Nobody ran the Key Bridge's collapse math until six were dead
Summary
On March 26, 2024, the containership Dali lost power and struck a pier of Baltimore's Francis Scott Key Bridge, which collapsed and killed six construction workers. The bridge, built in 1977, predated a 1991 industry standard requiring owners to calculate their bridges' collapse risk from vessel strikes -- and no one ever had. In March 2025, NTSB calculated it after the fact: the Key Bridge's risk was almost 30 times the acceptable threshold. NTSB also identified 68 other bridges nationwide, owned by 30 different entities across 19 states, built before that 1991 standard with no current risk calculation on file -- including the Golden Gate, Brooklyn, and Mackinac bridges, and the Sunshine Skyway, the bridge whose own 1980 collapse prompted the standard in the first place. NTSB's November 2025 final report confirmed the technical cause -- a loose wire -- and checked in on those 30 owners: 8 months after the urgent recommendation, more than half hadn't finished the math, and of those who had, more than one in three found a bridge over the threshold, including the same Maryland agency that owned the Key Bridge, for a second bridge.
A risk nobody calculated until it was too late
The Key Bridge opened in 1977, fourteen years before AASHTO's 1991 guidance told bridge owners to calculate their collapse risk from vessel strikes. No one ever performed that calculation for the Key Bridge. After the collapse, NTSB did: using AASHTO's own method, summing the risk across the piers supporting the bridge's main channel span, NTSB calculated the Key Bridge's annual frequency of collapse at 0.002921 -- against an acceptable threshold of 0.0001 for a bridge of its classification. That's almost 30 times over the line: a roughly 1-in-342 annual chance of collapse from a ship strike, instead of the 1-in-10,000 the standard allows.
View data as table
| AASHTO acceptable risk threshold (critical/essential bridge) | 1 |
|---|---|
| Key Bridge's calculated risk level (2024) | 29.2 |
68 more bridges, same blind spot
NTSB didn't stop at the Key Bridge. It asked FHWA to identify every other bridge in the country crossing a waterway used by ocean-going ships, then filtered for bridges built before 1996, with piers in the water and enough vessel traffic to matter. That left 72 bridges across 19 states, owned by 30 separate entities. Four of those owners had already run the numbers. The other 68 hadn't -- meaning their real risk of collapse from a vessel strike was as unknown as the Key Bridge's had been. The list includes the Golden Gate Bridge, the Brooklyn Bridge, the Mackinac Bridge, and the Sunshine Skyway Bridge -- the same bridge whose 1980 collapse, from a ship strike, is what prompted AASHTO's standard in the first place.
NTSB issued four urgent recommendations: that FHWA, the Coast Guard, and the Army Corps of Engineers form a team to help bridge owners assess and reduce their risk, and that the 30 owners calculate their bridges' risk and report back -- developing a reduction plan if any bridge came in over the threshold.
Eight months later: a loose wire, and half the owners still not done
NTSB's final report landed on the technical cause: a signal wire had disconnected from a terminal block because wire-label banding had been installed in a way that kept it from seating properly, triggering the blackout that left the Dali without steering or propulsion in the seconds before it hit the bridge. The report was explicit that the missing risk calculation was a contributing cause of the collapse and the deaths -- not just the wire.
It also checked in on the 30 bridge owners. Responses accounted for all of them, but only 13 had actually finished the calculation NTSB asked for. Of those 13, 8 came in under the threshold. Five did not -- and one of the five was the Maryland Transportation Authority itself, which found that both spans of its separate Bay Bridge, over the Chesapeake, also exceeded the risk threshold. The other 14 owners were still in progress: some had hired a consultant, some hadn't gotten further than starting the math themselves.
View data as table
| Completed the calculation | 13 |
|---|---|
| Still in progress (hired/hiring a consultant or started independently) | 14 |
View data as table
| Below the AASHTO threshold | 8 |
|---|---|
| Above the threshold on at least one bridge | 5 |
The final report also issued 18 new recommendations -- on redundant ship propulsion systems, terminal-block labeling standards, voyage data recorder requirements, motorist warning systems for bridges, and emergency communication procedures for workers on them, among other things.
The takeaway
- The Key Bridge's risk was calculable all along -- it just was never calculated. Using a math the industry adopted in 1991, NTSB found the bridge's actual collapse risk from a vessel strike was almost 30 times the acceptable threshold. Nobody ran that number until after the bridge was already down.
- The Key Bridge wasn't unusual -- it was just the one that collapsed first. NTSB found 68 more bridges nationwide, across 19 states and 30 owners, with the identical gap: built before the 1991 standard, never assessed. That list includes the Golden Gate, the Brooklyn Bridge, and the Sunshine Skyway -- the bridge whose own collapse created the standard being ignored.
- Eight months after an urgent warning, the response is still partial. More than half the responding owners hadn't finished their calculation by the final report. Of the ones who had, more than a third found a bridge over the threshold -- including Maryland's own transportation authority, discovering the same problem in a second bridge it owns.
Findings on the Key Bridge's calculated collapse risk, the 68 unassessed bridges nationwide, and the four urgent recommendations are from NTSB Marine Investigation Report MIR-25-10, 'Safeguarding Bridges from Vessel Strikes: Need for Vulnerability Assessment and Risk Reduction Strategies' (March 18, 2025, reissued March 24, 2025 with footnote corrections only), read directly in full from NTSB's website. Findings on the probable cause and the bridge owners' response status are from NTSB Marine Investigation Report MIR-25-40, 'Contact of Containership Dali with Francis Scott Key Bridge and Subsequent Bridge Collapse' (November 18, 2025), also read directly in full. The two reports are NTSB's interim warning and final report on the same investigation, about eight months apart -- the first calculating a risk nobody had measured before six people died, the second confirming the technical cause and reporting that the nationwide gap it identified was, eight months on, still only partly closed.
Sources(2) ▾
- National Transportation Safety Board, Safeguarding Bridges from Vessel Strikes: Need for Vulnerability Assessment and Risk Reduction Strategies (2025-03-18) — MIR-25-10, an interim safety recommendation report issued during NTSB's investigation of the Dali/Key Bridge collapse, finding the Key Bridge's calculated collapse risk was almost 30 times the AASHTO safety threshold and identifying 68 other bridges nationwide (30 owners, 19 states) with no current vulnerability assessment -- the warning facet. Distinct from doc-ntsb-mir2540 (the final report, published about 8 months later, confirming the technical cause and reporting on owners' responses to the urgent recommendations). Reissued March 24, 2025 with footnote corrections only (no substantive change). Direct fetch succeeded (HTTP 200). ntsb.gov · original document
- National Transportation Safety Board, Contact of Containership Dali with Francis Scott Key Bridge and Subsequent Bridge Collapse (2025-11-18) — MIR-25-40, NTSB's final accident report on the Dali/Key Bridge collapse, confirming the probable cause and reporting on bridge owners' progress responding to the March 2025 urgent recommendations -- the confirmed-outcome facet, published about 8 months after doc-ntsb-mir2510's warning. Direct fetch succeeded (HTTP 200). ntsb.gov · original document
Comments
Always open. Logged-in readers can annotate paragraphs in place.
On March 26, 2024, the containership Dali lost electrical power and propulsion while leaving Baltimore Harbor and struck a pier of the Francis Scott Key Bridge. The bridge collapsed. Six construction workers died. In March 2025⧉, NTSB calculated, for the first time, what the bridge's actual risk of collapse had been. In November 2025⧉ -- about eight months later -- NTSB's final report confirmed the technical cause and checked on what had changed since.