Elon Musk’s Tunnel: It Doesn’t Scale, so it Doesn’t Matter

Elon Musk just gave the media a tour of his 1.5 mile prototype tunnel under Los Angeles, which he spent US$10m to build.  Why are Elon Musk’s tunnels so cheap?  Because they’re tiny.

As media photos of the event will show you, the tunnel is just slightly wider than a car.  That means that if you used it for a train, it might have room for one seat per row. I suppose you could fit two if there was no way to move through the train while it was between stations, but that’s almost unimaginable once you add a required emergency exit plan.

So despite Musk’s occasional noises about using his tunnels for public transit, this thing is for moving cars, which means it is for moving trivially tiny numbers of people.

As we’ve discussed before, a car-based tunnel also requires elevators.  You zip your car into a parking space and it descends to the tunnel.  Cool, but have they run the numbers on how many of these they would need, assuming it takes, say, a minute to do a full cycle of the elevator?  How much real estate would it require to get cars into the subway at a rate that even maximizes the tiny capacity of the subway?

Anyway, those are some questions to ask today.

And yes, it would be great if this dalliance produces genuine improvements in tunnel technologies useful for building actual train-sized tunnels that can move the number of people who need to move.  But Musk’s prairie-dog burrows are mostly hype, confusion, and elite projection.  While delivering almost nothing useful, they are confusing elite opinion about whether we still need to build mass transit, which we do.  Is any marginal benefit worth the resulting delay in getting the infrastructure we really need?

Two lessons to remember: 

  • If it doesn’t scale, it doesn’t matter.  The media are easily excited by demonstration projects, but this idea doesn’t scale.  You could build lots of tunnels, and they would each move so few people that they wouldn’t make a dent in a city’s transport needs.
  • If it doesn’t scale, it’s for the rich.  Or to put it another way:  Inefficiency is inequality.  Anything that spends a lot of money to serve small numbers of people raises the question “why are those people so important, and what about everyone else?”

Does this remind you of other transport fantasies, such as replacing transit with “service to your door”?  These rules about scalability are pretty good tests to bring to all the fun new inventions, including whatever’s coming next.

24 Responses to Elon Musk’s Tunnel: It Doesn’t Scale, so it Doesn’t Matter

  1. James Garner December 20, 2018 at 1:02 pm #

    Some things never change. Years ago I was doing some research on fare policies and structures and came across a bit of histories where urban transit was cut and dismantled in order to run highly inefficient suburban express buses. Today, the examples are different (microtransit, Musk’s tunnels) but the results are exactly the same…..small-time, big expense unscalable projects that take large amounts of money away from successful urban transit, and put it in the hands and the pockets of the powerful and the connected.People always are amazed at the byzantine rules that FTA (and no-one else) has. I just say the number of rules are directly related to the number of grifters idling by.

  2. Georgist economist December 20, 2018 at 1:18 pm #

    To be extremely fair to Elon Musk here, the Glasgow Subway does have a tunnel diameter of 11 feet (3.35 m)—and yes, that is extremely limited, but just enough to have two benches lengthwise, facing each other.

    I just hope he doesn’t propose paternoster lifts for the cars!

    • Max Wyss December 20, 2018 at 1:47 pm #

      And the London Underground tube tunnels are around 12 feet in diameter.

      Actually, a paternoster system would probably be the best to get the cars down to the tunnel and back up… note, you read it here first…

      • Henry Chin December 25, 2018 at 1:17 pm #

        These days, you need emergency egress walkways and the like to build a subway. The Jubilee Line tube extension has tunnels 14 feet in diameter. And Crossrail’s using reasonably sized train cars will be 20 feet in diameter.

    • Dave December 28, 2018 at 11:15 am #

      The Glasgow Subway is cute – I’ve ridden it myself as a transit fan – but it also only carries about 39K people a day. A metro area of over 10 million (like LA) will need a transit system that can scale up to carry hundreds of thousands or even millions of passengers a day if it wants to relieve road congestion.

  3. Max Wyss December 20, 2018 at 1:47 pm #

    And the London Underground tube tunnels are around 12 feet in diameter.

    Actually, a paternoster system would probably be the best to get the cars down to the tunnel and back up… note, you read it here first…

    • Max Wyss December 20, 2018 at 1:48 pm #

      note to admin: please delete this message…thanks

    • orulz December 21, 2018 at 8:33 am #

      The idea of Teslas in Tunnels and McMansion Garage Elevators is absolutely ridiculous. No way the business case will pan out. There is no way they will make their throughput targets of 1 vehicle per second at 150 mph. The hubris it takes to make these claims is absolutely mind boggling. (Although we have certainly seen it before, for decades, from various PRT hucksters, although none with a public profile to match that of Musk.) The energy efficiency of traveling at 150mph is not good and will significantly harm the range of vehicles inside these tunnels. Their claims for improvement in tunneling speed and cost are completely untested and nobody should take them seriously until they actually have something to demonstrate.

