Skyways for the Suburbs

How Targeted Technology Could Tame the Endless Emergency

Readers of Changing Lanes will be familiar with the Endless Emergency: the fact that public-transit agencies lose money on every trip they provide, and consequently are obliged to depend on public subsidy to operate, a situation shot through with warped incentives, and which places ceilings on how good service can be. The way out of this trap is for operators to make money on each trip. That’s the best way to unwarp the incentives and remove the ceiling. In previous entries, we investigated possible solutions: changing how operators are organized, building more density around transit hubs, and automating vehicles.

Today, let’s investigate a variant on that last theme: building new systems only possible thanks to the advent of automation.

Consider Houston. It's a metropolitan area defined by pockets of significant density—notably downtown Houston and the Medical Center, but also The Woodlands, Pearland, Clear Lake, and Sugar Land—sprawling out in all directions, challenging traditional transit: the polycentric nature of the urban area makes it difficult to build a successful system. The light rail network links downtown Houston with the Medical Center and a few near-downtown neighborhoods, but not the outlying nodes. Surface bus does the rest of the work, and stews in the region's notorious road congestion. Of necessity, highways are everywhere. Interstates I-45, I-69, and I-10 connect here (with I-10 on the west occasionally reaching 26 lanes wide), combined with two concentric ring highways. These link some areas but divide others, creating islands of development that don’t connect.

Sugar Land, part of greater Houston, exemplifies the problem. Once a sugar plantation, then a company town, the city grew from 3,000 residents in 1970 to 110,000 today, making it the sixth-fastest growing city in the United States. Now, with only 4% of buildable land remaining and traffic projected to increase 40% to 60% by 2045, Sugar Land faces a mobility crisis. Buses get caught in traffic, while rail infrastructure wrestles to connect the necessary nodes. More traffic lanes or more surface buses will not help much, as both would only increase road congestion. Conventional solutions struggle.

Enter an unconventional system: the gondola.

Specifically, Sugar Land is considering an automated elevated system, implemented by Swyft Cities, that would connect key local destinations. This approach—combining the gondola with ‘personal public transit’—is a creative solution that, at least in some environments, offers a new tool to tackle the Endless Emergency and make urban transit a success.

Last November I spoke with Jeral Poskey, the founder of Swyft Cities, about my ‘Endless Emergency’ series and how Swyft’s approach might help to address it. But before we explore that conversation, let’s explore the intellectual foundations upon which the firm’s technology rests.

Artist’s impression of a Whoosh gondola. Photo credit Whoosh Hold LP.

Urban Gondolas and Personal Public Transit

Urban gondolas and personal rapid transit are two separate solutions for urban mobility problems that fall outside conventional transit models. Swyft Cities' innovation lies in combining them.

While gondolas as a mobility tool for skiers and other alpine enthusiasts have existed since the early twentieth century, urban gondolas did not appear until much later. The first was the Roosevelt Island Tramway in New York, opened in 1976. Originally intended as a stopgap while subway connections were developed, it demonstrated the viability of urban aerial transit, and in fact continues operating today as a supplement to the subway it presaged. Still, it took another thirty years for the technology to come into its own, with the introduction in 2004 of Medellín, Colombia’s Metrocable. As the world's first fully integrated urban gondola public transit system, Metrocable dramatically improved access to jobs and services for previously-isolated hillside communities. Medellín's success inspired the implementation of similar systems across South America, with major deployments in La Paz, Caracas, Bogotá, and—less successfully—Rio de Janeiro.

Personal rapid transit (PRT) traces its origins to the 1960s, when planner Donn Fichter, in his 1964 book Individualized Automatic Transit and the City, argued that people would only abandon their cars for transit that offered similar flexibility and significantly better end-to-end travel times. To that end, he proposed small automated vehicles running on dedicated elevated guideways, providing nonstop point-to-point service.

Fichter’s insight was sound: personal vehicles offer, on the one hand, privacy and more-or-less direct travel from one’s origin to destination, and on the other, relatively low costs to operate on a per-trip basis.¹ Other urban-mobility modes do not have these advantages: buses and rail transit offer the latter set but not the former, taxis and ridehail the former set but not the latter. A system that offered both would be a significant achievement.

