Canceling a big construction job normally means paying to put the ground back the way you found it. It is a real line item with its own budget, and on federal work it gets audited like everything else.<\/p>\n
Then there is Waxahachie, Texas. The Energy Department had 14 miles of finished tunnel sitting under the prairie there, priced out what it would cost to fill the thing in, and worked out that the cheaper move was to let the water take it.<\/p>\n
That tunnel was bored for the Superconducting Super Collider, a proton machine that would have dwarfed anything CERN has ever switched on. Congress killed it in the fall of 1993, after about $2 billion had already gone into the ground. The water went in afterward, deliberately, and it is still down there.<\/p>\n
Which is worth revisiting this summer, because in May the CERN Council finally picked the ring that will succeed the Large Hadron Collider. It is bigger than the one Texas started.<\/p>\n
The SSC was designed as a ring 87.1 kilometers around, or 54.1 miles, looping underneath the town of Waxahachie about 30 miles south of Dallas. Two proton beams would run in opposite directions and meet head-on at 40 trillion electron volts.<\/p>\n
For scale, the LHC ring at CERN is 27 kilometers. Texas was building something more than three times that circumference.<\/p>\n
It never got close. When the money stopped, roughly 20 percent of the project was finished: two dozen kilometers of tunnel drilled, 17 access shafts sunk, and about 18,600 square meters of buildings up, according to Scientific American<\/a>.<\/p>\n The tunnel itself sat around 150 feet down, cut by boring machines running 14-foot heads. Steven Baldauf, the supervising structural engineer on the design team, told the University of Connecticut that the magnets meant to line it “looked like torpedoes, each about 60 feet long,” connected end to end. Something on the order of 10,000 of them were planned for the ring.<\/p>\n Almost none were ever installed. What Texas ended up with was the civil engineering and none of the physics.<\/p>\n Here is the part that never survives the retelling. The popular version says the tunnels were flooded to preserve them, like someone was keeping the powder dry for a revival. The federal paperwork says something duller and more interesting.<\/p>\n In February 1994, the architecture and engineering subcontractor handling site restoration put that job at $110 million. Its plan involved refilling the tunnel.<\/p>\n The Energy Department then replaced that subcontractor, and by mid-March the site restoration estimate had dropped to $25 million. The General Accounting Office audit of the shutdown<\/a> spells out why: technical consultants advised DOE it was “preferable and cheaper to allow the tunnel to fill with water.”<\/p>\n So an $85 million line item disappeared because the government decided to stop fighting groundwater. The shafts got capped. The tunnel filled.<\/p>\n Three decades on, that is still the condition of the site. Shafts sealed, water below, and the surface buildings running as a working chemical plant under Univar Solutions, which absorbed the Waxahachie blender Magnablend at the end of 2012.<\/p>\n The decisive vote came on October 19, 1993, when the House went 264 to 159 against further financing. Sherwood Boehlert, the New York Republican who helped lead the effort to kill it, granted during that debate that the SSC was good science before adding that it “simply is not affordable science,” as recorded by the American Institute of Physics<\/a>. Of the 113 freshmen voting, 81 went against the collider.<\/p>\n The appropriations act that funded the shutdown became law on October 28, 1993. You will see October 21 cited all over the place, including in otherwise careful accounts. It matches neither the House vote nor the enactment date.<\/p>\n Killing the SSC then turned out to be its own construction project, running backward. DOE was handed $640 million to wind the thing down, and with $95 million carried over from the previous year it had $735 million to work with.<\/p>\n The department’s first guess at the cost, given verbally in October 1993, was about $1.1 billion. By February 1994 it had a real estimate of $695 million. By mid-March it was $568 million. GAO’s conclusion was that DOE already had more money than it needed and should not get the extra $180 million it had requested.<\/p>\n The scale of the unwinding is in the details. Of 1,278 subcontracts lined up for termination, all but two had been cut by February 1994. Severance went to 2,595 people across eight contractors, at a total of $78 million, with the prime contractor Universities Research Association accounting for 1,241 of them. Each got 60 days of pay and a $15,000 dislocation allowance.<\/p>\n The two surviving subcontracts are the best small detail in the whole audit. They were kept alive because quitting cost more than finishing. One was a $16,000 job that needed $5,000 of work to complete; terminating it would have run about $10,000 once the legal and administrative costs landed. So the government finished it.<\/p>\n On May 22, 2026, meeting in Budapest, the CERN Council updated the European Strategy for Particle Physics<\/a> and named the electron-positron Future Circular Collider as its preferred next flagship.<\/p>\n The FCC-ee would sit in a new tunnel of about 90.7 kilometers at an average depth of 200 meters, with eight surface sites and four experiments. A firm decision is targeted for 2028, with the machine potentially following the LHC in the 2040s.<\/p>\n They are not the same animal, and the distinction matters. The SSC was a proton-proton collider built for raw energy. The FCC-ee is an electron-positron machine built for precision, a Higgs factory first, with a proton version as a possible later stage in the same tunnel.<\/p>\n What they do share is the hole. And Europe’s hole would be about 3.6 kilometers longer than the one Waxahachie started and never finished.<\/p>\n The US did not respond by proposing another 54-mile ring, and that is not an accident.<\/p>\n In June 2025 the National Academies recommended that the country begin work on a muon collider<\/a>, with the aim of hosting the world’s highest-energy particle collider around the middle of the century. The 2023 federal prioritization panel had already framed it as the American shot.<\/p>\n The appeal is geometric. Muons are roughly 200 times heavier than electrons, which means they lose far less energy to synchrotron radiation going around a bend, which means you can reach very high energy in a much tighter circle. Advocates argue the whole thing would fit inside Fermilab’s existing campus in Illinois, a site of roughly 6,800 acres the government has owned for decades.<\/p>\n Nobody is proposing to buy 54 miles of prairie again. The physics is genuinely unproven, and the engineering problem of accelerating a particle that decays in microseconds is not small. But the land question, which is what sank the last one, has been engineered out of the proposal entirely.<\/p>\n That tension was the whole subject of a panel at the American Physical Society’s particle physics meeting held at Fermilab in July: whether the European ring and the American muon machine are complementary or competing for the same finite pot of money.<\/p>\n Texas has a fair collection of abandoned federal hardware, and the SSC tunnel is not even the loneliest example. Nevada still has an 860-ton boring machine parked outside Yucca Mountain<\/a> because DOE could not find a buyer willing to pay its price.<\/p>\n The difference is that Yucca’s tunnel is dry and you can walk into it. Waxahachie’s is not, by design and by invoice.<\/p>\n The machines that did get built elsewhere put the loss in perspective. Sweden’s spallation source has a five-ton tungsten wheel that has been spinning in a vacuum for years<\/a> waiting on its proton beam, and the largest pulsed superconducting magnet ever wound was stacked in California and shipped to France<\/a>. American industry still builds this class of hardware. It just does not get to keep it.<\/p>\n The tunnel under Ellis County remains the largest hole ever cut in North America for particle physics, and the cheapest thing anyone ever did with it was nothing at all. That $85 million saving is the only line in the entire project that came in under budget.<\/p>\n","protected":false},"excerpt":{"rendered":" Canceling a big construction job normally means paying to put the ground back the way you found it. It is … <\/p>\nFilling the hole back in was the expensive option<\/h2>\n
The shutdown had a payroll of its own<\/h2>\n
Europe picked its ring in May<\/h2>\n
America’s answer is a machine that fits on land it already owns<\/h2>\n
What is actually down there<\/h2>\n