Qualcomm’s "Automotive Redefined" wasn’t just an event where we witnessed a vision of the future, but where we heard the automakers telling the tech giants to collaborate or go...
Qualcomm’s “Automotive Redefined” wasn’t just an event where we witnessed a vision of the future, but where we heard the automakers telling the tech giants to collaborate or go.
I have only experienced three truly disruptive moments in more than 25 years as a tech analyst. The first was April 2006 when Jean-Anne Booth of Luminary Micro launched the first Arm Cortex-M3 microcontroller. The second was August 2017, when Kevin Tanaka of Seeing Machines gave me a demonstration of their driver monitoring system (DMS). The third was Qualcomm’s event this week. I don’t say this lightly. “Automotive Redefined” was an earthquake.
The Silicon Valley style of “moving fast and breaking things” was simply never going to succeed in automotive. Silicon Valley excels at disruption. AI, apps, compute, datacenter, graphics, the list just goes on and on. But the entire business model fails at mission-critical, functionally-safe system development, of which automotive is one example.
Wasn’t the tech industry supposed to disrupt the auto industry into oblivion? Weren’t those dinosaur automakers and their network of sclerotic Tier 1 suppliers dead meat, destined for a future as subservient “metal bashers” as Silicon Valley’s smartest and finest ruled the world. Then Uber killed Elaine Herzberg and the dream died.
At first glance “Automotive Redefined” looked to be about automotive technology, and it was. But look deeper and it signaled the start of a shift in the power balance between the auto and tech industries. Via Qualcomm, the automakers just gave big tech a gigantic punch in the mouth.
For automotive, Silicon Valley’s macho, braggadocio style of command and control is now deader than a deer in the headlights. The future is what I will call the “San Diego model” of share, embrace, empower. Tech companies must learn to collaborate with the automakers, or go.
Vertical hierarchy systems
You probably don’t expect to read much about the root system of redwood trees in EE Times, but I draw on all kinds of sources, so bear with me. In this blog, the author writes:
So, redwoods do not survive alone…ever. They form “tribes” or communities. Sometimes they grow so close to each other they merge at the base into one tree. The first thing they provide each other is strength and support: intertwining roots. Not deep, but wide, living in an embrace of others.
There is a beautiful description from nature encompassing the broad and rich ecosystem an automaker needs to successfully traverse the technological revolution ahead. The auto industry is embracing technology, not being replaced by technology. That is the distinction which practically everyone has failed to grasp and which Qualcomm just demonstrated it understands precisely.
The biggest problem facing Silicon Valley is the end of the road for what could be called “vertical hierarchy systems.” Intel’s challenges are structural and cultural, not technological. I’ve sat in an airless, windowless meeting room in the Robert Noyce Building (Intel’s HQ). The layout and atmosphere are what my wife would call “designed by men.” It is set up for aggressive, combative “winner-takes-all” debate.
How many decades have now passed since Intel created a truly revolutionary new technology? Where is its skunkworks? Why risk radical innovation or reinvention when you can buy and own whichever company you like? Nokia started life making rubber boots, and Nintendo making playing cards.
A vertical hierarchy system hasn’t worked for Intel in the automotive sector and nor will it work for Nvidia. At both companies, “winning” means power, authority and control, demonstrated by Intel’s ownership of Mobileye and Nvidia’s desire to own Arm. In automotive, power, authority and control rest with the automakers, but Silicon Valley CEOs simply can’t grasp that truth. It is anathema to them.
As we have also witnessed with the transition from the Trump to the Biden administration, the worlds of politics and business are changing. Macho empire building is out, the era of the cult CEO is over and success for any tech company in the automotive sector requires embracing the San Diego model and collaborating with the automakers. This is the new normal.
As we saw this week, Qualcomm doesn’t just have a technological advantage over Intel and Nvidia in automotive, it has a corporate one. Let’s look at how it is different.
Share, embrace, empower
Qualcomm believes having an open platform with qualified software from key partners — rather than spending time developing its own software — is crucial for success in automotive.
Qualcomm’s automotive strategy for the Snapdragon Ride platform is partnership, in particular with Arriver for vision policy and driving policy, with Valeo for parking (Park4U) and with Seeing Machines for driver monitoring. Speaking to EE Times, Nakul Duggal, senior vice president & GM, automotive, at Qualcomm commented:
Our customers, especially in the automotive space, like choices. They don’t like to be locked down into just one solution. That’s the first thing I’ve learned in the automotive business. You watch all the choices your customers have and make sure that you are choice number one.
Which is the very embodiment of the San Diego model. As of this writing, Qualcomm’s market capitalization was just north of $150 billion, compared with about $0.6 billion for Seeing Machines. Why hasn’t Qualcomm already grabbed the driver monitoring minnow from Canberra, Australia all to itself?
Because driver monitoring has a fundamental safety role in multiple transport verticals beyond automotive, such as aviation. Because buying Seeing Machines would up-end its corporate structure at a critical time for the automotive industry. Because Qualcomm views Seeing Machines as an equal.
Just as Arm showed us two decades ago with an IP licensing model for CPU cores, sometimes the most empowering business strategy is for multiple competing parties to share and embrace a technology, not for one party to own it outright. Arm’s main asset was its independence and allowing itself to be purchased by Softbank was one of the greatest acts of corporate irresponsibility I have witnessed in my career.
Let me show you the truly revolutionary new technology that Kevin Tanaka of Seeing Machines showed me in August 2017 and then let’s compare it to the DMS offerings from two companies with vertical hierarchy systems.
This video shows where Seeing Machines’ DMS capability was just over three years ago. In particular, notice how the eye-gaze vector can be used to determine precisely where the driver is looking:
Seeing Machines published this video in collaboration with Qualcomm for the Automotive Redefined event this week. I have screenshotted this frame to demonstrate how the state-of-the-art for the technology has moved on from driver to whole-cabin (occupant) monitoring.
(Source: Seeing Machines)
I invite you to compare that with this video from Nvidia and to draw your own conclusions about technology leadership.
And so on to this slide from Intel, showing that Mobileye has effectively left its automaker customers to work out DMS integration all by themselves. In so doing it has probably handed a sizeable portion of its ADAS business straight to Qualcomm.
What have we learned? That Qualcomm has gone all around the world, found the companies with the best technology and collaborated with them to offer their automaker customers a complete solution.
Collaborate, collaborate, collaborate
Take a close look at the redwoods anytime you drive around Silicon Valley and they will show you the road to success in automotive: Stronger together. Perhaps the tech giants were so busy plotting domination they forgot to engage with the world around them.
Meanwhile, along the pacific coast highway it seems Qualcomm spent the last five years simply listening to the automakers and understanding what they actually want. What is automotive redefined? That’s the San Diego model of share, embrace, empower.