Creating Muni’s First True Rapid Subway Line

Putting the Metro in Muni Metro

A Flawed Model

One of the fundamental flaws of the Metro is having a single track on to which all 6 Muni Metro lines funnel. The tunnel is over-capacity at peak hours, restricting service and capacity increases below Market Street—and throughout San Francisco.

While the subway is automated and can theoretically support below 2-minute spacing, trains also run above ground competing with traffic, stop signs, and driver variability, and then attempt to merge underground. Plus, a single delay underground can wreak havoc across the entire city.

The result? Traffic, delays, and overcrowding underground...and everywhere else.

Introducing M-Market

Muni’s first true urban subway.

Let’s remove the variability: no traffic, no stop lights, just pure automated subways running below Market Street. With longer 4 or 5 car trains running every 2 minutes (or less), we quadruple the capacity of the Muni Metro.

Let’s have a real urban subway every two minutes down Market Street, all day.

The Benefits

The M-Market will drastically reduce congestion and delays, while setting the stage for bigger and better things. Despite introducing potential transfers for certain routes, the maximum wait for a subway becomes 2 minutes.

The efficiency improvements underground also increase the number of trains on the remaining surface level lines while keeping the same number of trains in the system, reducing the time between trains above ground by over a third.


The net result is a decrease in commute time of almost 30% for the major rail corridors today, decreasing the time spent waiting at stops throughout the system, more reliable systemwide service, and—most importantly—the stage is set to take Muni into the 21st century.

This simple step, which we can take as soon as today, is the first shift toward a citywide true rapid transit network that goes beyond commuting and brings world-class urban transit to San Francisco.

Subways That Act Like Subways

M-Market Metro trains can be true subway cars with high-platform boarding, wide efficient doors, and smooth but rapid acceleration.

A Smoother, More Comfortable Ride

Above-ground trains can be low-floor trams to facilitate boarding, improve accessibility, minimize mechanical failure, and decrease weight.

Low-Floor Trams

Trains running on street-level can switch to low-floor trams. These provide better accessibility, improve boarding times, increase vehicle options, and are lighter-weight. There are fewer mechanical parts to fail and no need for platform infrastructure.


Underground Subway

Trains on the M-Market Metro line can be multi-car length single or dual-cab subway trains with high-platform boarding, wide and efficient doors to minimize time spent at stations, and improved automation and reliability options.


Step 1: Simplify to 5 Lines

The N-Judah and J-Church terminate at Church Station, while the J-Church takes over the the old K-Ingleside from Balboa Park to West Portal. The L-Taraval extends along the old M-Oceanview route from Balboa Park to West Portal.

Step 2: Network Enhancements

The beginning of a true transit network. By bringing the J-Church underground (cut-and-cover) along Church St. between Dolores Park and Market, we have an underground transfer at Church, with continuing service to Lower Haight.

The N-Judah can be brought underground at the eastern edge of the Sunset Tunnel, keeping the N underground from Carl and Cole through Church, with an underground transfer to the J & M. Continuing across Market, the N comes above ground and continues eastward along Duboce.

Step 3: Building on Existing Plans

When you add in current plans for Van Ness, Central Subway, E-Embarcadero, and the (beleaguered) Geary Line, a true transit network begins to form.

By continuing the N-Judah east to the new Showplace Square District and southern SoMa, it connects to Caltrain and T-Third at 4th & King. The J-Church continues north along Webster to meet the Geary Rapid Line.

Step 4: Going Beyond

Continuing this vision beyond, you can start to see the potential benefit to rapid transit in San Francisco when radically altering how we think about the role of the Muni Metro.

The J continues up to Fort Mason on the east side and begins terminating at Balboa Park once more. A new K line takes over the western side of the J and extends up 19th Avenue to the N-Judah and Geary. The Van Ness Rapid Line continues down Mission to Balboa Park. The Central Subway continues to Aquatic Park and the Marina.

