Such capabilities matter because zero‑days are among the most valuable assets in cyber operations. Discovering them traditionally requires extensive manual research by highly specialized security experts. AI models capable of scanning massive codebases and systems could dramatically accelerate that process.

The model’s capabilities have raised concern among policymakers because they are inherently dual‑use: the same technology that helps defenders find vulnerabilities can also help attackers exploit them.

OpenAI’s Cybersecurity Model Enters the Race

Anthropic is not alone in developing AI tools for cyber defense. OpenAI has introduced its own advanced cybersecurity model aimed primarily at defenders.

Access to this system is restricted to vetted organizations through a controlled program, allowing approved security professionals to identify vulnerabilities, analyze malware, and strengthen defenses while limiting potential misuse.

The restricted rollout reflects growing concern across the technology and national security communities that highly capable cyber AI systems could be abused if released broadly.

China’s Push to Scale AI Cyber Defense

While U.S. companies currently appear to lead in cutting‑edge cyber AI models, China is moving aggressively to build its own ecosystem.

Reports indicate China is expanding its AI‑driven cybersecurity market and research capacity in response to the rapid advances from American developers.

Industry forecasts suggest the country’s AI cybersecurity sector could reach about $8.7 billion by 2030, reflecting a massive expansion as organizations adopt AI‑based defense tools.

This push includes increased investment in automated vulnerability discovery, defensive monitoring systems, and AI‑assisted cyber operations designed to operate at large scale.

How Soon Could China Match U.S. Capabilities?

Despite the current U.S. advantage, experts caution that the technological gap may not last long.

Some analysts warn that adversaries could develop systems comparable to Mythos within roughly six to twelve months, given the rapid pace of AI research and the availability of open research techniques.

Former U.S. National Security Council technology official Chris McGuire has also warned that China could soon produce an AI cyber capability similar to Mythos, potentially narrowing America’s lead and increasing pressure to strengthen critical infrastructure defenses.

If that happens, the race could shift from a technological lead to a competition in deployment scale and operational integration.

How AI Could Change the Balance Between Attackers and Defenders

The most important shift may not be who builds these models first, but how they change the tempo of cyber conflict.

AI systems capable of automated vulnerability discovery dramatically reduce the time between:

• discovering a software flaw
• creating an exploit
• deploying it in real‑world attacks

This compression of the cyberattack lifecycle could make both offensive and defensive operations far faster than traditional human‑driven workflows.

For defenders, AI could automate detection, patching, and threat analysis across massive digital environments. For attackers, similar tools could enable rapid scanning of global infrastructure for weaknesses.

The Emerging AI Cyber Arms Race

Taken together, these developments point toward a new phase in cybersecurity competition.

The United States currently appears ahead thanks to advanced models such as Anthropic’s Mythos and new defensive tools from OpenAI. But China’s rapid scaling of AI cyber capabilities—and the speed at which comparable systems might emerge—suggest the advantage could narrow quickly.

Ultimately, the decisive factor may not be which country builds the most powerful model first. Instead, the outcome may depend on who can deploy AI for cyber defense fastest, integrate it safely into infrastructure, and maintain control over its dual‑use risks.

Such capabilities matter because zero‑days are among the most valuable assets in cyber operations. Discovering them traditionally requires extensive manual research by highly specialized security experts. AI models capable of scanning massive codebases and systems could dramatically accelerate that process.

The model’s capabilities have raised concern among policymakers because they are inherently dual‑use: the same technology that helps defenders find vulnerabilities can also help attackers exploit them.

OpenAI’s Cybersecurity Model Enters the Race

Anthropic is not alone in developing AI tools for cyber defense. OpenAI has introduced its own advanced cybersecurity model aimed primarily at defenders.

Access to this system is restricted to vetted organizations through a controlled program, allowing approved security professionals to identify vulnerabilities, analyze malware, and strengthen defenses while limiting potential misuse.

The restricted rollout reflects growing concern across the technology and national security communities that highly capable cyber AI systems could be abused if released broadly.

China’s Push to Scale AI Cyber Defense

While U.S. companies currently appear to lead in cutting‑edge cyber AI models, China is moving aggressively to build its own ecosystem.

Reports indicate China is expanding its AI‑driven cybersecurity market and research capacity in response to the rapid advances from American developers.

Industry forecasts suggest the country’s AI cybersecurity sector could reach about $8.7 billion by 2030, reflecting a massive expansion as organizations adopt AI‑based defense tools.

This push includes increased investment in automated vulnerability discovery, defensive monitoring systems, and AI‑assisted cyber operations designed to operate at large scale.

How Soon Could China Match U.S. Capabilities?

Despite the current U.S. advantage, experts caution that the technological gap may not last long.

Some analysts warn that adversaries could develop systems comparable to Mythos within roughly six to twelve months, given the rapid pace of AI research and the availability of open research techniques.

Former U.S. National Security Council technology official Chris McGuire has also warned that China could soon produce an AI cyber capability similar to Mythos, potentially narrowing America’s lead and increasing pressure to strengthen critical infrastructure defenses.

If that happens, the race could shift from a technological lead to a competition in deployment scale and operational integration.

How AI Could Change the Balance Between Attackers and Defenders

The most important shift may not be who builds these models first, but how they change the tempo of cyber conflict.

AI systems capable of automated vulnerability discovery dramatically reduce the time between:

• discovering a software flaw
• creating an exploit
• deploying it in real‑world attacks

This compression of the cyberattack lifecycle could make both offensive and defensive operations far faster than traditional human‑driven workflows.

For defenders, AI could automate detection, patching, and threat analysis across massive digital environments. For attackers, similar tools could enable rapid scanning of global infrastructure for weaknesses.

The Emerging AI Cyber Arms Race

Taken together, these developments point toward a new phase in cybersecurity competition.

The United States currently appears ahead thanks to advanced models such as Anthropic’s Mythos and new defensive tools from OpenAI. But China’s rapid scaling of AI cyber capabilities—and the speed at which comparable systems might emerge—suggest the advantage could narrow quickly.

Ultimately, the decisive factor may not be which country builds the most powerful model first. Instead, the outcome may depend on who can deploy AI for cyber defense fastest, integrate it safely into infrastructure, and maintain control over its dual‑use risks.