Half a Mile, $23 Million: The Math on Glydways’ Cost Claims Doesn’t Add Up
Glydways raised $170M touting 90% cost savings over rail. But its Atlanta half-mile pilot is running $36-46M per mile — right at the light rail average. The full contract terms have never been made public.
Half a Mile, $23 Million: The Math on Glydways' Cost Claims Doesn't Add Up
Glydways says it can build transit for 90 percent less than rail. Its own flagship American project is about to test that claim with a half-mile pilot in Atlanta costing up to $23 million. The numbers do not cooperate.
The company, which announced a $170 million Series C this week, has been explicit about its core value proposition: cut infrastructure costs by up to 90 percent compared to conventional rail. That is the sentence on every press release, the line Vinod Khosla used to justify his board seat, the reason Sam Altman put it in his Series B check. But the Atlanta project, Glydways' most advanced American deployment, is already producing figures that complicate the pitch.
Construction Review Online, a trade publication that covered the Atlanta groundbreaking, puts the pilot cost at approximately $18 million. A BriefGlance article describes the same public-private partnership as costing more than $23 million. Both numbers cover the same 0.5-mile guideway. Split the difference and you get roughly $41 million per mile. The North American average for light rail construction is $35 million per mile, according to a survey of built systems. Glydways is not building light rail. It is building small autonomous pods on a dedicated elevated track. But the cost per mile is landing in the same neighborhood as the technology it claims to undercut by an order of magnitude.
The company did not respond to a request for the Atlanta pilot budget breakdown.
The 90 percent figure comes from a cost-per-passenger-mile comparison that Glydways and its advocates have circulated widely. The company's website claims its system runs $0.35 per passenger mile versus $4.49 for light rail and $1.14 for bus rapid transit. That calculation lives in a document published by the Advanced Transit Association, an industry advocacy group. It is not an independent audit. Glydways cites it the way a restaurant cites a review it wrote itself.
Chris Van Eyken, a transit analyst at Transit Center who has studied personal rapid transit systems, is not persuaded by the capacity math either. Glydways says its system can move 10,000 people per hour per lane. Each pod carries four to six passengers. That requires a pod to depart every 1.4 seconds, Van Eyken told Streetsblog USA in 2023, and boarding times make that target unrealistic at any station with more than a handful of users. "PRT would have to run at incredible frequencies to ensure no waiting," he said. The 2026 pilots in Atlanta, the UAE, and the New York area will be the first real test of whether the frequency problem has been solved in software.
The 2023 memorandum of agreement that launched the Atlanta project was signed among MARTA, Fulton County, Clayton County, the City of College Park, and two airport-area community improvement districts. The formal public-private partnership contract, the document that would lay out who pays what and when, has not been made publicly available. The governance structure exists. The detailed financial terms do not appear in any public filing type0 could locate.
Glydways is not a small bet. The Series C, co-led by Suzuki Motor Corporation, ACS Group, and Khosla Ventures, brought total capital raised to over $250 million. Bloomberg separately reported that Glydways is already in discussions to raise another $250 million at a valuation above $1 billion. The company operates eight offices globally. It has been building toward this moment since 2016.
The pods themselves are modest in scale. They travel at a maximum speed of 31 mph on a dedicated 2-meter-wide guideway, according to BriefGlance. The Atlanta pilot is scheduled to open in December 2026 as a free, on-demand service. Free is an attractive price. But someone is paying $18 million to $23 million for half a mile of infrastructure, and the taxpayers in Fulton and Clayton counties have not seen the line item.
Transit projects routinely run over budget. Light rail systems routinely cost more than their initial estimates. The pattern Glydways is asking cities to bet on is that its technology breaks that pattern permanently, not just on the first half-mile. The 2026 pilots will show whether the pods can run at frequency, whether the maintenance costs stay low, and whether a 31 mph pod network makes sense for distances longer than an airport terminal loop. What they will also show, quietly, is whether the Atlanta numbers were a genuine early-stage efficiency or a demonstration project premium that the company will never quite escape.
The promise of cheap transit infrastructure is not unique to Glydways. Personal rapid transit has been sold to American cities as the solution to cost overruns since the 1970s. Every few years the technology gets another hearing. This time it has autonomous driving software, sleek pods, and $250 million in the bank. The 2026 pilots will be the most serious evidence yet about whether that is enough.