San Francisco has easily the second highest per capita transit ridership in the country after New York, but if you've sat and watched train after train pass by without room to fit, you know we're not even serving the capacity of today, to say nothing of planning for tomorrow. San Franciscans continually say, "We want more transit!"

We're adding thousands of new housing units to the market, many with limited parking spaces due to lower demand for cars, but our system itself—not just funding— prevents us from adding any meaningful new transit capacity. It's time to think outside of the current model and build a true network of rapid transit throughout the city.

The latent demand for transit in San Francisco is at an all-time high and growing, but we have no means of handling the capacity we need, and we're artificially keeping transit ridership down by providing inadequate service.

Full Vision

The M-Market proposal is a first step in a series of recommendations by Nextransit for the growth and expansion of the San Francisco Muni Metro system. This small step has fundamental effects on how we think about a metro transit network and how we design for the next generation of the Muni Metro.

To learn more, we'll be releasing our Full Muni Metro Vision and our Market Street Bus Plan at the site. The former describes how we envision the city growing into a true world-class city network, one that would be the envy of the world over. The latter looks at the unique challenges and opportunities afforded to buses by both the M-Market vision as well as the Better Market Street project that is part of the city's efforts to improve the livability, safety, and vibrancy of Market Street.


I love this! How can we get started?

We want to switch the conversation from complaining about Muni to discussing ideas and solutions, and then pushing our leaders enact the changes. We'd love to gather a group of folks who are interested in talking about how to take Muni from trying to pick up the pieces into a system that is worthy of envy. Please sign up for our mailing list on the main page, and we'll reach out for gatherings or meetings. We're also on twitter at @nextransit and would love to hear your thoughts.

I hate this. Muni should just <insert Muni complaint>!

This isn't really a question, but it's a somewhat valid observation. Yes, Muni has a lot of issues, but sitting around complaining about them doesn't get us anywhere. (Believe us, there have been plenty of decades of that.) Until we start fundamentally re-evaluating some of the premises Muni is built on, we're stuck with some pretty daunting issues.

Let's imagine a world where Muni vehicles don't break down and maintenance is fully funded, a world where Muni can afford twice as many Metro drivers, and owns a much larger fleet of trains. This still won't solve the basic problems we face today, though it's definitely a given that proper maintenance would have a noticeable impact on our existing system (please vote YES on Props A and B this fall!) With 6 lines converging onto a single track, even with 1.5-Minute headways, each of those 6 lines could not be at frequencies better than every 9 minutes or we're stuck with major traffic issues underground. If a single incident happens underground, going from Castro to Noe Valley on the J or Stonestown to Balboa Park on the M are both now nightmares as well. Our above ground stations can't handle 3 or 4-car trains (some lines can't handle over 1 car!) which remains a limiting factor underground as well, severely restricting our ability to improve capacity.

Unless we rethink our network, even in best-case scenarios, we'll still have low frequency trains above ground, system-wide meltdowns periodically, and overcrowded trains throughout. Remember, that's best-case if everything else at Muni is magically fixed.

We think it's time to re-evaluate this broken model.

How does the connection at Church and Duboce work?

We view this transfer in stages. If we're up for it, we can simply have an above-ground transfer at the current J-Church stops along Church Street. Or, to facilitate this transfer, we could build a short pedestrian tunnel to add an additional exit right to the platforms on both sides of Market. Currently, the transfer speed is about 1 minute from platform to platform northbound, 1.5 minutes southbound. Extending the entrances to the platform (and adding additional east-side underground staircases and faregates to the platform) could cut this time down to 30 seconds. Even today, with no infrastructure improvements, if a K, L, or M will arrive within 4 minutes, it's faster to disembark the J at Church, and walk down to the underground platform than to wait for the J to merge in.

Either way, ultimately we view bringing the J and the N underground as essential to maximizing the transferability between lines, as well as to speed up trains between Sunset Tunnel and Church Station on the N and Dolores Park and Church Station on the J.

