Thread by @asymmetricinfo

Guys, let's talk Tesla! And specifically, Elon Musks's impromptu, maybe-not-so-legal idea to take the company private: washingtonpost.com/blogs/post-par…

Getting some pushback from Tesla fans. But I'm a hardened veteran of the bitcoin wars. I'm ready.

Many, many people are mad about this passage: "taking the company private, ...would give him some breathing room to get better at making cars and developing a battery-charging infrastructure that might actually allow more than a handful of aficionados to buy the things."

Cue the derisive laughter. Don't I *know* that Tesla has a network of charging stations? Don't I *know* that they're already the most popular electric car? Didn't I even, like, Google "Tesla" before I sat down to write?

Friends, we are talking past each other.

(We apologize for the interruption; we had another column to edit. No, we aren't quite sure why we are speaking in the first person, plural. Hopefully, it will pass. Onwards.)

Tesla has a market cap higher than that of General Motors. General motors sells ca 10 million cars a year. Tesla is on track to produce ca 200k cars a year. That is 1/50th of GM's production.

Also, GM makes profits. Tesla doesn't.

(Yes, GM lost a boatload money in 2017 but that was an accounting loss, a weird artifact of tax-law changes): bloomberg.com/view/articles/…

For you to think Tesla is worth this kind of money, you basically have to think it's going to put GM out of business. Or take an equivalent share of business from a broad swathe of automakers.

I mean, maybe! But not proven.

So the auto industry is an industry where extremely high capital costs of manufacturing require a lot of output to break-even at mass-market prices. Functionally, you should either be manufacturing millions of vehicles, or you should be selling into the six-figure luxury market

The six-figure luxury market can obviously support smaller runs. But by the same token, it's not a very big market. It doesn't support Tesla-sized valuations.

Musk has two challenges: to prove that there is demand for electric cars at that scale, and to prove that he can actually make and deliver cars profitably at that scale.

He has done neither.

Making cars at that scale is really, really, hard, and really, really capital intensive. I once stood under a stamping plant the size of a small tract home on the GM floor. Really really cool. Really, really creepy. Really, really, really, really, really, really costly to buy.

But it's not just that you need a lot of cash to buy all these fancy machines. The financial markets have a lot of money! Maybe too much! If you're cool enough, they will give it to you, even if you haven't, you know, actually shown a profit.

It's that as the scale of your operation goes up, your problems scale geometrically rather than linearly.

The best way to think about this is by thinking about getting lunch.

If it's just you, you just, you know, get lunch.

With three friends, you spend some time arguing about where to get lunch.

Neither method works with 150 people. Totally different organizing principle required

That's why there are inflection points as organization size grows--startups have to figure out how to run things when you're too big hash it all out over the water cooler, and medium-sized growing firms have to figure out what to do when even the c-suite can no longer do that

But importantly, all of the methods for making operations work as they grow involve information loss. They have to! The problem you have is that there's too much information coming from too many people, and now no one can hear themselves think. So you silo problems, and info

But as you get more complex, there are more ways for problems to develop without upper management having the least problem they exist. More ways for supply chains to break down, silos to work at cross-purposes with each other, etc Scale is efficient, but difficult to achieve

And this is stuff you should be especially worried about with a CEO who is a superstar genius at the center of all major decisions.

I'm really just touching the surface of all the challenges of scale here, but we all have limited bandwith, so I'll move onto the demand side: how many people want electric cars?

Case in point: me. I am like the ideal use-case for an electric car early adopter. I care about global warming. I have owned a Mini for ten years. In that time, I have put a dainty 10k miles on it.

Just one tiny problem: I live in a rowhouse with no garage. Where do I charge?

There are a lot of folks like me, who might be electric-curious, except for the fact that the very thing that makes us good potential users--low mileage folks living in dense, walkable (read: built before 1929) neighborhoods, also leaves us without garages.

For most people, who regularly go very long distances and have no alternative when they need a car RIGHT NOW, NOT AFTER IT'S DONE CHARGING, electric has to be, say, 95% as convenient as internal combustion.

So it's non-responsive to say that Tesla has a HUGE NETWORK of charging stations compared to other companies running networks of charging stations. They need a HUGE NETWORK OF charging stations compared to other companies that are selling gasoline.

Also, it can't take hours. medium.com/@DonB/how-fast…

It has to take under 10 minutes.

Tesla isn't there. Not clear it's technically possible to get there.

If they can't, that leaves complicated battery-swapping schemes i.e a massive capital investment sitting around unused, because every station must always be able to swap every customer's battery; "come back tomorrow" won't fly. Implying massive excess capacity.

Now, maybe self-driving cars change all this. Maybe. Also, maybe not!

Not clear to me that the future of self-driving cars is the fleet model, though that's an argument for another tweetstorm. Suffice it to say, if it's NOT the fleet model, the charging problem remains.

But even we end up with fleets of electrics, the question remains: when does the self-driving car revolution arrive? I mean true doesn't-have-a-steering-wheel-tell-me-what-it-was-like-to-actually-steer-a-car-grandpa automation.

Before, or after, Tesla runs out of money?

Also, they're not, like, the only people who know how to make electric cars. Maybe they make the best ones! But "best" or "most innovative" doesn't always win the market.

"Knows how to produce efficiently and cheaply" might well win. Like, Toyota, or even boring old GM.

Because here's the thing about a fleet of electric cars owned by operators who deliver them on demand by some uber-like interface: those guys are probably not paying for top-of-the-line design and state-of-the-art features.

(And the features they will pay top dollar for won't necessarily be ones that riders would notice, or car owners will pay for now, like "can be cleaned super quickly and easily").

How much money do you drop on super-cool technology you don't really--let's be honest--need?

If you're the kind of person who puts a deposit on a Tesla, probably a lot. Also, maybe you're the kind of person who complains about the boring old mid-range computers your corporate IT department stuck you with.

(Especially if you're a mac user being forced to use Windows *merely* because the hardware is cheaper and it has a lot of features that make it easier for corporate IT departments to integrate with their servers and manage the workstations)

In other words, if we do develop a self-driving electric car infrastructure that replaces individual ownership of internal combustion engines, I'm not clear that Tesla actually ends up owning that market.

But to circle back to the beginning of this thread: the first thing that Elon Musk has to do is demonstrate that he can produce at scales closer to mass-market automakers than specialty luxury operations. If he can't do this then the rest of the discussion is moot.

And also with the same kind of safety/repair standards as other automakers--it won't do, in this day and age, to be going back to touch up each vehicle by hand for manufacturing faults, and it also won't do for your cars to catch fire much.

One drags up your manufacturing costs and makes you uncompetitive/unprofitable, the other drags up your liability costs, which also scale less linearly with production than you'd think.

But let's say he reduces his waitlist to something reasonable, and everyone loves their Tesla--now he's got to grow the market a lot, to justify current valuations. Really a lot. In a reasonably short timeframe, before people stop being willing to give him money.

That's why I said he needs to prove he can make cars at scale, and find some way to charge them at scale.

And that's the end of this tweetstorm. Thanks for your patience, read the column, which per usual, makes ENTIRELY DIFFERENT points from this thread washingtonpost.com/blogs/post-par…