Your browser is not supported. Please update it.

14 October 2021

Rivian puts huge service revenue expectations into IPO document

The Rivian S1 document filed at the SEC at the beginning of September was a first chance to see how well it was doing financially, but Tesla fan-boy publication CleanTechnica this week highlighted the ambition Rivian has for subscriptions, so we took another look at it – and it raises even more question marks over its forthcoming IPO.

This is especially true given the hammering we gave General Motors for saying that it would double revenues in the next decade, partially based on the sale of subscription services, in our lead last week.

On the one hand we would prefer to believe a pure EV play that it will attract extra revenues for software updates and services both sold and delivered over the air to an Electric Vehicle, rather than a “recent convert” to EVs such as GM. But this S1 shows the extent to which it expects to attract such revenues – and for the first time it is out there, available for scrutiny. The diagram below pretty much sums it up.

The largest part of this revenue expectation is the resale – which means that Rivian is expecting to be involved in the vehicle experience throughout its lifetime – that’s not a trivial claim.

The Technology industry is littered with hardware and software suppliers that expected to continue to “own” all the data from a platform and the resell rights, simply because it owned the software platform; and expected to have complete control of related services – in this case insurance and financing.

We go all the way back to IBM mainframes to show than an architectural hardware platform which defined an instruction set, complete with security software etc… is always a great target for innovators once leadership is established. The legendary mainframe wars which IBM won in the 1970s and 80s, from the BUNCH (Burroughs, Univac, NCR, Control Data and Honeywell) was triggered back into life by Fujitsu, Amdahl, Magnusson and Hitachi, building computers which ran the same software – the term plug compatible was born. In all these cases, IBM reacted with illegal account control, got sued for it, and got heavily fined for anti-trust activities. All this did was stretch out the period in which rivals could operate – but with the coming of Unix and subsequently Linux, the entire reason for exerting this control evaporated – smaller more flexible systems came along and then moved to the cloud, replacing mainframes entirely.

The same thing happened with PC operating systems, and today you can download a Linux for your PC and run it with alternative applications. The same for Intel – the 8086 instruction set was set in concrete in 1978, and since that date AMD has made money out of building chips which operate identically to Intel chips, and which could run the same software.

All we are saying is history tells us that the winning platforms will have copycats and that will be part and parcel of what sets the price for subsequent subscriptions – set them too high and you trigger a rival into activity. Could Tesla offer its autonomous driving to Rivian owners in the future or vice versa? Of course. While right now it seems counter intuitive and unlikely – it is a certainty that someone will and if Rivian starts to grow in the same way as Tesla, this will also happen to Rivian.

It’s not that this is so much of a problem and may only lead to the company  losing 10% of its customers to rival services, but it does tend to put a ceiling on the premium that can be charged for services. If companies put technical barriers in the way, as Hewlett Packard has done with printer ink, the outcome is almost always successful anti-trust action.

The best way to prevent a rival copycatting your service is to keep the margin small, and with that in mind it’s worth us looking at what Rivian thinks it can charge. A price of $10,000 has already been charged for autonomous driving software by Tesla, quite successfully. So we think that if Rivian develops software which is at least as good, it will be able to charge the same.

It hopes that it can charge $2,100 for accessories over a ten year period – which is only $210 a year, or $18 a month. While we are not clear what it means by accessories, this is a minor part of its lifecycle revenue, but we assume this is software triggered functional enhancements. However $8,700 for insurance over a vehicle’s 10 year lifetime is considerably more, closer to $870 a year, over $70 a month. Comprehensive car insurance in Europe is nowhere near that high, not with full no-claims discount, but is quite a lot higher in the US on average, some $1,592 a year or $133 a month.

We think it is a given that insurance companies who we have spoken to feel they should charge more for insurance on unproven designs like EVs, and even more still for self-driving cars which are supposed to be significantly safer to drive in than manually driven ones. That’s the insurance companies seeing what they want to see, and it is now fairly standard practice for EV firms to supply their own insurance – but this is a price based on the US market? Does that mean that Rivian does not see a future outside the US market? Surely not. Perhaps the $8,700 over ten years is an average of both US and European prices over that time.

But the truth is that Rivian cannot have it both ways. One of the dilemmas in Europe in particular, is how to make money from service information – brought about by connected cars which sum up the condition of all components and advise on scheduling a service. Again this is not something Europeans have previously paid for – right now almost none of them shell out for this.