The most simple option is the cut-and-cover model to bring the J-Church underground just after Dolores Park. The hill is steep along Church, and the train could enter a flat tunnel immediately north of 18th Street while the road climbs upward. When done effectively, cut-and-cover subway models have been built for less than $30M/mile of track, and we're talking less than a half mile here. Another option is, should the city succeed in working with Safeway to redevelop the triangle between Duboce, Church, and Market, any new development (as well as the grocery) could receive direct subway exit access in exchange for allowing the J church to veer north along Webster from Church as part of the infrastructure construction. This could set the stage for an underground tunnel from Church Station all the way to Hayes and Webster (with an underground stop at Haight & Webster), where the J could emerge as a surface-level tram on the rather large Webster Street.

For the N-Judah, we see the current eastern edge of the Sunset Tunnel as a place to descend the N just below the surface toward Church Station. A relocated Duboce Park station could be partially or fully underground just west of the current station. As the train intersects with Church (where the J would cross perpendicularly), underground transfers are in order, with moving walkways to express to the M-Market platform. Continuing just below the surface, the N would head east along Duboce just past Market before potentially emerging as a surface-level tram along Duboce to Mission Street.

How does the connection at West Portal work?

We view this transfer in stages. In stage one, you simply add a platform along the side where passengers can alight and walk to West Portal Station. Ultimately, having platform-platform transfers is key to making this work well.

I hate transfers! Why not just improve what we have?

If you've sat and watched train after train pass you by without room to fit, you know we're not even serving the capacity of today, to say nothing of planning for tomorrow. Maybe you've had to disembark a train that's become disabled? Kicked off a train at Church and Duboce because the N-Judah is turning into a J-Church? Our tunnel is over-capacity downtown already with literally no way to significantly increase that capacity. Even if Muni had the budget, we can't run trains more frequently above ground because of the traffic jams it would exacerbate underground. As we grow as a city, we need to increase the capacity and efficiency of transit under Market Street, but also throughout San Francisco.

Transfers are an essential part of any good transit network, not a single-purpose one. One only needs to look to New York or Montreal in North America, or cities like Tokyo, London, Paris, or Moscow internationally to see that transfers are precisely what allow a city's transit system to grow into a comprehensive network.

What makes transfers work, however, is very frequent service and the ease of transferring. Transferring from, say, the 38 Geary to the 24 Divisadero can be an exercise in inner zen, waiting for up to 20 minutes. The trick here is frequency: The M-Market would run every 2 minutes, making the worst-case scenario a two minute wait. By contrast, it takes the N-Judah an average of 4 minutes just to enter the tunnel from Church and Duboce, or 6 minutes for the J-Church from Church and Market.

If you really want, you can try it today: disembark the J at Church and Market, enter Church Street Station, catch an inbound K, L, or M, and if one arrives within a few minutes, you will handily beat the very same J to Van Ness. And that's with an incredibly inefficient transfer.

Even accounting for added time of transfers, based on our model, between the 30-40% increase in vehicle traffic above ground and the frequent underground capacity service, riders stand to save anywhere from 5-40% of the time it takes today to get downtown. Saving time, much improved reliability, increased capacity, and a broader network are the benefits which drastically outweigh the negatives of a transfer.

How much would this cost?

If we talk about it in phases, here's a rough overview:

Phase 1 : Low Cost
On one hand, we could literally start phase 1 today for essentially no cost. Ideally, however, we would extend the Church Station entrances to the Muni above ground stops, so there's a quick transfer option. This would involve 2 hallway extensions, with 2 optional staircases and faregate construction to shorten platform distance. At West Portal, the ideal would involve a construction of a second boarding platform for rapid transfer from/to the L-Taraval. This would involve no additional track work.

Phase 2 : Estimated $230 – $500 million
The extension here would call for underground sections of the J and N through the Church and Market area. Most cheaply, this would happen via cut-and-cover from just south of Market on Church to above ground at Hermann and from Duboce Park to Duboce and Guerrero, with the remaining sections above ground. If we maintain subways the entire route, from Dolores Park along Church to Haight and Duboce Park all the way to 16th & Mission, we reach the high end of the cost range. Potential variabilities include routing the J church north along Webster instead of Fillmore and leveraging any development that occurs on the current Safeway site as combination subway + foundation work, which could be done in conjunction with a site developer.