There were attempts in the later 1960s and 1970s to introduce it. As I’ve written before, the automated RT that Vancouver, Detroit, and until recently Toronto used for parts of their transit systems stemmed from an attempt to commercialize PRT. But it never succeeded at scale: the technical limitations of the time meant any automated system had to be rail-based, and building new, elevated, branching rail-based guideways, and procuring appropriate vehicles for them, proved so costly that no large city ever pursued it in a serious way.

What makes Swyft Cities interesting is how it draws lessons from both traditions. From urban gondolas, it takes proven guideway technology and the ability to traverse challenging topography with minimal infrastructure. From PRT, it adopts the focus on direct point-to-point travel and automated operation. But crucially, it avoids the pitfalls that derailed many earlier PRT projects by focusing on specific use cases where its advantages are most compelling.

Learning from Silicon Valley's Mistakes

Jeral Poskey founded Swyft Cities in 2021, but his interest in transport goes back much further than that. In college, he says, “I remember the day reading a magazine and seeing monorail-style, small autonomous vehicles that deliver people on demand and going, ‘that's the future. That's what I want to do.’" He went to Stanford, got a business degree, and started the path that took him to founding Swyft.

That path led through Google, where rather than selling transport solutions, he was interested in evaluating them for licensing or purchase (a role analogous to the one I played later at a different Alphabet company, although he and I never met before last fall). "It's totally different when you're the customer," he explains. As an entrepreneur, "We were always selling a dream and a future vision. But if you're a customer, you want steel and concrete delivered on schedule and on budget." This awareness of what the customer needs has fundamentally shaped Swyft's approach: what matters most are reliable, cost-effective solutions delivered on schedule. "It killed me," says Poskey, describing a different start-up he was involved with for a time, one that raised tens of millions for a different type of automated-mobility solution before discovering basic regulatory barriers that made the idea impossible. "They spent all that money focusing on technology before understanding the real world."

If we take Swyft seriously, then the company’s aspiration must be to solve a real-world problem with an approach that is affordable and feasible. What is the problem, and what is the solution?

Swyft Origins

Swyft got its start with Poskey’s attempt to solve a problem for Google: its transition from suburban campuses to mixed-use developments. “As a large developer building millions of square feet, Google was shifting to wanting not the suburban campuses [it had] always had, but real places where people wanted to live and work.” To attract and retain talent, Google wanted to ensure that its employees could “live nearby, [easily reach] shopping, recreation, restaurants… a whole mixed-use community.”

But this imperative clashed with the NIMBY attitude of the Bay Area: “the transit stations down the peninsula is exactly where the cities don't want you to build, because they want to protect the cute downtowns. So they push all of the new development, which they want for tax purposes, away from transit, away from anything. So you'd have to build these future campuses in a car-dominated way, which kills the economics… you're losing money and space to various forms of parking and roads. The conclusion was, everyone designed these beautiful campuses first, and then asked us to figure out the transportation afterwards.”

And so Project Swyft was born: not to invent anything new, but to find an existing technology that could solve this problem. Specifically, the problem of short-haul travel: at the time, most Google employees lived within ten miles of campus, and most drivers were making trips to work of fewer than five.

Poskey and his team tested a variety of solutions: electric vehicles, monorails, and more. And what they found was that a “cable-based solution offered the most flexibility, the lowest cost, the highest sustainability, the highest attractiveness to riders, and then offered this incredible experience at the same time. We had about eight criteria, and I guess it won on seven of them. It was very close on the eighth. It just became a no-brainer that wow, this can really change things simultaneously.”

Their engineering and development partner built the first prototypes of the system, known as “Whoosh”, but—like another Alphabet exercise in mobility innovation that I could mention—the advent of Covid in 2020 generated so much risk in real estate, and changed Google’s plans for its own development, that implementing ‘Project Swyft’ for Google itself no longer made sense. Notwithstanding that, says Poskey, “this product is far better and lower cost than we had even scoped it out to be. So let's spin it out and take this to every developer, every city, every corner of the world.”

Swyft Solutions

But what is it, exactly?