If Rivian cannot charge customers for that data, it will sell it to insurance companies – who will then know which drivers drive on their brakes, too close to the car in front and accelerates too aggressively. Insurance firms can then pick off the best drivers to insure cheaply, in the process competing with the in-house insurance supplied by Rivian.

But it’s the $5,500 of software subscriptions that worries us the most – that’s around $550 a year or £46 a month. Now my phone doesn’t cost me $46 a month, and if you add up Netflix, Spotify and Google Maps (which is free) that makes up less than 50% of all the supposed monthly software subscriptions Rivian expects. So on the one hand it wants to be valued like Netflix, on the other it believes it can use its dedicated hardware and software as a “lock in” so it can compete with Netflix and Spotify, at double the prices they charge.  We’re not sure that will really happen – in the end if will partner Netflix in order to sell infotainment to the cars.

Let’s look at the commercial side of the equation. In the past Fleet Management software has had to prove itself in the open market, but Rivian assumes that it will just bulldozer its way into accounts like Amazon, merely because it is supplying the EV platform. Is that realistic?

What it likely CAN do is sell the vehicles themselves as subscriptions, with monthly pricing changed depending upon which online service you need access to. It can put its fleet management software in the cloud and yes it may get picked up by some big organizations – such as Amazon itself.

We conclude that in the US it may well get some of these subscriptions at these price levels during its first few years, but that as it becomes a bigger target to aim at, rivals will single it out for offering rival service offerings. This will drive down margins and increase the number of people who want to do without these services, or want to pay less for them.

If someone else asks for access to telematics data to sell a rival servicing platform and Rivian won’t sell it to them at a reasonable price – there will be legal actions, because after all the company is trying to steal their existing servicing revenues and rivals offering such services will either fight or die. Will it have to separate pout a price, one for the service data and another for actions that are based on it?

The idea that Rivian’s FleetOS will add more features over time, including leasing, financing, insurance, driver safety packages, smart charging and routing to chargers, plus remote diagnostics, and 360° collision reports – surely this data once again belongs to the driver, not Rivian, and they should be able to sign it over. Imagine an EV self-driving company that’s won’t even let you have access to all the cameras on your car during the run up to a collision? You’d dump them pretty fast.  Same for any EV firm which would not let you join a rival app for alerting you to when you needed a recharge, in plenty of time.

This entire idea is like GM saying, “you must buy your petrol, tires and oil from us,” and going into the oil business, bypassing Exxon Mobil. It can start in the tire and oil and petrol business, but it starts from a disadvantage because it has never done it before, and has no reputation for it. We believe that software service sales will be slower to build than everyone thinks, and will involve a smaller premium than everyone thinks and will lead to businesses which compete across the board and across all brands for all of these.

If all of this goes into a vehicle subscription, paid for by a finance house, then what’s to stop the finance house switching to a different telematics firm or different insurance firm for its customers “en masse?”

The other great takeaway from the S1 of Rivian was how quickly it plans to ramp capacity from the 150,000 units annually that it can manage right now. Already it says it has 48,390 R1T and R1S preorders in the US and Canada, and in commercial vehicles it will deliver 100,000 EDVs to Amazon by 2025.

We always thought that Microsoft would have trouble turning itself from a product business which sold operating systems and apps for lifetime use on a PC, into a subscription business, and it took something like 20 years to get to the point where if you don’t like it, it can just show you an ecommerce page, and ignore your complaints, and force you to sign up. But it may also take the entire vehicle industry the same amount of time to do this.

So on the one hand this is great for showing upside to investors and a different way to look at EV makers, but on the other it may end up taking 20 years to stick and even longer to internationalize.

Rivian also said that it expects each vehicle to have a life of 10 years – when cars tend to last 15 years to 20 years across different economies and that a consumer will drive 12,500 miles a year – when the average varies considerably from country to country, and for EVs, they are driven about half the distance of other vehicles currently, closer to 5,500 miles per year. If Rivian has been that fast and loose with all of its assumptions, then expect this IPO to fall flat on its face.

We do however expect it to, at least initially, maintain control of all of its vehicles through trade in and resale, at least until that represents a multi-$billion opportunity to invade, and then expect fireworks and legal action – if it ever gets that far.