Google is intensifying its push into the cybersecurity sector at its I/O conference, signaling direct competition with Anthropic's recent breakthrough in AI-driven security. The company invited select expert groups to test CodeMender, an AI agent designed for code security that was first unveiled in October. While initially internal, Google is now making the tool available to external partners to flag and fix vulnerabilities, aiming to secure global codebases. This strategic move follows a wave of industry reactions sparked by Anthropic's surprise announcement of its Claude Mythos Preview. The release of Mythos, a powerful model capable of identifying unknown security gaps in high-stakes systems, shocked the tech world and financial regulators alike. Unlike previous tools, Mythos is reportedly too powerful for public release, yet it has helped Anthropic regain favor with the US government following past supply chain risk issues. The model's potential to generate significant revenue from enterprise and government clients has pressured other AI labs, including OpenAI and Google, to accelerate their own cybersecurity offerings to maintain market leadership and prepare for potential future monetization or IPOs. OpenAI was among the first to respond with its own security-focused tools, and Google has now followed suit. Koray Kavukcuoglu, CTO of Google DeepMind, confirmed in an interview that the company is already in discussions with government agencies and enterprises to use CodeMender for auditing critical systems. The tool represents a shift toward leveraging large-scale AI models not just for creation, but for defensive purposes. During a press briefing on Monday, Google CEO Sundar Pichai acknowledged Anthropic's influence on the sector. He stated that Mythos demonstrated the clear value of large models in security use cases. Pichai expressed confidence in Google's capabilities, noting that the company is equally positioned to deliver similar results. By expanding CodeMender's reach, Google aims to capitalize on the growing demand for AI-enhanced cybersecurity as companies face increasing threats to their digital infrastructure. The competition highlights a broader trend where AI labs view cybersecurity as a key revenue driver. As the race for AI dominance intensifies, these specialized tools offer a practical application that combines immediate utility with substantial financial potential. Google's decision to invite external testers suggests a strategy to refine the tool while building partnerships ahead of a wider public launch. This move solidifies Google's position in the security landscape, challenging Anthropic's early momentum and addressing the urgent needs of governments and corporations seeking robust protection for their codebases.