A Warning From 20 Feet Below
In 2010, workers rebuilding New Yorkās World Trade Center uncovered an unexpected piece of the past. About 20 feet belowĀ ground,Ā excavation machinery struck weathered timber buried at the edge of the site. It turned out to be part of an 18th century sloop that once sailed the Hudson River and was later used as landfill as Manhattan expanded.Ā
For historians, it was a remarkable find. For engineers, it was a reminder of a more practical problem.Ā Sometimes infrastructure is still designed with incomplete knowledge of what lies below the surface, and that uncertainty can be expensive.Ā
“Sometimes we design in the dark. Sometimes we find things and they are massive, and we start to bleed,” said Rizwan Baig, chief engineer of the Port Authority of New York and New Jersey.āÆ“Not knowing what’s below the ground is what concerns me.”āÆĀ
That concern framed one of the clearest messages at theĀ Transforming Infrastructure Performance (TIP)Ā NYC Summit 2026 in New YorkĀ organized by Bentley Systemsā Infrastructure Policy AdvancementĀ (IPA)Ā team andĀ theirĀ partners. Better infrastructure outcomes start with better information, captured early enough, structured well enough, and shared widely enough to influence real decisions.Ā
A $100 Billion Challenge
Held in mid-April, the second annual TIP NYC Summit brought together more than 170 infrastructure leaders from engineering, government, finance, and technology. The conversation ranged from permitting and project finance to mega-project delivery, fragmented information, and the growing role of AI.Ā
The backdrop was hard to ignore. The New York metropolitan area is expected to see roughly $100 billion in infrastructure work over the next decade. The Port Authority alone is managing a $45 billion project budget spanning airports, terminals, rail, and bridges.Ā
Connected data was the dominant theme. The industry knows that information should move cleanly from planning through design, delivery, operations, and maintenance. The deeper issue is that too little of that data becomes operational intelligence that decision-makers can use in time.Ā
“There’s trillions of dollars’ worth of infrastructure assets across the U.S., but you go to a lot of states and local governments, and they don’t even know what infrastructure exists in the first place,”āÆsaidāÆRory Linehan, director for infrastructure policy advancement at The Infrastructure Policy Advancement, Bentley Systems’ in-house global think tank.āÆ“We need to find a better way to captureāÆthatāÆdata people can trust and useāÆin real-timeāÆ, andāÆleverage it to inform better decision-making. Otherwise, theāÆoutcomeāÆis friction, rework,āÆdelaysāÆand lost value.”āÆĀ
Without that, the result is familiar: friction, rework, delay, and lost value.
Money Follows the Data
That matters not only to engineers and operators, but also to investors.Ā
As public budgets tighten, private capital is taking a larger role in infrastructure finance. And private investors prize predictability. They want evidence that projects will be delivered on time, on budget, and with credible governance around risk. “Everything we do is grounded in data, that’sāÆ100%,”āÆsaid Anmay Ditman, managing director and head of the Climate Finance Partnership at BlackRock.Ā
Ditman moderated a panel on funding next-generation transit that included Allison L.C. de CerreƱo, chief operating officer for bridges and tunnels at the Metropolitan Transportation Authority; Sabrina Sussman, chief program officer for Choose How You Move, Nashvilleās transportation improvement program; and Samantha Biddle, deputy secretary of transportation at the Maryland Department of Transportation.Ā
The examples wereĀ real.Ā DeĀ CerreƱoĀ described how data and advanced algorithms supported Manhattanās congestion-pricingĀ program.Ā Biddle is helping oversee the reconstruction of Baltimoreās Francis Scott Key Bridge after its collapse in 2024.Ā
WhatĀ linkedĀ the panelists was not simply a belief in data, but a belief in evidence-backed deliveryĀ supported byĀ stakeholders. New financing models and ambitious infrastructure programsĀ work only when leaders can show the basis for decisions and build confidence across institutions and communities.Ā
From One Time Collection to Operational Intelligence
For Biddle, the issue is notĀ just betterĀ data. It isĀ faster decisions.Ā “We have to get beyond one-time data collection”, she said after the panel.Ā It has to be operational in a way that impacts decision-makers, impacts stakeholders, and drives the business forward.
That is a useful test for the industry. If data does not improve the speed and quality of decisions, it is not solving the right problem.Ā
Biddleās point was blunt. Costs are rising constantly. Teams may have the right people and the right intent, but performance depends on whether the information ecosystem is integrated enough to let those teams act quickly.Ā “We are in this climate right now where every minute, every hour things are costing more,” she said. We want to be as efficient as possible. “We have great teams. We have the right talent. It is just a matter of having all the pieces together and letting them run.”
