And when it does go wrong, the fallout is immediate. The collapse of the Francis Scott Key Bridge disrupted one of the East Coast’s most important shipping routes. Sinkholes at LaGuardia delayed hundreds of flights and revealed how vulnerable critical systems can be. In Hawaii, levee failures left communities exposed to flooding and long-term damage.

But the biggest warning signs are not always headline-grabbing disasters. More often, infrastructure failures develop gradually, hidden beneath day-to-day operations until the economic and social costs become too large to overlook.

In Fayetteville, Georgia, a data center campus consumed nearly 29 million gallons of water over 15 months through two pipe connections the county didn’t know existed. At the same time, local officials were urging residents to conserve water during severe drought conditions. Water pressure dropped, yet there were no early warnings, no clear visibility into rising demand, and no practical way to intervene before the system was pushed to its limits.

That kind of pressure could become more common as data centers expand to support AI. EPA estimates show U.S. data centers used 17.4 billion gallons of water in 2023, and that total could reach 73 billion gallons by 2028.

As these facilities expand into drought-stressed regions like the American West, that growth has real consequences for local systems and communities. Water operators need modern tools that provide better visibility to prevent these incidents. A system with real-time metering integration would have caught Fayetteville’s drain immediately — before 29 million gallons disappeared.

Real-time response alone is no longer enough. Infrastructure operators also need the ability to anticipate failures before they happen and evaluate “what if” scenarios before systems come under strain. Digital twin technology already makes this possible. In New Orleans, the 17th Street Canal pump station implemented a digital twin to improve decision-making during storm events, helping protect 635,000 people, assets, businesses, and critical industries while strengthening resilience against climate-related flooding.

These virtual models allow operators to simulate how infrastructure systems will respond to severe droughts, unexpected population growth, or the addition of a hyperscale data center. It shifts operations entirely from break-fix reactions to proactive maintenance.

Predictive models are only as effective as the data behind them. Although operators collect more infrastructure data than ever before, that information often remains trapped in disconnected systems. Digital twins help solve this problem by bringing those systems together into a unified operational view. Without that visibility, operators remain information-rich but operationally blind — lacking a clear, real-time understanding of how their infrastructure systems are performing.

Bridging this visibility gap requires policy as much as technology. Policymakers, operators, and the industry are increasingly working together to advance digital requirements that give all stakeholders a clearer understanding of how assets can be managed more effectively. During the recent Infrastructure Week in Washington, D.C., the House Transportation and Infrastructure Committee advanced the BUILD America 250 Act, which includes a key provision for digital infrastructure. By integrating digital delivery into federal transportation policy, this Act helps operators adopt the same digital twin technologies that have transformed the private sector.

Beyond mere modernization, adopting these predictive tools helps solve infrastructure’s persistent optics problem. Public investment in infrastructure creates an opportunity to strengthen accountability and transparency around how these systems perform. Taxpayers supporting these major upgrades should be able to see how infrastructure works in real time and how those investments are improving reliability and resilience. Policymakers can help accelerate that progress by encouraging modern approaches that prioritize visibility, predictive insight, and better long-term outcomes for federally funded projects.

The U.S. has committed over half a trillion dollars to rebuild its foundation. The next challenge is ensuring those investments are as intelligent as they are historic. Though largely invisible to the public, infrastructure increasingly determines America’s economic competitiveness. The question is whether the country will modernize these systems before the next failure forces its hand.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.