AI-First Organizations Face Longer Recovery Times and Higher Costs After Cyber Incidents, Fastly Report Finds Fastly’s latest Global Security Research Report for Australia and New Zealand reveals that organizations identifying as "AI-first" experience significantly longer recovery times and higher financial impacts from cybersecurity incidents compared to their peers. On average, AI-first businesses take nearly seven months about 100 days longer to fully recover, a gap Fastly terms the "AI Speed Tax." The financial toll is also stark: incident costs for AI-first organizations exceed those of non-AI-first counterparts by 135%. The report highlights AI’s role in expanding attack surfaces, with 48% of AI-first organizations reporting that AI was directly exploited in their most recent incident, compared to just 10% of non-AI-first firms. Key vulnerabilities include agentic workflows and decentralized data flows, which introduce complexity for security teams. Additionally, 42% of AI-first organizations cited AI-related blind spots as contributing factors in incidents, versus 29% of non-AI-first respondents, underscoring challenges in visibility and control. Beyond direct breaches, AI adoption is driving operational costs. 75% of organizations report that AI scraping has become a material expense, with average annual infrastructure impacts exceeding AUD $595,000. Half of those surveyed noted increased infrastructure costs due to AI activity, while 48% experienced operational disruptions and 33% faced issues affecting online visitors, such as slow load times or broken functionality. In response, organizations are prioritizing security investments, with 59% focusing on agentic discoverability, 53% on API security, and 51% on web application firewalls. Concerns about AI-targeted attacks are high, with 85% of respondents worried about DDoS attacks on AI agents and 56% recognizing a need for AI-specific security expertise. Fastly’s Chief Information Security Officer, Marshall Erwin, emphasized the need for security modernization to keep pace with AI adoption, noting that the challenge now extends beyond traditional threats to managing rapidly expanding, often invisible infrastructure. The report, based on a survey of 2,000 IT decision-makers across the Americas, Europe, and Asia-Pacific, was conducted in Q4 2025.

Fastly, a leading edge cloud computing and content delivery network (CDN) provider, was identified as one of the major vendors affected by CVE-2025-8671 (MadeYouReset), a critical HTTP/2 protocol-level vulnerability. The flaw allows threat actors to exploit server-sent stream resets, enabling devastating denial-of-service (DoS) and distributed denial-of-service (DDoS) attacks. By manipulating HTTP/2 stream cancellation mechanisms, attackers force Fastly’s infrastructure to process an unbounded number of requests while bypassing concurrency limits, leading to severe CPU and memory exhaustion.The vulnerability stems from a mismatch between HTTP/2’s specification and real-world implementations, where reset streams are marked as inactive in protocol accounting but continue consuming backend resources. For Fastly, which powers high-traffic websites, APIs, and streaming services, this could result in catastrophic outages, disrupting global digital services, e-commerce platforms, and real-time applications reliant on its CDN. Prolonged exploitation might degrade performance for end-users, trigger financial losses for dependent businesses, and erode customer trust in Fastly’s reliability.While Fastly has likely released patches, unpatched systems remain at risk of large-scale disruptions, particularly if targeted by state-sponsored or criminal hacking groups seeking to weaponize the flaw against critical infrastructure. The vulnerability’s potential to halt or degrade internet-facing services positions it as a high-stakes threat to Fastly’s operational integrity and its clients’ digital continuity.