126 Years to Replace Every Bridge We Have

Beams buckled in a Manhattan tower this week. That building was brand new construction. The real problem is the half that isn't.

On the morning of July 7, construction workers converting the old Pfizer headquarters at 235 East 42nd Street into apartments watched the structural support beams on the 21st and 22nd floors begin to buckle. They self-evacuated. Floors 21 through 26 started to cave under the stress. The FDNY evacuated eight surrounding buildings, including a Hampton Inn, and shut down two full blocks of Midtown East traffic. As of this writing the investigation is ongoing and no cause has been confirmed.

This is worth being precise about, because precision is the whole point of this essay: that building is not old. It is a 38-story tower undergoing active conversion, meaning whatever failed, failed during construction, not from decades of neglect. It is not, on its own, evidence of the thing this essay is actually about. What it is, is a jolt, the kind of headline that makes you look up from your phone and ask an entirely reasonable question: how old is everything else, and how many of the buildings and bridges and pipes holding this country together are one bad week from doing the same thing without the courtesy of self-evacuating construction workers to notice first?

The answer, when you actually pull the numbers, is not reassuring, though it is also not the cliff-edge collapse scenario the question implies. It is something slower and in some ways more insidious: a country that has been quietly and knowingly deferring maintenance on itself for half a century, doing just enough to avoid the worst outcomes, and calling the result progress.

The number that matters most: fair

The American Society of Civil Engineers has graded the nation’s infrastructure every four years since 1998. In March 2025 the overall grade rose to a C, up from a C-minus in 2021, the best mark the country has ever received. This sounds like good news, and in a narrow sense it is: eight of eighteen categories improved, none scored in the D-minus range for the first time ever, and $1.2 trillion from the 2021 Infrastructure Investment and Jobs Act has visibly moved the needle on roads, bridges, dams, and ports.

But look at what a C actually means in ASCE’s own grading language: “mediocre and requires attention,” infrastructure in fair to good condition that “shows general signs of deterioration” with “some elements exhibiting significant deficiencies.” That is the best grade America’s infrastructure has earned in the twenty-seven years anyone has been formally grading it, and it translates to mediocre with significant deficiencies. Nine of eighteen categories still sit in the D range. Stormwater and transit both earned a D.

The gap between what needs fixing and what’s funded to be fixed stands at $3.7 trillion over the next decade. ASCE calculates $9.1 trillion is needed to bring the country’s infrastructure into a genuine state of good repair by 2033. Only $5.4 trillion is currently expected under existing funding trajectories. If federal investment reverts to pre-2021 levels, the gap widens to $4.4 trillion. This is not a hypothetical bill. It is a bill that has already been incurred and is accumulating interest in the form of deteriorating steel and concrete every single day it goes unpaid.

The bridges

There are 623,000 bridges in the United States. The average bridge is now 47 years old against a traditional 50-year design life, which means the typical American bridge has, on paper, about three years of designed service life left, a fact that is somewhat less alarming than it sounds only because “design life” is a conservative planning benchmark, not an expiration date, and because minimum design standards have since been extended to 75 years for new construction. Still: 45 percent of all bridges in the country now exceed their originally intended 50-year design life. Of the 623,000, 6.8 percent are rated in poor condition and 49.1 percent are rated fair, meaning essentially half the bridge inventory in America is aging and one maintenance cycle away from sliding into poor. ASCE explicitly warns that fair-condition bridges are cheaper to save now than to replace later, and that failing to act on them is how a country ends up with more poor bridges, not fewer.

At the current replacement rate of roughly 4,900 bridges per year, it would take 126 years to replace every bridge currently in the national inventory. The funding gap for bridge rehabilitation alone is $373 billion over the next decade, even after $40 billion in dedicated IIJA bridge funding.

The 2023 collapse of a section of I-95 in Philadelphia, after a tanker truck fire, is the kind of acute failure that makes news. The slower story, the one that doesn’t generate a single dramatic headline, is the one where a bridge quietly slides from fair to poor over a decade of underfunded maintenance and nobody notices until the day it does something dramatic.

The buildings themselves

This is the category most directly relevant to what happened in Midtown this week, and it’s also the category where the data is thinnest, because unlike bridges, dams, and roads, there is no single federal agency tracking the condition of America’s building stock the way ASCE tracks transportation infrastructure.

