Intel read my mind and bought Mobileye for $15.3 billion. I always knew they were smart.
I was looking at the Intel/BMW/Mobileye autonomous vehicle demonstration at Mobile World Congress 2017, and I wondered why Intel just didn’t buy Mobileye outright. They had been doing the groundwork already, with both working together through BMW and Delphi, so they clearly liked each other. Compatibility examples were stacking up: vision is a critical enabling technology for advanced driver assistance systems and autonomous vehicles; Mobileye is already the leader in the space; and automotive manufacturers need a good, reliable, end-to-end solution – from camera to data center and data analysis.
Of course, I didn’t think it would end up being a $15.3 billion deal. Clearly Intel and Mobileye had been working on it long before it crossed my mind at their booth. It’s the biggest deal yet for an Israeli company, and while it’s tempting to call Mobileye a “startup,” it’s far from being so.
Specializing in making sense out of visual data with the goal of making driving safer and eventually autonomous, Mobileye has been focused on automotive safety for 17 years. It is a Tier 2 supplier, working with all Tier 1 automotive suppliers and through those established–and trusted–relationships is either working with or has production deals with 27 car manufacturers, including BMW.
Also, while the average consumer tends to be the first thing people think of when autonomous driving comes up, the big opportunity for safety, efficiency and cost savings is with fleets, from taxis and municipal services, to 18-wheelers. Any margin of improvement in terms of collision avoidance saves lives and lowers overhead costs in terms of vehicle maintenance–as well as insurance.
Mobileye’s 17-year history in automotive safety made it a Tier 1 supplier, making Intel an end-to-end data solutions provider, from sensors to the cloud. (Image source: Mobileye NV)
That said, everyone around the Intel-Mobileye-equipped vehicle stands to benefit too, including pedestrians, cyclists and other motorists.
Mobileye’s expertise is in both hardware and software as it is only by optimizing both together that it can achieve the performance and low power required. Its EyeQ® processor is on its fourth generation (EyeQ4) and is shipping into vehicles this year. EyeQ4 is 8x faster than EyeQ3 and can process data from more than eight camera sensors, in addition to radars and LiDARs. EyeQ5 will ship in 2020. At least that was the plan.
Mobileye’s EyeQ architecture is a melding of proprietary software and hardware optimization for maximum performance at the lowest possible power. (Image source: Mobileye)
The purchase by Intel might accelerate the roadmap, as the deal is structured such that Mobileye’s fast-paced culture will be kept intact. Instead of Mobileye going through the process of being absorbed by Intel, for example, Mobileye will absorb the Intel Automated Driving Group, and be headquartered in Israel. “Combining forces will help accelerate our plans and lower our execution risks,” said Ziv Aviram, Mobileye co-founder, president and CEO to the company’s employees. The other founder is Professor Amnon Shashua, CTO and chairman.
While it has been a supplier to Tier 1 companies, Mobileye has seen itself more as a partner. To that end, it has already moved beyond supplying chips and associated vision software. Along with sensors and sensor fusion, it also includes high-definition mapping through crowd sourcing and driving policy.
Stepping back a bit, the deal is not a big surprise. Intel has made it clear that it sees automotive systems as a critical part of the IoT and a bit part of its own vision of itself as a data-based company. It bought Itseez, a computer vision software company just last year. In a statement to employees, Intel CEO Brian Krzanich could not have been clearer, “We are a DATA company. The businesses we focus on, and deliver solutions to, create, use and analyze massive amounts of data.”
Autonomous vehicles, he predicts, will gush 4 terabytes of data per day. For IoT solutions providers, this is where the combination of Intel and Mobileye makes the most sense. The Intel ecosystem now runs end to end, from the Intel® Xeon processors, FPGAs and 3D Xpoint memory, to edge processors like the Intel® Atom™ Processor E3900 Series and Intel® Joule™ Compute Modules, and 5G modems, right out to cognitive automobiles and other road vehicles. That’s end-to-end processing, connectivity and software compatibility. There’s a caveat though: EyeQ, according to Mobileye, is a completely proprietary design. How compatible it is with x86-based software is yet to be determined.
In the meantime, IoT solution providers should probably start to absorb what it means for Intel to say it’s a data company, while leveraging the expanding hardware and software capabilities of its ecosystem to apply data gathering and analysis technology across all industries, not just automotive.
Many customers may be simply grappling with the real-world problems of just getting their first GPS or temperature and humidity sensors deployed and provisioned on a network simply to get that first byte of insight.