      The hype over this is way over the top and is entirely driven by the personality cult around Elon Musk.

      Elon Musk himself has some character flaws. He is brash and impatient. Many say that he is difficult to work for and harsh with his employees. His itchy twitter finger is a big problem. Some people treat him like the Messiah when clearly he isn’t. He sets overly ambitious targets for time and scope of what he can deliver and states them like fact. Most of what he does, tunnels included, is just mashing up existing technology in different ways, rather than revolutionary cutting edge research. I would compare him to Steve Jobs in many ways, perhaps even more polarizing.

      There are plenty of gaping holes here, but that said, it is critically important to make rebuttals based on fact, and I think that much of what I have read so far, unfortunately including this blog, don’t quite do that. Mostly I have seen people criticize this based on the fact that this is not new tech (ie, curb guided buses have been around for decades; Paris Metro already runs on tires; this is just a conventional TBM, etc). Those are true statements but to use it as grounds for dismissal is a red herring. The first iPhone used no new technologies and yet it repackaged existing tech into something that was nevertheless new and useful.

      Arguments based on his obsession with using these tunnels for private vehicles are 100% valid and I agree fully. My opinion is that not only is this a social step backwards, the business case for doing this with private cars is doomed to fail. At that point, BoringCo will either shut down, or pivot more toward transit. The fact that their first attempt at building something for revenue service in Chicago will be a closed system serving passenger traffic only shows that they actually do take transit seriously, even if it doesn’t seem like the Chicago Airport Express itself will be full scale grown up transit that operates at a large enough scale to make a dent in actual urban transportation needs. If they can demonstrate the technology and demonstrate that there is a *path* toward scaling, that is enough.

      As for this blog, this is an argument based on tunnel size, which I unfortunately fear misses the mark.

      The tunnels are supposed to be about 14 feet in diameter. Much larger than a car. The pictures show a Model X, which is not a small car, with plenty of room to spare. The roof of the car barely even makes it halfway to the ceiling. You could probably fit a conventional low floor bus, about 8ft wide by 9.5ft high, through this, though just barely. Lower the vehicle’s roof by 6 inches and you have more room to work with. In addition, plenty of existing subways work with smaller tunnels than this, as posters have already mentioned.

      What I do like about this is that they are pushing the limits of the minimum infrastructure that can work for a tunnel in a transportation context. If this looks more like a water tunnel, that’s because it is, and that’s not an insult. They are exploring whether you can:
      -replace wayside current collection with onboard batteries
      -replace conventional wayside signals with autonomous vehicle tech based on cameras, lidar, sonar, and whatever.
      -reduce the amount of ventilation required compared with conventional subways and tunnels
      -I suspect they want to do without safety walkways by making AV tech failsafe enough that the tunnel floor can be used for evacuations
      -build no finished spaces underground whatsoever, by handling all passenger access on the surface (this is their plan but they haven’t demonstrated it yet; if they complete the Chicago project we’ll get a glimpse of what they have in mind and be able to figure out if they have a plan to make it scale)

      Even if the demonstration here is useless they are nonetheless investigating concepts that could definitelty be salvaged into something useful in a transit concept when the initial business model fails.

      At $10 million per mile ($20m for a second tunnel in the other direction) this would be cheaper than rebuilding a surface street to add a median busway. You would need battery buses with guide wheels, and autonomous vehicle technology good enough to detect obstacles in the tunnel and stop in time. Those are almost available off the shelf. I am glad somebody is rethinking whether we really need all the conventional wayside infrastructure inside tunnels anymore.

      • orulz December 21, 2018 at 9:10 am #

        I guess what I’m mostly saying is that I’ve seen a rather disturbing amount of unwavering, unquestioning acceptance and enthusiasm from the unwashed masses (and the media, sadly), and quite a lot of outright dismissal, or even condemnation on moral grounds, from urbanists, but a paucity of level-headed, skeptical but open-minded, critical thinking and analysis.

        That sort of level-headed analysis and commentary is what I have come to expect from this blog. Mr. Walker, you have earned my respect and that of many others by seeking out and engaging with your adversaries in a reasonable manner, including Randall O’Toole. Even on some occasions with those who don’t return the favor, including some of the groups in Dublin.

