Anthropic AI advances stir cybersecurity debate, but market risks may be overblown Proactive uses images sourced from Shutterstock
Shares in cybersecurity firms took a hit last week following news about Anthropic’s latest language model, Claude Mythos, which Jefferies analysts say represents a “step change” in capabilities for reasoning, coding, and cybersecurity applications.
The market reaction comes amid broader excitement over the potential for large language models (LLMs) to disrupt existing cybersecurity services. Jefferies noted that while Mythos has demonstrated impressive capabilities, its reach is likely to remain concentrated in specific areas of cyber, including threat intelligence, code scanning, vulnerability management, and automated response operations. Identity management, network security, and endpoint protection appear relatively insulated, the analysts said.
“LLM advancements are not without any risk of cyber encroachment, but the most impactful debates are likely around incremental cyber budgets rather than broad disruption,” Jefferies wrote, citing survey data from 30 chief information officers showing that organizations see threat intelligence and code security as the areas most open to AI solutions, while core cybersecurity spend is expected to remain steady.
Anthropic has pre-released Claude Mythos to select cybersecurity vendors, Jefferies highlighted, suggesting a collaborative approach aimed at improving robustness rather than competing directly with incumbents. “This signals partnership rather than competition, which is positive for overall cyber demand,” the analysts wrote.
Jefferies also pointed to hiring trends at Anthropic as an indicator of the company’s cyber ambitions, noting positions focused on code scanning, investigations, penetration testing, and vulnerability management. The bank argued that while LLMs may eventually move into other cybersecurity segments, broad disruption is limited by technical differentiation, the value of first-party data, and the need to gain CISO trust.
The firm cautioned investors to expect persistent headline risk, as leading LLM providers continue to announce major model improvements. “While these headlines create uncertainty, model upgrades should not come as a surprise,” Jefferies wrote.
The analysts suggest that names with strong fundamentals and defensible AI positioning in Identity, Network Security, and Endpoint sectors could present attractive opportunities amid temporary market volatility.