What we do know: the average U.S. commercial building was 53 years old as of the most recent comprehensive survey, up from a stable average of roughly 50 years across surveys conducted in 2011, 2014, and 2017. Seventy-two percent of American commercial buildings are 20 years or older, the age at which structural and mechanical systems typically begin needing serious attention. Sixty-eight percent of national office inventory was built before 2000. In gateway cities, more than 70 percent of office space is over 30 years old.

Facilities management research describes a fairly consistent pattern in how buildings age: light maintenance and cosmetic work for the first 40 years, then a threshold around 40 to 50 years old at which point hidden distribution systems, plumbing, electrical, elevators, fire alarms, and building facades all begin failing more or less simultaneously, requiring the kind of comprehensive renovation that gets deferred precisely because it’s expensive and invisible until it isn’t.

The median residential building in New York City, a useful proxy for the oldest and most infrastructure-stressed housing stock in the country, is nearly 90 years old. Nationally, 37 percent of all housing structures were built before 1959, meaning they are now 65-plus years old. New York had two prominent building collapses in 2023 alone, and structural engineers researching the trend point to the same conclusion the bridge data points to: age alone doesn’t cause failure, but age combined with deferred maintenance is a near-guarantee of eventual failure, and deferred maintenance is what happens by default when nobody is required to budget for it.

What is, and isn’t, being done

The honest answer is: something, but not enough, and the something that exists is politically fragile.

The 2021 Infrastructure Investment and Jobs Act was the largest federal infrastructure investment in generations, and the improvement from a C-minus to a C is real and traceable to that spending. As of late 2024, over $568 billion of the law’s $1.2 trillion had been allocated across 66,000 state-level projects. Bridge programs alone received $40 billion, funding more than 12,300 new bridge projects.

But that law expires. Its funding is not permanent, and the political consensus behind renewing it at scale does not currently exist. The House Transportation and Infrastructure Committee introduced its 2026 reauthorization framework in late June, described by its own chairman as a return to a more “traditional” and narrower approach focused on asphalt and concrete rather than the broader resilience and climate provisions of the 2021 law. Nobody involved in that process is promising the money will keep pace with the $3.7 trillion gap. Most of the public conversation around infrastructure in 2026 has been consumed by other things: wars, tariffs, an administration actively defunding the scientific agencies that measure the climate risks ASCE says are now the single biggest threat multiplier to aging infrastructure, given that 27 separate billion-dollar weather disasters struck the US in 2024 alone, causing over $182 billion in damage and 568 deaths.

There is also no equivalent of the ASCE Report Card for the country’s private building stock. Bridges and dams get systematically graded because they’re public infrastructure with public accountability built into the funding structure. A 90-year-old apartment building in Queens or a 53-year-old office tower in Cleveland gets graded by nobody until an inspector, an insurance adjuster, or a construction crew on the 21st floor happens to notice something buckling.

What this actually adds up to

Nobody who has looked closely at this data is predicting an exponential wave of collapsing buildings, and this essay isn’t either. That’s not how deferred maintenance typically fails. It fails the way it failed in Philadelphia in 2023, and the way it may have just failed in Midtown, and the way the Champlain Towers South condominium failed in Surfside, Florida, in 2021, killing 98 people: isolated, sudden, geographically scattered, each one initially looking like a local anomaly rather than a symptom of a systemic condition, until you zoom out far enough to see the pattern the ASCE report card has been documenting in flat, bureaucratic letter grades for almost three decades.

The country’s infrastructure got its best grade ever this year. The best grade ever is a C, defined in the grading agency’s own words as mediocre with significant deficiencies. Half the bridges in America are aging out of good condition into fair, and a quarter of that half is one deferred inspection away from poor. The typical commercial building is over 50 years old, sitting past the threshold where hidden systems begin failing in earnest, in a country with no systematic mechanism for tracking that risk the way it tracks bridges. And the federal spending that has been the single documented cause of the recent, modest improvement is neither permanent nor currently being renewed at the scale its own overseers say is necessary.

This is not a country on the verge of falling down. It is a country that built an extraordinary amount of infrastructure in a fifty-year window after World War II, is now watching nearly all of it arrive at the same age simultaneously, and has spent the last several decades treating the resulting maintenance bill as an optional expense rather than the fixed cost it actually is.

The beams on the 21st floor buckled Tuesday morning. Nobody was hurt. The workers noticed and walked out before anything came down. That is what worked this time. It is worth asking, plainly and without hysteria, how many more times the country intends to rely on somebody noticing in time

.