Swyft offers what I call a PUPPET system: Personal, aUtomated, Point-to-Point, Elevated Transit. PUPPETs recognize that any successful new transit system needs automation, and its own right-of-way, and the ability to take patrons from their starting point directly to their destination without transfers, and ideally by a direct route. The latter three elements cannot be accommodated below ground due to the cost of tunneling, nor at surface level because of competition for space, so they must be accommodated above ground, as part of an elevated system reserved for its own use. Glydwys is another example of a PUPPET currently in active development; another, now defunct, was SkyTran.

Swyft’s approach to offering PUPPET service is to combine PRT principles with the gondola. The infrastructure is simple: elevated cables supported by posts every 300 feet. On these cables run automated pods. The pods are electric, compact, can carry up to five passengers, and can travel up to 30 MPH, faster than typical urban traffic and traditional buses. Unlike ski-lifts or other gondola systems with which you may be familiar, the cable is fixed and the pods move along it. Rail guideways atop the support posts enable turns and line switching. As such, branching is simple, which means that pods don’t have to run on fixed routes, but can travel directly from any point in the system to any other. At the same time, the system can typically move up to 3,000 passengers per hour per direction along each guideway, with higher capacities possible under certain circumstances. That’s comparable to a busy bus corridor, but without any traffic delays. This throughput, combined with point-to-point routing, means that Swyft’s network can function as genuine transit rather than just a niche solution.

Artist’s impression of an implemented Swyft Cities network. Photo credit: Whoosh Hold LP.

Such a system has a variety of advantages:

Dedicated right-of-way. Only Whoosh vehicles will use the cable network, meaning that it is not in competition with other uses. Operating in one’s own right-of-way is the defining feature of higher-order transit.

Point-to-point travel. Swyft fulfills the PRT criteria: it takes riders from any stop on the system directly to another without transfers or intermediate stops, one of the crucial features that rapid transit trips usually lack.

Reliability. While a traditional ski-lift gondola can leave passengers stranded when there's a mechanical issue, Swyft's Whoosh vehicles can operate independently and it has a web rather than a line of stations, meaning that other vehicles can route around a disabled one. And Swyft has tools for reaching and moving a stopped pod in a straightforward fashion.

Sustainability. The system is all-electric, which is the future of everything. Further, it’s competitive on this score with its rivals: Whoosh pods will each use less than half the energy of an electric car and less than a quarter of an electric bus.

Permission from regulators. Swyft already has it, because it builds on established frameworks for airport people movers, ski gondolas, and utility franchise agreements, i.e., the ones that let telecom companies run wires through cities. Swyft is also compliant with international regulations, so deployments outside the USA are equally feasible.

Speed. The system’s regulatory clarity enables rapid deployment. For Swyft’s initial deployment in Sugar Land, the timeline from final approval to revenue service is expected to be just 18–26 months, which is a fraction of what traditional transit infrastructure requires. (For context: construction on Toronto’s Eglinton LRT began in 2011 and is not yet complete.)

Scalability. Swyft’s system can be built incrementally, starting with a few key nodes and a single connection between them, and expanding incrementally. Whereas traditional rapid-transit infrastructure requires extensive planning and construction to extend the service, Swyft can grow quickly, station by station, as demand and funding allow.

Ground-level stations. Traditional elevated transit systems require expensive vertical circulation, including elevators, escalators, and stairs, to get passengers up to platform level. These elements add significant cost and complexity while creating barriers to accessibility. Swyft is more fortunate: "We can keep our stations and our stops down at ground level," explains Poskey, "so you don't have to have the elevators and escalators that the Vancouver SkyTrain or the L has to have." The system can accommodate both approaches: street-level stations integrated with the public realm, and, where appropriate, direct elevated connections within buildings. This flexibility allows each station to be optimized for its specific context and user needs.

But these potential advantages will require proper execution. The key question is: where would such a system be most valuable?

Where It Might Work

I mentioned above that Poskey said ‘every city’ and ‘every corner of the world’ would benefit from this technology: the high spirits of a determined entrepreneur. For myself, I think that the Swyft approach is worth examining precisely because it does not solve every transit problem, but instead might solve one specific problem very well: how to provide good rapid transit to the North American suburb… the places that relied on greenfield expansion, never made transit plans, and now find themselves nearly built out with mounting traffic problems.

