Downtime costs reach US$19 million per hour (approximately £14.07 million) on average when endpoint devices become inoperable, according to Absolute Security’s latest research: 88% reported that autonomous recovery capabilities would enable a measurable reduction in the cost of downtime incidents, while only 4% described their company’s capabilities as “broadly autonomous.”
Absolute Security has published the third volume in its State of Enterprise Cyber Resilience research series: Autonomous Cyber Resilience in the Era of AI.
The report features results from a survey conducted by Censuswide that includes 1,000 Chief Information Security Officers (CISOs) based in the United States (500) and the United Kingdom (500).
Christy Wyatt, President & CEO at Absolute Security commenting on the findings said:
“Downtime is the real cost of a breach, it often exceeds every other related cost and is capable of permanently damaging a business. With threat actors now leveraging AI and agentic tools to launch attacks and spot vulnerabilities at rates previously unimaginable, security and risk leaders are faced with either continuing to fall victim to costly disruptions or deploying solutions that operate with the same speed and efficiency as those being used against them.”
Revealed in the report are additional findings that highlight the state of autonomous security and recovery solutions in use and the trust challenges and barriers causing organisations to lag cybercriminals. This research suggests that closing that gap will require cyber resilience capable of detecting, responding to, and recovering from disruptions at the same speed as the attacks themselves.
Among several top takeaways:
- 41% of organisations experienced AI-accelerated vulnerability exploitation in the past 12 months
- 58% of leaders are in agreement that AI-assisted attackers are weaponising vulnerabilities faster than patching processes can address them.
- For production-system autonomy, trust ranks as the top-cited barrier (38%), ahead of integration complexity (33%) and budget (18%).
When asked whether Anthropic Mythos and similar tools presented too much risk for their organisations to consider using, 43% agreed, while 32% disagreed and 25% had no opinion.
Further underscoring these concerns, 41% agreed the tools should be blocked within their organizations until more is known about how they work, while 30% disagreed and 29% had no opinion.
“Trust in autonomous solutions remains among the top barriers standing between enterprises and the recovery speed they need to minimize downtime. This research will help security leaders see the risk they’re up against and give them the confidence needed to cross the bridge to autonomous recovery,” continued Wyatt.