        While you do raise some good questions, I would like to see you go a little bit deeper. Maybe this is not worth your time, but I leave that decision up to you.

  4. Daan stevens December 20, 2018 at 3:49 pm #

    “This is, by definition, a way to serve very few people at very high cost, compared to fixed routes“ not by definition, somewhere there is a turning point between viable routes and non viable routes. On demand flexibel realtime microtransit should beat Classic public transport with fixed lines ( meaning fixed costs/losses)

    • RossB December 20, 2018 at 5:01 pm #

      >> somewhere there is a turning point between viable routes and non viable routes.

      Right, and that turning point moves towards more transit in more places as both are automated. Places that currently only make sense for microtransit become appropriate for regular, fixed route, fixed frequency transit. That would allow those cities to expand microtransit to areas that aren’t even appropriate for that right now. Of course that means that a much higher percentage of people would use regular transit.

  5. JJJ December 21, 2018 at 7:49 am #

    Jarrett, of course it scales! According to Musk, the answer to every problem pointed out has been to simply build more. One tunnel not enough capacity? Build five! The elevator takes too long? Build 20!

    Sure, sure, that completely destroys any cost advantage, and yes, you could always do the same with any real transit line by building a parallel route, but clearly the man has it all figured out.

    When your only tool is a TBM, every solution involves digging more holes.

  6. Jack December 21, 2018 at 9:07 am #

    I like your work, Jarrett, but I think you are being preemptively disparaging on this one. As noted, this is a demonstration project that has been used to test tunnel boring technologies and test out concepts. Musk has demonstrated previously that he is very willing to change his plans based on what works better (how many design concepts has he gone through for rockets?). The basic premise of reducing the cost to create tunnels that better open up the vertical dimension for multi-modal transportation is a good premise, even by your geometry argument. There’s nothing wrong with modes of varying capacities and costs to use. There’s nothing wrong with innovations that aren’t equally accessible to everyone financially. Putting that requirement on all innovation would stop innovation. This all just has the feel of criticizing the Wright Brothers because their first airplane couldn’t fly far enough to be useful or wasn’t accessible for transportation to people of all income levels.

    • Jarrett Walker December 21, 2018 at 9:52 am #

      Jack. Wright Brothers is not a good analogy because i’m obviously not criticizing tunneling. With Elon, the small size of the tunnel IS the innovation. If Elon were to scale his idea up to accommodate trains that use the space efficiently, he’d have — a subway! He wouldn’t have invented anything.

      And as Elon never stops reminding us, a key idea of his invention is these little car elevators, which definitely do not scale in an urban context.

      I acknowledged that if the Boring Company does something to improve the cost of digging tunnels big enough for transit, they’ll deserve credit for that.

      • orulz December 21, 2018 at 10:13 am #

        Notable other than the size of the tunnels (which aren’t quite as uselessly small as you suggest) is what they have left out of them: namely, power systems, wayside signal systems, complicated ventilation, safety walkways, and even underground stations (which they claim they won’t need.) If modern technology has made it so these things are no longer as necessary as they used to be, and the infrastructure can be entirely passive (or much more passive than it is today) then that, to me, is a worthwhile development.

        Have they proven any of this yet? Presumably they’re close to showing that batteries will work, and that AV tech can replace signals, at some scale anyway. Regarding ventilation and walkways, I suspect they have a ways to go before they can convince regulators of this, but that’s the obstacle, not technology. As for the stations and how all that will work, we’ve only seen renderings so far so it’s a big fat question mark.

        It won’t be Siemens, Alstom, etc. (signal, power, and rolling stock vendor) or Parsons, etc (engineering and construction contractor) to tell us these things aren’t needed anymore. They have a profitable business building and selling this stuff, at very high cost (even in countries where costs are reasonable.) And they benefit if people keep on thinking that it’s all needed, and in fact they have benefitted as the requirements have gotten more and more complicated over time. In a way, this transit-industrial complex is a more insidious enemy to public transit than conservatives who make values-based arguments against it. I will be overjoyed if the Boring Company can shake this world up, even just a little bit.

        Say what you will about Elon Musk, but SpaceX and to a lesser degree Tesla have both had that effect on their respective industries. This seems to me like more of a longshot, but sometimes even a longshot is worth taking.

        Don’t cancel your subway extensions yet, folks. But keep watching this space.

      • Jack December 21, 2018 at 10:33 am #

        Orulz, agreed completely.

