A major leak tied to Anthropic has triggered sharp reactions across global markets, with software stocks and cryptocurrency related assets facing sudden declines amid concerns over the capabilities of its unreleased AI model, Claude Mythos.
The leak, which exposed internal materials and draft announcements, revealed that Anthropic has been testing what it describes as its most powerful artificial intelligence system to date. According to reports, the model significantly outperforms previous iterations in areas such as coding, reasoning and cybersecurity, raising alarm among investors and industry players.
The immediate market response was swift. Shares of major cybersecurity and software firms dropped notably, with companies such as CrowdStrike and Palo Alto Networks recording losses of up to 6 to 7 percent as investors reacted to the perceived threat posed by advanced AI systems.
The selloff reflects a growing fear that increasingly sophisticated AI models could disrupt traditional software and cybersecurity business models. If AI systems like Claude Mythos can autonomously detect vulnerabilities, write secure code or even exploit systems, the role of existing software tools could be fundamentally altered.
At the same time, cryptocurrency markets were also caught in the turbulence. While there is no evidence of direct manipulation, the broader concern lies in how powerful AI systems could interact with digital finance ecosystems. AI tools are already being used in crypto trading, analytics and decentralized finance platforms, meaning any leap in capability introduces both opportunity and risk.
The leak has also intensified concerns about security. Internal documents suggest the model may possess advanced cyber capabilities that could be misused if released widely. Anthropic itself has reportedly delayed full deployment due to fears that the technology could enable sophisticated cyberattacks or exploitation of vulnerabilities at a scale beyond current defenses.
This is not a theoretical risk. Previous incidents involving AI systems have already demonstrated how such tools can be leveraged in cybercrime, from automating attacks to identifying weaknesses in digital infrastructure. As AI systems become more autonomous and capable, the line between defensive and offensive use cases becomes increasingly blurred.
However, not all analysts are convinced the panic is justified. Some argue that while AI will transform the cybersecurity landscape, it is more likely to increase demand for security solutions rather than eliminate them. As threats become more complex, organizations may need even more advanced protection systems, potentially benefiting the same companies currently experiencing market pressure.
There is also a broader structural shift at play. The leak highlights how AI is beginning to challenge the foundations of traditional software models, particularly those based on subscription services. If AI agents can perform complex tasks that previously required multiple tools, entire segments of the software industry could face disruption.
For the crypto sector, the implications are equally significant. AI driven automation is already embedded in trading systems, smart contract analysis and fraud detection. A more powerful model could accelerate innovation but also amplify systemic risks, especially if used maliciously or without adequate safeguards.
Ultimately, the reaction to the Claude Mythos leak underscores a deeper reality. Markets are not just responding to what AI can do today, but what it might soon become capable of.
For now, the volatility reflects uncertainty rather than confirmed disruption. But one thing is clear. The race to build more powerful AI systems is no longer just a technological competition. It is increasingly a financial and security issue with global consequences.
Anthropic and OpenAI CEOs condemn ICE violence while praising Trump