It’s noteworthy that to date, gondolas as a transit solution have mostly been deployed in South America. (Though there are a few in the United States, and they have been proposed in Edmonton and in Paris.) The reason they enjoy such favour in South America is they neatly solve the problem of elevation. While it’s possible to run rail-based transit up steep grades—something the Swiss have been known to do—most rail systems avoid grades steeper than 4%. Similarly, for comfort and safety, roads in hilly or mountainous regions aren’t laid out directly, but along switchback paths. Consequently, traditional rapid transit is stymied in such environments, while gondolas—which can travel in straight lines largely irrespective of grade—are more successful.

Put another way, gondolas excel at creating cheap, reliable, and direct connections where roads and tunnels can’t go.

Most North American suburbs aren’t built on the slopes of mountains, but they do have other barriers that frustrate easy trips within them. One of these is highways: these function as arteries that facilitate some trips… and moats that block all others. Imagine two suburban commercial districts separated by a freeway. They might be less than a mile apart, but traversing them would take 10–20 minutes by car due to road congestion and limited crossing points. Traditional transit has no good solution to this problem: buses get stuck in the same traffic, while rail, with its need to maintain grade, requires stupendously lengthy (and expensive) tunneling. Swyft's elevated system solves the problem of direct connection. A cable-based guideway can span freeways at a fraction of the cost of traditional infrastructure, enabling point-to-point trips that would be impractical by other modes.

Another barrier is development patterns. As we’ve seen, density and rapid transit require each other to succeed. Unfortunately, adding new transit to existing density is often prohibitively expensive; while adding new density to existing transit is often prohibitively difficult, given the strength of NIMBYism. Developers then face an impossible choice: they need density to make projects profitable, but density requires concomitant amounts of parking, which keeps projects from penciling out.

Swyft's approach could break this deadlock in two ways. Firstly, their target cost, of as low as $10 million to $20 million USD per mile, puts projects within reach of private developers and smaller cities. Secondly, by connecting existing commercial nodes with frequent service, it would enable infill development without massive parking structures.

This has the potential for powerful feedback loops:

• Start small, with a few stations, to prove viability

• Private-sector partners can finance new stops and stations for their real-estate projects

• Each new station multiplies connection possibilities, improving system economics…

• …increasing demand for new stations

For this to work, Swyft will have to be able to extend the system quickly and at low cost where there is demand. In this context, low cost means, as noted, $10M–$20M per mile, which Poskey seems confident is possible. As he puts it, "Transit-oriented development is about development being put where the transit is.” But in many places, the development has already happened. “You can't change it. It's there now. You have to build development-oriented transit… The goal isn't to make the suburbs into urban areas, but to serve them better as they exist today.”

The Bus Question

Why not just run buses?

This is the first and most reasonable objection to any new urban-transit technology. Buses are proven, flexible, and relatively cheap. If you need to cross a freeway, build a bridge. If you need more capacity, buy more buses. To do anything else risks falling into the gadgetbahn trap, as the Hyperloop and so many other gimmicks have… not to mention some PUPPETs I’m familiar with. Why complicate things?

The reaction of one senior transit-agency deputy, meeting with Swyft, illuminates the deeper problem. Unhappy from the start, they objected that their agency was already in fiscal crisis. When Swyft explained the goal was to help transit by acting as a connector to get more riders, the response was telling: "Well, if you haven't seen, we lose money on every rider we take. So the more riders we get, the worse our problem gets."

This perfectly encapsulates the fundamental issue. As we have seen, driver wages represent 40% to 70% of bus operating costs. This creates a harsh equation: either run frequent service and lose money, or run infrequent service and lose riders. No amount of bridges or additional vehicles can solve this fundamental problem.

Swyft's automated, elevated system may be able to escape this trap. Because its vehicles will be automated and grade-separated, they can run frequently without incurring unsustainable labor costs or getting stuck in traffic. Combined with lower infrastructure costs and targeted deployment in high-demand corridors, this creates a path to cost recovery that eludes traditional transit operations; the path that, I have argued, is necessary to achieve progress in public transit.

