Technology and the American Trucking Industry

I have kept an eye on trucking over the last few years. Trucks are not technically cars, but from a semiconductor point of view, they are part of the automotive market segment. We're talking about big trucks here, not pickups or Amazon delivery vans. These are officially known as "class 8 vehicles." 

I have written about trucking in passing before, mainly in the context of autonomous trucks. See my posts:

• O Lord, Won't You Buy Me a Mercedes Benz...Truck
• JP Morgan Report: Dude, Where's My Stuff?
• Level-4 Autonomous Trucks Are on the Road Already

On Trucking

What prompted this blog post is that on Hacker News, I came across a very understated reference among the top few posts to something called "On Trucking." This turned out to be a fascinating document by a guy called Richard Kong called...you guessed it...On Trucking. I highly recommend reading the whole piece, but I will warn you that it is 6,500 words long (our blogging platform estimates it is a 30-minute read). Richard was involved in some sort of tech startup for something to do with trucking, although he doesn't really go into much detail. The whole piece is a brain dump on how the US trucking industry works.

Although these days Richard is in tech, he got his start in trucking as a dispatcher, someone who finds and books loads on behalf of drivers.

First, his definition of trucking is simply the process or business of transporting goods on trucks. It is over 80% of US freight costs and 4X larger than rail, air, and ocean freight combined. A lot of government (at all levels) investment goes into roads, and a lot of that is maintenance. A lot of the reason for that maintenance is trucking since a truck does nearly 4,000X as much damage to the road as a regular automobile.

A lot of shipments of goods are in the 10,000 pound to 45,000 pound range. That's too much for air freight, not enough for rail freight (it won't fill a railcar) but perfect for a truck. And trucks have the flexibility to go anywhere the roads go, and that is basically everywhere. Air freight, ocean freight, and rail freight go where the airports, sea ports, and rails go, which is, in comparison, not many places (and generally, then requires a truck to get the goods to where they are really going). This flexibility compensates for the fact that trucking is not the fastest (air) nor the cheapest (rail).

I grew up in Britain, and there is very little rail freight. By the time you've put the goods on a truck, you might as well drive them to their final destination rather than drive them to any sort of intermodal terminal. Britain is just not that big. London to Glasgow is only 400 miles. Also, Britain (and Europe in general) uses their rails for passenger traffic, which doesn't really mix with freight. By contrast, the US runs its rails for freight, which doesn't really mix with passenger rail (Amtrak) which is why in UK there is little freight except for bulk goods moving at night, and in the US, there is little passenger traffic (except on the East coast where Amtrak has some dedicated lines).

The reason for doing something techy for the trucking market — maybe the Uber of trucking — is that trucking is massive, although just how massive is unclear. It seems to be somewhere between $730 B and $1.3 T. That's a big range, and some of that depends on the boundaries of the industry. It includes not only the obvious trucks and drivers, but fuel stations, freight brokers, truck tire companies, and more. Trucking companies vary from an owner-operator with a single truck, up to large companies (such as Yellow Freight, whose trucks are, confusingly, bright orange).

There are ~3M truck tractors on the road, each doing about ~$340K of transactions per year. Just multiplying those together gets you over a trillion. That's a big number. As Richard says:

It’s natural that every year, founders like me decide to try building a business in trucking.

The industry is pretty fragmented. Richard's numbers are:

Right now, there are 1,946,308 registered motor carriers in the US. Of these, 1,868,399 have fewer than 10 trucks. In other words, 96% of carriers in the US operate fewer than 10 trucks. Another 4% operate 11-100 trucks. Only 5,538 companies have a fleet size greater than 100 trucks, 0.2%. However, of the ~2 million registered carriers, only 10% or ~200,000 carriers are legally active. When a carrier decides to pause or quit trucking, they’ll submit a form to revoke their motor carrier authority...FMCSA will simply mark the authority as inactive.

Paradoxically, given that 96% of carriers operate fewer than 10 trucks, most of the trucks on the road (85%) are part of the very big fleets.

Between the shippers (the people who want stuff trucked, like Walmart) and the carriers (who own or subcontract to the truckers) are brokers. They quote shippers a fixed price and then have to find a way to execute. There are 17,000 brokers, most of them small and regional. But the biggest are very big, pushing 60,000+ loads per week. The total brokerage market is about $100B per year.

Trucks need fuel, of course, and so there are truck stops all over the country. They need a lot of fuel since trucking is not very fuel-efficient, 9X worse than rail for the same tonnage. A truck will purchase 10-15,000 gallons of fuel per year. Some truck stops are one-offs, but a lot are owned by big truck stop companies with as many as 500-700 stops around the country. These businesses are huge. For example, Richard points out:

Truck stops are such great businesses that Pilot Flying J is the 5th largest private company in the US, with $42 billion in revenue annually. Love’s is the 10th largest private company with $26 billion.

How much does a truck cost?

An average class 8 truck costs anywhere from $150k to $250k new. A 53’ trailer costs $50k to $100k. All told, a standard dry van truck costs the carrier around $250k. Most carriers don’t have a cool quarter million just lying around. So, they’ll go and get a commercial truck loan with a payback period of 24-84 months. Interest rates for truck purchases are far above market. Usually, they’re 8-10% APY with a 20% down payment. That means that just the interest payments on the truck are around $15-$20k a year. If we factor in principle, that number becomes somewhere closer to $25-30k per year.

Opportunities

Richard's suggestions of places ripe for tech disruption are:

1. Virtual Carriers
2. Fuel Cards
3. Digital Brokers
4. Load boards
5. Factoring

I won't go into details. Read the whole piece if you want to double-click on any of these for more (or even to find out what they are!). He has a whole section on each.

He doesn't mention autonomy or any of the electronics/semiconductor opportunities. He's focused purely on software. But one thing that I've pointed out before when discussing autonomy and electronics in general, is that it is a lot easier to add a few thousand dollars of electronics to a $250K truck than it is to a $15,000 Chevvy Spark.

Obviously, another big trend (not discussed by Richard) is the move towards trucks with electric traction powered by batteries. The highest visibility is the Tesla Semi, for which production has started. The lead customer is Pepsi. Pepsi is a good match since it does not generally ship soft drinks huge distances coast-to-coast but is more local, so it does not require charging infrastructure everywhere (I'm assuming that the existing chargers are adequate from an electrical point of view, but generally don't have room for such a big vehicle). Tesla claims the truck can go 500 miles on a single charge.

Read the Whole Article

So that gives you an idea of the scale of trucking. There is a lot more detail in On Trucking. Even the footnotes are three pages long!

Plus, if you've not seen it already in the post I linked to at the start of this post, here's a fully autonomous truck driving through a storm:

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.

.