The security of software is increasingly determined by how much AI compute was spent trying to break it. Companies use diverse AI agents to autonomously attack their own code. The amount of money spent on APIs from labs like Anthropic to find vulnerabilities is becoming the best indicator of a system's robustness.

The high cost of defending against advanced AI cyber threats could bankrupt small and medium-sized businesses in the defense industrial base. Their inability to afford next-generation security, like dedicated hardwired networks, threatens to cripple the military's supply chain for critical components.

Defensive AI systems deployed in the real world must use approved, often older models. Meanwhile, attackers (or models in testing) can leverage the newest, most powerful frontier models, creating a fundamental and dangerous asymmetry where defense always lags behind offense.

Nikesh Arora argues that using large, token-based AI models to inspect massive data flows in cybersecurity is cost-prohibitive. The viable strategy is deploying specialized small language models (SLMs) for widespread monitoring, where marginal cost is zero, reserving expensive models only for confirmed threats.

Hugging Face found that leading commercial AI APIs were unusable for incident response. Their safety guardrails blocked the analysis of real attack data, unable to distinguish a defender from an attacker. The team had to use a less-restricted, open-weight Chinese model on their own infrastructure to perform the necessary forensic analysis.

A call from tech companies for government-funded AI cyber defense is being challenged. Critics argue that companies generating massive profits from AI, such as NVIDIA, should bear the financial responsibility for mitigating the risks they create, especially given the government's poor fiscal situation, rather than shifting the cost to taxpayers.

While large firms use AI for defense, the same tools lower the cost and barrier to entry for attackers. This creates an explosion in the volume of cyber threats, making small and mid-sized businesses, which can't afford elite AI security, the most vulnerable targets.

The rise of AI dramatically increases the 'quantity and quality' of cyberattacks, allowing bad actors to automate attacks at scale. This elevates security from a compliance issue to an existential risk for startups, who often lack dedicated teams to combat these advanced, persistent threats. A severe hack is now a company-killing event.

Despite staggering costs—some testers spent over $1M in tokens in weeks—cybersecurity firms are not hesitating to expand budgets for Anthropic's Mythos model. The platform's ability to find critical code vulnerabilities provides a return on investment that makes the extreme expense a necessary cost of doing business in an AI-driven threat landscape.