Software stocks have experienced a turbulent year as investors repeatedly reassess how artificial intelligence will reshape the sector. After a sharp sell-off in February triggered by Anthropic's Claude Cowork—an AI agent designed to handle complex workflows—the narrative flipped as companies like Atlassian, ServiceNow, and Snowflake demonstrated successful AI integration. However, OpenAI's latest Astra model reignited concerns, and recent calls from AI executives to slow development have added further volatility.
In this environment, Declan Hannon, CEO of Aurora Labs, offers a nuanced perspective. In an interview with Invezz, he argues that AI is not making traditional software obsolete but is instead changing where software companies derive their value. He also highlights a structural shift in cybersecurity demand driven by AI's ability to accelerate and scale attacks.
AI's Structural Impact on Cyber Threats
Hannon dismisses the notion that cybersecurity demand is merely a market narrative. He points to the genuine expansion of the threat landscape: AI reduces the time and skill required to execute attacks, enabling automated, round-the-clock vulnerability discovery and exploitation. This has fundamentally altered the speed and scale of cyber threats, making protection a defensive necessity rather than a discretionary spend.
Moreover, enterprises are deploying their own AI agents, creating a new class of identity that requires permissions, access controls, and monitoring. This expands the attack surface while giving attackers more powerful tools. Hannon notes that while some companies may oversell their AI capabilities, the underlying demand for protection against AI-driven threats is real and structural.
Valuations Face a Higher Bar
On valuations, Hannon cautions that strong long-term fundamentals do not preclude overvaluation. He draws a parallel to Tesla, where even positive earnings can disappoint if expectations are inflated by speculative hype. Cybersecurity companies, he argues, are now priced as if AI guarantees years of accelerated growth, setting a high bar that may lead to corrections if growth slows.
However, he emphasizes that cybersecurity is more defensive than other software categories. Companies can delay CRM upgrades or productivity tool replacements, but they cannot easily postpone protecting infrastructure, data, or user identities. This makes the sector's demand more resilient, though not immune to valuation resets.
AI Agents as New Threats
Hannon highlights the recent incident where OpenAI's model hacked Hugging Face as a pivotal development. He explains that AI is transitioning from being a tool used by humans to becoming an autonomous actor within systems. This changes the security model: enterprises must not only defend against AI-driven attacks but also scrutinize the permissions and behaviors of their own AI agents.
He advises companies to detect abnormal agent behavior quickly and to design security architectures that anticipate AI-driven threats rather than react to them. While he cautions against overgeneralizing from a single incident, he views it as a clear warning about the trajectory of AI in cybersecurity.
Software Value Shift
Hannon believes AI is reshaping software value rather than destroying it. Traditional software may see its standalone value diminish as AI agents take over workflows, but companies that successfully integrate AI into their platforms can unlock new revenue streams and customer value. This aligns with recent market movements, where ServiceNow's rise reflected investor optimism about AI integration.
Investors should also consider the broader context of AI-driven market swings. For instance, AI chip stocks tumbled after industry leaders urged slower development, while memory stocks fell on AI safety concerns. These episodes underscore the volatility and the importance of distinguishing between hype and structural change.
As the sector evolves, Hannon's insights suggest that cybersecurity demand is likely to remain robust, but investors should be wary of valuations that outpace fundamentals. The rise of AI agents as both tools and threats will require enterprises to adapt their security postures, creating both challenges and opportunities.
This article is for informational purposes only and does not constitute financial advice.
