Apple Car: Software and Money
by Jean-Louis Gassée
After years of speculation, the Apple Car project boils down to two challenges: How to write world-beating Autonomous Vehicle (AV) software, and how to generate Apple-like Gross Margin dollars making and selling a car.
[I have just completed a full draft of my Grateful Geek book and can now return to the Monday Note habit.]
From its beginning, the still unacknowledged Apple Car has been the source of two pleasant fantasies. First, many of us dream of an electric vehicle that provides a quintessentially Apple user experience. Behold the CarPlay UI that Apple revealed at the June 22 WordWide Developer’s Conference:
Second, company executives and shareholders would love an Apple-sized bite of the auto industry’s $3T (as in Trillion) annual revenue. A successful Apple Car would be an answer to the perennial Next Big Thing question: Where will the next big wave of money come from now that margins from the saturated smartphone market are becoming depleted?
Last month, The Information published Inside Apple’s Eight-Year Struggle to Build a Self-Driving Car, a meticulous and sober summation of the Apple Car project’s long history. You won’t regret taking the time to read this 4,100-word piece by reporter Wane Ma.
My initial, top-level impression of Ma’s portrayal is that the project known as “Titan” has habitually failed to converge. Apple is a famously patient company, but eight years and counting of constantly revolving doors at the top and in the trenches tells us that the project continues to struggle. It isn’t about to enter the reassuring phase where only execution remains between the team and the big prize, the market.
Ma identifies a symptomatic sin that he calls the “Demoware Trap”: Make a great demo designed to impress top execs, and then convincingly assert that the team is close to a Minimum Viable Product. But while a specially-rigged Apple Car prototype successfully negotiating a heavily rehearsed itinerary might convince the blind faithful, it isn’t going to fool a seasoned executive. “Great job guys, we’ll circle back later!” (Having spent decades giving and receiving demos, I sympathize with both sides.)
This must worry some Apple execs. Ma reports that Senior VP Craig Federighi (head of Software Engineering but not for Titan) is openly skeptical. Others have reproached Tim Cook for delaying, wisely if I may interject, the green light that would move the project into the production phase. Given Apple’s reputation and brand loyalty, they believe that the challenges can be handled, even with prices in the unusual-for-Apple $50K range.
Certainly, marketing would be easy. Two words, Apple Car, would provide a crisp and effective positioning statement, even more concise than Steve Jobs’ memorable three-sentence 2007 iPhone proposition: an iPod, a phone, and an Internet navigator.
Contract manufacturing would not want for willing, competent partners. Nor would distribution logistics anchored by Apple’s powerful network of 520 stores and a newer cohort of cherry-picked automotive service providers.
These challenges are hardly out of reach for a well managed, determined, and immensely rich Apple.
Still, one mountain remains: Software, the code that would make the Apple Car a true Autonomous Vehicle.
According to the Society of Automotive Engineers, ADAS (Advanced Driver Assistance Systems) comes in six stages, from Level 0 with nothing more than parking sensors and the like, to the nirvana of Level 5 in which truly Autonomous Vehicles take us from anywhere to anywhere else, under any circumstances, without driver assistance, and at a fraction of today’s road fatalities.
Creating a Level 5 AV is a challenge that the finest minds and deepest wallets of the world have failed to meet. Today’s best attempts have gotten no further than Level 2.
Google was an early AV pioneer and prophet because of its AI prowess and financial resources. More recently, however, Alphabet/Google spun off its AV development into a company called Waymo and pushed it into raising money from outside investors.
With this in mind, I applaud the authorities whom we pay to protect us from charlatans. Specifically, government agencies have finally challenged Tesla’s “Full Self-Driving” promise. The company charges $10K or more for “driving assistance” that certainly isn’t full autonomy. Actually, in conversations with regulators, Tesla has privately admitted that it only offers Level 2 ADAS.
(To be sure, we are in Tesla’s debt for pushing the auto industry into the age of EVs. Regulatory leniency did help the company in its early, struggling years, but its now time for our protectors to do their job.
More context: for the past six years, our household kept two vehicles. My wife Brigitte went through two successive 3-year Tesla S leases and has graciously led me drive her on numerous occasions. My regular drive is a SUV of the German persuasion. I regularly wish someone would combine the best parts of both.)
The real challenge for Apple is jumping from Tesla’s best-selling Level 2 implementation to Level 5 automation — not marketing, not production or distribution.
The rumors sometime go as far as evoking an Apple Car without steering wheel and pedals, passengers sitting around, conversing, or even sleeping on the floor. Great fantasies, but the software to even approach any version of full autonomy is nowhere in sight. This might explain why rumors now refer to a launch date no sooner than 2025.
Then, we must ask why is Apple willing to bet against such long software odds? Why invest in a disordered project that costs as much as a billion dollars each year?
A look at the financial side of the Apple Car challenge gives us a clue.
Initially, one asks why Apple, whose average Gross Margin is in the 54% range (40% for hardware, more than 60% for services), would want to enter a hundred-year-old entrenched industry whose gross profit margins are in the 7% range, climbing to the mid teens for premium brands. But a closer examination reveals an exception: Tesla’s Gross Margin recently jumped from 26.5% to 33%. These numbers explain why its market capitalization often exceeds the combined value of its five largest competitors: Toyota. Volkswagen AG, Daimler Benz, Ford, and GM. It also explains why Elon Musk is one of the world’s richest individuals.
In this light, the Apple Car’s Gross Margin prospects look more favorable. And, unlike Tesla, contract manufacturing would unburden the financial equation from company-owned manufacturing plants, a formula that has worked for all Apple products for more than twenty years. The business model assumes an ample supply of contract manufacturing companies will be willing to help the Cupertino company make a dent in the world of cars (to badly paraphrase Steve Jobs).
We now have a possible explanation for Apple’s enduring effort to make a car against such long software odds. A profitable share of the $3.8T global car industry is well worth the estimated $1B/year it costs to move the Titan project forward. And even if Level 5 automation remains out of reach for the entire auto industry, Apple still might decide to compete using its asset-light, software-heavy business model.
— JLG@mondaynote.com
