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks. This project leverages the capabilities of Anthropic's unreleased frontier model, Claude Mythos Preview.

Claude Mythos Preview has demonstrated advanced coding capabilities, identifying thousands of high-severity software vulnerabilities that are challenging for humans to find, including those in major operating systems and web browsers. Anthropic is committing up to $100 million in usage credits for Mythos Preview and $4 million in direct donations to open-source security organizations to promote its defensive application.

The model autonomously discovered and developed exploits for a 27-year-old vulnerability in OpenBSD, a 16-year-old flaw in FFmpeg, and multiple vulnerabilities in the Linux kernel, all of which have since been patched. Project Glasswing aims to give defenders a lasting advantage in the coming era of "AI-driven cybersecurity," recognizing that while AI can escalate cyber threats, it also offers powerful tools for identifying and fixing vulnerabilities.

👉 Read the full article on Anthropic

OpenAI has reportedly sent a letter to the attorneys general of California and Delaware, alleging that Elon Musk and Meta are coordinating "attacks" against the company. The letter, while addressing allegations against OpenAI CEO Sam Altman in a New Yorker story, specifically requests an investigation into Musk's actions.

Chris Lehane, OpenAI's chief of global affairs, told CNBC that Musk and Meta CEO Mark Zuckerberg are engaging in "highly questionable and sharply worthy of investigation" conduct. The letter accuses Musk and Zuckerberg of "colluding as Musk sought to dig up incriminating information on Altman," characterizing this as "improper and anti-competitive behavior" that could hinder OpenAI's ability to develop artificial general intelligence (AGI).

Musk is currently suing OpenAI for $134 billion, a lawsuit OpenAI has described as part of an "ongoing pattern of harassment." Reports from the New Yorker suggest that intermediaries connected to Musk circulated detailed opposition research on Altman, including personal contact information and tracking of his activities. OpenAI claims these "attacks" are designed to transfer control of AGI's future from those committed to benefiting humanity to competitors lacking mission-driven principles and responsibility for safety.

👉 Read the full article on Gizmodo

Anthropic has announced a new agreement with Google and Broadcom to secure multiple gigawatts of next-generation TPU capacity, expected to come online starting in 2027. This significant expansion of compute infrastructure is intended to power Anthropic's frontier Claude models and meet the extraordinary demand from customers worldwide.

Krishna Rao, Anthropic's CFO, stated that this partnership continues their disciplined approach to scaling infrastructure, enabling Claude to define the frontier of AI development while accommodating exponential customer growth. Anthropic's run-rate revenue has surged past $30 billion from approximately $9 billion at the end of 2025, and the number of business customers spending over $1 million annually has doubled to over 1,000 in less than two months.

The majority of this new compute capacity will be located in the United States, further expanding Anthropic's November 2025 commitment to invest $50 billion in strengthening American computing infrastructure. Anthropic trains and runs Claude on diverse AI hardware, including AWS Trainium, Google TPUs, and NVIDIA GPUs, allowing them to optimize workloads for the best-suited chips and provide superior performance and resilience to customers.

👉 Read the full article on Anthropic