The Federal Highway Administration issued a record of decision for improving this crossing in 2001. Full notice to proceed for construction followed on 11 September 2020. The nineteen years in between are not an administrative footnote; they are the clearest available answer to why adding capacity across Hampton Roads is so hard.
Virginia’s own financial plan sets out the sequence. A major investment study finished in 1997, a draft environmental impact statement followed in 1999 and a final one in 2001, with the record of decision that year. A 2003 re-evaluation validated the earlier decisions. A 2011 environmental assessment was revised in 2013 and never received a federal decision at all, and was not advanced because of fiscal constraint. A 2012 draft statement found that the retained alternatives would cause high impacts to historic and private properties, and those impacts, together with a lack of public and political support, led the agency to rescind its notice of intent. Only when the regional transportation commission put the crossing on its priority list in 2014 did the work restart, this time as a supplemental statement prepared in part because environmental regulations and conditions in the region had changed substantially in the fifteen years since the first decision. A second record of decision arrived on 12 June 2017.
Eight lanes of water crossing, four of them new
The scope is easier to read once the trick in it is visible. Twin new tunnels carrying two lanes each are being bored just west of the existing crossing, giving eight lanes of capacity across the water in total. The new facility takes eastbound general-purpose and high occupancy toll traffic, and the existing eastbound tunnel is converted to carry westbound tolled traffic. The project also replaces the trestles that connect the tunnels to land, expands both of the artificial islands, and widens the four-lane landside sections in Hampton and Norfolk to six full-time lanes, four of them free general-purpose lanes and two variably priced, plus two part-time shoulder lanes that open under heavy congestion. Accommodating that widening means rehabilitating or rebuilding 27 interstate bridge structures along a corridor of nearly ten miles.
Delivery is design-build under Virginia’s public-private legislation, with no private capital in the deal. The department’s finding of public interest concluded that the size and complexity of the work justified that route, and the procurement produced a comprehensive agreement in April 2019 with Hampton Roads Connector Partners. The trade-offs behind that choice belong to project delivery methods rather than to this crossing alone.
The tunnelling method was a competition, not a decision
One line in the plan tends to get passed over. The commissioner’s finding permitted proposals under either an immersed tube or a bored tunnel methodology. Both were treated as credible here, which is itself informative: an immersed tube would have been trenched into the harbour floor across a working shipping approach, while a bored tunnel keeps the work underground and the water open.
The bored option won, and it produced the first bored roadway tunnel in Virginia’s history. The machine, named Mary, stands 46 feet tall, runs more than 430 feet long and weighs over 4,700 tons. It was built in Germany, shipped in 170 components, and advanced at rates of up to 50 feet a day. Each tunnel is about 8,000 feet long, its deepest segments roughly 150 feet below the water surface and about 50 feet below the existing tunnels. Boring began in April 2023. The machine broke through on the north island on 17 April 2024, turned around, and completed the second drive on 24 September 2025.
At concept level the method is a pressure-balancing exercise. The face of the excavation is held up by pressurised slurry rather than by the ground itself, and the excavated material leaves as slurry to a treatment plant on the surface, which is why the project’s risk register tracks clogging and soil stickiness as cost items rather than as curiosities.
The risk register is the real project description
Virginia’s plan devotes a section to risks and response strategies, and it is the most instructive part of the document. The contract is priced against a geotechnical baseline report, which defines the soil properties the design-builder is entitled to assume. Everything the register lists is a way that reality might depart from that baseline.
Buried obstructions appear repeatedly: rock containment dikes, scour protection and steel casings that the baseline report either does not locate correctly or does not describe at all, any of which can interfere with support-of-excavation walls, approach structures, ground improvement or the bore itself. Approach excavations carry the risk of basal instability and heavy groundwater inflow, mitigated by mandatory dewatering and water-tight walls with enough toe embedment. The points where the machine breaks out of its launch and reception structures are singled out for flowing soils, and the register notes that the north island was the less favorable of the two because no seal was provided at the breakout. Ground gas gets its own entry, with ventilation and monitoring required by federal safety regulation.
This is the level at which a megaproject is actually specified. A reader who wants to know what a project is buying will learn more from twenty rows of a risk table than from any rendering.
The ledger, and the two conventions behind it
A federal cost estimate review workshop for the project ran on 5 and 6 November 2018 and was updated on 12 December 2018 after further analysis. The workshop modelled 37 risk items; the final model carried 38, comprising 34 threats and 4 opportunities. Federal guidance requires the result at the 70th percentile in year-of-expenditure dollars, and the review produced a design-build contract cost of $3,282,000,000, a departmental cost of $524,000,000 and an overall project cost of $3,784,000,000.
Both conventions deserve unpacking. Year-of-expenditure means each cost is stated in the dollars of the year it will be spent, so escalation is inside the number rather than a footnote to it. The 70th percentile means the estimate is a confidence statement about a distribution, not a price: there is roughly a three-in-ten chance the project finishes above it. Announcements of megaproject costs almost never say either thing.
The actual bid came in at $3,299,997,227 of maximum compensation, 0.55 percent above what the review had modelled for the contract. Stacking the $335,000,000 construction contingency and the $90,000,000 of incentives on top brings the construction line to $3,724,997,227, and the total project cost stands at $3,965,451,641, unchanged from the previous year’s update. A separate bridge repair option of $73,454,414 is funded from maintenance allocations rather than from construction money. Ninety-one work orders have been executed for a cumulative $17,461,712, drawn from contingency.
The commodity clause is where the market risk actually sits. The original contract allowed price adjustment for asphalt, fuel and steel. After the first amendment to the comprehensive agreement on 7 February 2024, adjustment was extended to tunnel and marine materials because of what the plan calls the hyper-inflation experienced in those areas. Adjustments have reached $130,346,188 against a contractual maximum of $373,119,941, all of it paid from contingency. Funding comes overwhelmingly from regional sources, with a federal share of $303,222,543 that now includes PROTECT formula money added when state board funding was reduced. That a resilience formula program is quietly financing part of a harbour crossing is a small signal of a larger shift, one traced in sea level rise and coastal highways.
The sentence a department did not have to publish
Under the first amendment the contractual final completion date is 27 August 2027. The plan then says something rarely found in project communications. The design-builder has struggled to meet projected production rates and near-term completion dates, consistently across marine, bridge, roadway and tunnel outfitting work, and has responded by compressing future activities in its monthly schedule updates to preserve the delivery date. The department records that it has raised its concern about the contractor’s ability to hit those rates at project and executive level.
That paragraph exists because federal law requires the filing and federal guidance requires the comparison against the baseline. The regime that compels it was built after a different tunnel project in a different harbour city stopped reporting reliably, a history traced in the Big Dig case study. The cost figures in more on major projects come out of filings that regime compels.
The boring is finished and the machine has been taken apart. What is left is fit-out, systems and the calendar, and on the department’s own account those are the items running behind.