AI Is Eating the Insurance Industry. Can Verisk Survive?

The rise of artificial intelligence is forcing a reckoning at Verisk Analytics, the data analytics giant that powers much of the global insurance industry [219545]. The core question is no longer whether AI will change insurance, but whether Verisk’s own tools will become obsolete [219545]. For decades, Verisk has built its business on proprietary datasets and complex actuarial models, with insurers paying a premium for its catastrophe-risk maps, fraud detection, and underwriting scores [219545]. But new generative AI models can now process vast amounts of public data—weather patterns, building codes, and claims history—without needing Verisk’s legacy infrastructure [219545].

The threat is not hypothetical. Several startups are already using machine learning to price risk in real time, undercutting Verisk’s slower, batch-processed models [219545]. Meanwhile, large cloud providers are offering AI tools that let insurers build their own analytics in-house [219545]. If insurers can generate the same insights for a fraction of the cost, Verisk’s pricing power will erode [219545]. Verisk is not standing still, having invested in its own AI capabilities, including machine-learning models that predict claim severity and detect fraud faster than human analysts [219545]. It is also bundling its data with new AI-driven “decision engines” to keep clients locked into its ecosystem [219545].

The competitive window is closing, however, and the next 18 months will be decisive [219545]. Verisk must prove that its proprietary data remains the “ground truth” that AI models need to be accurate [219545]. If it succeeds, AI becomes a tailwind; if it fails, Verisk becomes just another data vendor in a world where intelligence is free [219545].

The pressure on the industry comes as new data shows the number of breach notifications has blown past the full-year figure for 2025 in record time [219516]. Security experts point to two main factors: the increasing use of AI by cybercriminals to automate and scale attacks, and a spike in so-called "malicious insider" incidents, where employees or contractors deliberately leak or steal data [219516]. AI tools are now being used to craft more convincing phishing emails, find system vulnerabilities faster, and launch attacks that adapt in real time [219516]. Meanwhile, the rise in insider threats suggests that access to sensitive information remains a critical weak point for many organizations [219516].

Beyond security, AI is also having a negative impact on the climate. According to recent studies, AI is currently increasing greenhouse gas emissions from fossil fuels—not reducing them [219188]. The problem is especially visible in North Rhine-Westphalia, Germany’s industrial heartland, where data centers that power AI systems are consuming massive amounts of electricity, much of which still comes from coal and natural gas [219188]. AI models require huge computing power for training and daily operations, pushing energy grids to their limits and keeping older, dirtier power plants running longer [219188]. The studies conclude that, so far, AI’s net impact on the climate is clearly negative, and experts warn that without urgent policy changes—such as stricter efficiency standards for data centers or a faster shift to clean power—the AI boom could undermine global climate goals [219188].