        Jarrett, certainly no analogy is perfect. My general point is that this is too quick a rush to judgment when I think there is room for genuinely useful changes in how things are done to come out of this. Any final product probably won’t look just like what he’s pitching now, but it will probably also be different than the traditional subway. Let’s see where it goes.

        The elevators are clearly a weak point in capacity. I’m sure anyone who led a company to space has enough analytical skills to see the problem and think through solutions to it. I’m interested to see what he comes up with. Or perhaps someone else will pick up where he gets stuck/sidetracked and find a different application for what he’s started.

        • Dave December 28, 2018 at 11:23 am #

          A friend of mine is on the waiting list for a Tesla. He says similar things. “The autopilot software is clearly a weak point in the safe operation of the car. I’m sure anyone who led a company to space has enough analytical skills to see the problem and think through solutions to it.” I imagine every huckster in American history has had similar true believers loyally and uncritically hang on their every word and promise too. Just because someone accomplishes something in one field doesn’t mean they are a genius or omnipotent in any field.

        • Enron Must January 3, 2019 at 1:25 pm #

          Jack and Orulz make good, thoughtful points. Fortunately a lot of the debate will be settled/short-circuited by TBC’s claimed next gen boring machine, Linestorm. Either they deliver on the promised 15x speedup or they don’t. If they do, expect the number of media/transport lobby critics to diminish sharply.

          Historical note: we’ve been here before… https://www.telegraph.co.uk/news/uknews/road-and-rail-transport/9790918/They-said-the-London-Underground-would-never-work.html

          • Sailor Boy January 7, 2019 at 6:07 am #

            Let’s say that the tunnel can be dug 15 times faster: You are still better to put a train in the tunnel then have a Car drive down the tunnel.

  7. Enron Must January 2, 2019 at 4:53 pm #

    >>I acknowledged that if the Boring Company does something to improve the
    >>cost of digging tunnels big enough for transit, they’ll deserve credit for that.

    Even if they fall short, and ‘only’ manage to improve the cost of digging tunnels big enough for sewerage or water mains, won’t they deserve credit for that as well?

    • Sailor Boy January 3, 2019 at 4:24 am #

      They would, but they have a heck of along way to go if they are at $10m/mile!

      • Enron Must January 3, 2019 at 1:21 pm #

        @Sailor Boy: Not sure what you mean? Was talking about major city-scale mains. E.g. the Thames Water ring main came in at around $19m per mile…

        • Sailor Boy January 7, 2019 at 6:19 am #

          You have fallen for Musk’s trick. $10m is the *tunneling* cost of his project. $19m per mile is the *total* cost of the Thames Water project. The Thames Water cost includes land acquisition, decommissioning, joints, access points, service provisions, dumping costs for spoil, engineering services/design, and regulatory and planning approval.

          Actually building a tunnel is very cheap. In Auckland (home for me, though I don’t live there, the government is building a two 3.4km tunnels about 15feet in diameter to turn the commuter network into a metro with through routing and improved capacity and frequency. The actual tunneling cost about US$300m; about 400m of the tunnel is below sea level, next to the sea and the rest goes through basalt, it’s built in a very busy city centre. Even in a super complex environment like that, tunneling costs *including* design, management, and dumping are less than US$50m a kilometer.

          Gaining planning permission, building the stations, relocating street services around the tunnel box, laying tracks/access points/ventilation/traction, reinstating roads, and design/management are the really expensive bits. Musk wants to improve tunneling costs 15 fold from an already very low number achieved through 150 years of incremental improvements in technology.

          Either he will fail, or he will fail and lie about the numbers and he has already lied about the numbers.

          • Enron Must January 8, 2019 at 11:43 am #

            Some free advice! The disparaging rhetoric you wrap around your assertions, such as “Musk’s trick” and “he will lie”, will make readers doubt your objectivity.

            Anyhow, let’s deal with Musk’s dastardly trick shall we? (The fiend!) The Thames Water Ring Main was a fairly old project,begun in the 80s, so probably a poor choice by me anyhow. The Thames Tideway tunnel is newer and has had well-publicized budget woes:

            “The current £4.2 billion estimate consists of £3.2 billion to be undertaken by Bazalgette (£1.9 billion base construction costs, £0.8 billion indirect costs and £0.5 billion contingency), and Thames Water’s enabling works estimated at £1.0 billion.”
            https://www.nao.org.uk/wp-content/uploads/2017/03/Review-of-the-Thames-Tideway-Tunnel.pdf (page 30)

            i.e. base construction costs there are 1.9/4.2=45% of the total project cost. Which is similar to figures I’ve seen in the media. And definitely *not* “very cheap” as you claim.