In Sugar Land's vision, the system would connect key destinations like the Town Square and Smart Financial Centre, areas that currently require car trips despite their relative proximity. Rather than replacing the City’s existing transit services, it would complement them by connecting destinations that are too close for efficient bus service but too far to walk.

The integration potential extends beyond buses. At Google, the original vision included connecting to Caltrain stations, extending the reach of regional rail. "From San Francisco to Mountain View [Caltrain was] faster than the car, faster than the bus," Poskey notes, "but that last two-mile connection was killer." This highlights the system's potential to solve the first/last mile problem that plagues many transit systems. Rather than competing with existing transit, it could help make traditional transit more viable by expanding its effective catchment area.

Put simply: Swyft may be a sufficiently-cheap solution that, in the right environments, could escape the subsidy trap… in a way that more buses, with more bridges and more of an urban footprint, never will.

Other Objections

Swyft’s solution, in the right environments, would be better than simply investing in more bus and road infrastructure. But the key word is right. In what situations will Swyft, or other PUPPETs, succeed where conventional transit will not?

Firstly, it needs a strong market across a very few stations to succeed initially. I imagine the first deployment in any city will be just a handful of stations. To succeed sufficiently well to become actual transit rather than a tourist attraction—sufficiently well to attract investors in extensions—it will need to deliver big volumes of riders out of only a few initial stations. That means a collection of origin and destination points that are relatively close together but hard to connect in any other way. I hope Swyft chooses wisely here.

Secondly, it needs to complement existing transit. An obvious candidate for one of these origin/destination points is an existing higher-order station to collect transfer passengers from. To make this work, Swyft will need to have partner agencies willing to embrace this complementary role.

Thirdly, and importantly, it will need to have a community that can accept the visual impact of the system. Miami, Chicago, and Detroit have demonstrated that cities can accept elevated infrastructure, and Toronto seems to get along fine with the clutter of the catenary wires that power its streetcar network. That said, Swyft’s gondolas, cables, and posts won’t be invisible, and there will be some who don’t like it… and some who resent change of any sort. The right city will be one that can accept the costs in return for what Swyft offers.

Finally, the system will need to prove its weather resilience. Any elevated system faces unique challenges, as a moment’s thought will reveal. Swyft will need to demonstrate reliable all-weather operation and have robust protocols for extreme conditions. That said, its fixed guideway would provide an advantage over road-based transit in snow and ice conditions, as the system doesn't depend on road clearing operations. (Houston, of course, is rarely afflicted with snow and ice, but Swyft has broader ambitions than Texas.)

These limitations suggest careful consideration of where and how it's deployed. In many cities, traditional methods will serve better than what Swyft has to offer. The technology seems best suited for specific kinds of highway-oriented suburbs, rather than as a complete transit solution.

Taming the Emergency

One reason that the Endless Emergency persists is that we try to solve all transit problems at once. We build expensive systems that try to serve everyone, everywhere, equally. When they inevitably lose money, we either cut service or increase subsidies, neither of which solve the fundamental problem, such that it persists… which is why it is ‘endless’.

Small-scale, targeted systems like Swyft offer a different path. By starting with inexpensive capital and operating costs, serving profitable routes and customers whose needs would otherwise go unmet, and growing organically, they can break the cycle of subsidy dependence while still serving real transit needs.

This won't solve every transit problem. But it might solve one important problem well. And in transit, as in most complex systems, progress depends as much on precise improvements rather than grand gestures.

That's how we will solve the Endless Emergency: not all at once, but piece by piece, starting where we can make the biggest differences fastest.

1. Of course ownership of a personal vehicle imposes significant capital costs to purchase and maintain and insure; and often operating costs too, such as fuel. Alas, the nature of human psychology is that these costs are easy to miss. Often, the only per-trip cost that a traveler takes in view is parking, which is free in many places; and tolls, which are uncommon. This feature of our psychology explains much of the disproportionate rage that Manhattan’s new congestion charge has induced. I’ll be writing more about this soon.