Our national water treatment and pipe infrastructure has a value of over $3.6T with nearly 60% of the value underground. University studies suggest water distribution systems have increased in the number of water main breaks overall by 27%. These studies also highlight the growing uncertainty of the performance of metallic pipes in corrosive soils and the higher break rates with pipes 12 inches in diameter and less which size represents nearly 85% of all water pipe inventories. Now, with the abundance of pipe related data and analytical models, the industry is realizing that age-based methods are not effective. Many utilities strive for a goal of replacing 1% of their pipe network each year, representing a 100- year replacement cycle metric. The primary purpose of these replacements is to avoid service disruptions (leaks, breaks, sinkholes, etc.). These disruptions obviously have documentable economic impacts with revenue loss due to water outage as well as repair costs. To guide replacement priorities, most utilities currently employ some form of subjective strategy such as age-based or scoring/matrix approaches (i.e. assigning weights and factors of 1 to 5 for various inputs). The desired intent is to replace the riskiest pipes first to avoid the occurrence and impact of failures. These kind of models can successfully reduce pipe network breaks between 5% and 10% over three years at an annual replacement rate of 1% of the system. With the increase of water main breaks, the effectiveness of the scoring/matrix prioritization model will not keep up with new breaks. An advanced multi-model approach (not just a machine learning model only) has been shown to achieve a much higher level of break reduction, 35% for each mile of pipe replaced. This pipe risk assessment methodology can justify budget reductions while maintaining current service level expectations by reducing breaks in an equal amount as before and replacing far less pipe.