AI Is Only As Smart as it's Data
No infrastructure conference in 2026 was going to avoid AI. But the discussion in New York was notably grounded.Ā
For now, most practical AI use cases in infrastructure are focused on productivity, decision support, and visualization. That is important. But it is not yet the same as fundamentally reshaping how infrastructure is planned and delivered at the system level.Ā
David Goldwater, senior vice president for public policy at Stantec, told an AI panel that the “technology could be transformative for the sector, provided it comes with quality control and human oversight.”Ā
“The prospect of AI designing a bridge sounds a little terrifying,” he said.Ā
He said that Stantec is already applying AI to accelerate project delivery, predict soil conditions from sample data, streamline permitting, and improve public engagement through faster 3D modeling and visualizations.Ā If we can get better data, we can make significant efficiencies,āÆGoldwater said. “We canĀ better communicateĀ with the public. We can better show things because we can quickly create 3D models and visualizations,”Ā he said.Ā
Panelists also argued that faster delivery does more than cut costs. It can widen economic benefits. “By automating and streamlining and expediting the pain points along the way to that construction start, we as a society will have more and better infrastructure, but we will also create a lot more jobs that way,” said David Gilford, infrastructure strategist and managing partner at PolicyAlpha, who moderated the AI panel.Ā
The Unexpected Early Adopters
Gilford pointed to a notable pattern.Ā Some of what are seen as the most traditional industries, whether power generation or water infrastructure, where the workforce is aging, and there has been a shortage of new people entering the field, actually seem to be among the earliest adopters of these tools, he said.Ā
That may seem counterintuitive, but it makes sense. The sectors under the greatest delivery pressure often have the strongest incentive to use technology that can extendĀ expertise, reduce repetitive work, and help manage scarce talent.Ā
Still, the AI conversation kept circling back to the same prerequisite: data discipline.Ā Robert Kumapley, chief of the enterprise asset management division at the Port Authority, said that his agency was “piloting AI and natural-language search to analyze maintenance records and identify root causes of repairs on the 94-year-old George Washington Bridge.”Ā But withoutĀ standardization, heĀ warned,Ā AI will struggle to generate reliable insight.Ā
This is why data structureĀ standardization is important, because if you do not have that data structure, AI is not going to tell you anything,Ā KumapleyĀ said. We haveĀ maybe 50Ā consultants doing inspections. We do not want some to say “rust” and others to say “corrosion”. That is not going to work.Ā They all have to use the same acronym, the same nomenclature.Ā
It was one of the most practical observations of the day. Clean data is not an administrative exercise. It is the condition forĀ scale.Ā
Breaking Silos Is The Start, Not the finish
The summitās value was not only in the subjects discussed, but in the mix of people in the room. Engineers, financiers, public officials, operators, and technology leaders rarely get enough time together to understand each otherās constraints, incentives, and blind spots.Ā
“So, few times do we ever come together as an entire infrastructure industry to better understand what the challenges are and how we can work together to solve them,” Linehan said. “If we do not do events like this, we cannot break down our silos, we cannot learn from each other, and ultimately, we cannot deliver better outcomes for our communities.”Ā
Biddle agreed. For her, one of the most useful signals from the morning was that industry wanted to engage seriously with public-sector problems.Ā “Industry wants to be part of the equation”, she said.āÆ“What I really enjoyed this morning is that they are willing to listen and really understand where our challenges are, what our problems are, and how they can be part of bringing to market projects that are going to succeed.”Ā
That matters. But discussion alone is not enough.Ā
TIP puts engineers, investors, and officials in one room. That isĀ an important step. It is not the finish line. TheĀ real challenge, and the real opportunity, lies in converting infrastructure insight into execution.Ā
That is a central part of the mission for Bentley Systemsā Infrastructure Policy Advancement unit. Through TIP events in London, Toronto, Melbourne, and New York City, IPA is working to move the conversation beyond diagnosis and toward practical reform in how infrastructure is designed, delivered, andĀ operated.Ā
The message from New York was clear. AI may help the industry move faster. But it will not modernize infrastructure on its own. That will depend on whether leaders can build the data foundationsĀ requiredĀ to make better decisions, earlier and with more confidence.Ā
FAQ:
According to experts at the TIP NYC Summit 2026, the U.S. infrastructure sector currently has a massive volume of data, but it is often fragmented, outdated, or unstandardized. For AI to be transformative, the industry must prioritize data discipline.
In 2026, AI is mostly focused on productivity, decision support, and 3D visualizations rather than completely reshaping system-level planning. Engineering firms are currently applying these tools to accelerate project delivery, predict soil conditions from samples, and streamline the permitting process, creating significant efficiencies.
Private capital prizes predictability above all else, requiring evidence that infrastructure projects will be delivered on time and within budget. Since everything investors do is “100% grounded in data,” maintaining clean and trusted data frameworks is the only way to build the confidence needed to secure funding for next-generation transit.