Phase 3 : Estimated $200 – $600 million
Extending the J Church north to Geary would be cheapest along Webster as an LRT and the N along 16th. The costs of this 2.7-mile extension to each line would run up to $600 million for cut-and-cover subway the entire route. An alternate, which could provide cost reductions, would be the extension of the N along Duboce as part of a redevelopment of the existing raised freeway structure. Should this occur through alternate funding, an N-Judah alignment along Duboce would make for a shorter route, and could include a connection to BART through an additional BART stop between Civic Center and 16th/Mission.

Phase 4 : Estimated $900 million – $3.5 billion
Here we see the greatest cost, 6.5 miles of rail along Geary, 3.4 miles of extension along 19th Ave, 1.5 miles of the J up to the Marina, 1 mile of the E west to Crissy Field, 2.7 miles of the Van Ness rapid line extended to Balboa Park, and 2 miles of extension of the T to Chestnut and Fillmore from North Beach. This is a bigger-term vision, marrying current projects (and existing project visions) from the SFMTA with the network of plans we're discussing here. Phase 4 has a wide variability in cost due to options between LRT and subway.

Total, for the full vision, we're talking a wide range of anywhere from $1.4 billion to upwards of $5 billion. While this is a substantial amount of money, keep in mind that, at the low end, we're talking less than the cost of the Central Subway. That's astounding when one considers how much interconnectivity this vision could provide San Francisco. Everything we're discussing here is rail or subway at a minimum, with the low-end possibility for extending the Van Ness Rapid line as BRT (though we would prefer the rail option.)

Why can't the tech companies pay for this?

We never said they couldn't! (But we're not really talking about it here either...) We're currently evaluating a vision that looks at comprehensive commuter reform in the city, so stay tuned. In the mean time, there are some great funding concepts for better transit funding, which include things like payroll taxes on larger organizations that rely on transit for their employees to get to/from work. We'd love to hear from you and any input you might have on the tech company situation by emailing us at info [ at ]

Is San Francisco really ready for a true transit network?

San Francisco has the second highest per capita transit ridership in the country, but one of the slowest and most crowded transit systems in the country too. If you've sat and watched train after train pass you by without room to fit, you know we're not even serving the capacity of today, to say nothing of planning for tomorrow. San Franciscans continually say, "We want more transit!" Our residents are literally clamoring for a better system. When Vancouver replaced one of their bus lines with a BRT line, they saw ridership increases of almost 15%. When they replaced that line with an automated subway, the ridership increased 700%. We believe the latent demand for transit in San Francisco is at an all-time high and growing, but we have no means of handling the capacity we need, and we're artificially keeping transit ridership down by providing inadequate service.

At the same time, we're adding thousands of new housing units to the market all over the city, many with limited parking spaces due to lower demand for cars, but we're adding no new transit capacity. It's time to think outside of the current model, shake things up, and build a true network of rapid transit throughout the city—the kind of transit we know we deserve.

San Francisco is bus-focused. How does this integrate with that model?

We are strong believers in facilitating inter-modal transfers as the model for better converging these systems. Take a look at our Market Street bus plan, which aims to simplify how we move between modes and maximize the potential for better transfers. Buses are integral to the San Francisco transit network, and will continue to be for the forseable future. While we tend to be strong believers in subways in dense urban environments for speed, capacity, and space usage, the role of buses remains critical, particularly for local service. Better integrating our bus network with our subway and rail networks can help optimize scenarios where time savings can be maximized. For example, the 6-Parnassus takes around 20-30 minutes from Van Ness to its terminus along Market Street. A train underground (without traffic) could make that same journey in as little as 5 minutes. Having a platform-platform transfer to maximize that opportunity for riders who want speed is an important option, while riders who prefer slower service but frequent, accessible stops can ride the bus along Market Street instead. Our rapid network should have a high enough frequency to enable this sort of opportunity, and we should design what we can to facilitate the connection.