This story is written by [Saman Farid](https://medium.com/@samanfarid), founder of Comet Labs.
----------------------------------------------------------------------------------------------

### **TLDR;**

Intel’s acquisition of Mobileye is a reminder that fundamental transformations to the infrastructure of the auto industry are underway — mostly caused by the drastically increased ability of computing to **perceive, understand,** and **act** on the physical world. Where huge companies previously dominated, new startups are bubbling up that are challenging the way we think of automotive companies. Some of them will grow organically (e.g. Tesla) and some will be acquired by incumbents. Comet Labs believes that many more will be born in the next few years tackling each part of the stack — and this same trend will be replicated into every major industry — agriculture, professional services, retail, construction etc. In other words… the future is just around the corner.

**So, $15 Billion is a lot of money.**

When the news of Intel’s acquisition of Mobileye was announced, many in the tech-community (and world at large) scoffed at the price tag. Especially coming so closely on the heels of Intel’s acquisitions of Nervana (for $400m) and Movidius (reportedly also around $400m). Intel has been burning a lot of cash.

Despite the shock expressed by a lot of investors and automotive veterans, I (and the rest of the team at Comet Labs) think Intel deserves recognition for how forward thinking it has been.

In fact, we believe we just seen the beginning. The first companies to be built for this new AI ecosystem are only barely hitting adolescence — that phase where you don’t really know what you should be doing and instead just play video games all day. But what the internet revolution and mobile revolution couldn’t do, AI will. It’ll drastically transform the way every industry is operated, from Agriculture to Construction, Retail to Transportation, Security to Logistics. Everywhere, companies are being built right now that will either be acquired by the incumbents, or will become the next group of Fortune 500’s.

![](https://hackernoon.com/hn-images/1*yRUoBehGQyF4t1QdZk7OVw.jpeg)

Intel CEO Bryan Krzanich & Mobileye CTO Amnon Shashua | Source: [Intel & Mobileye Acquisition Announcement](http://intelandmobileye.transactionannouncement.com/)

**Infrastructure and Applications: The Technology Seesaw**

In every technology revolution, we’ve seen **infrastructure** built first: TCP/IP, mobile communications, chipsets, browsers, smartphone cameras, to name a few. Companies building these technologies tend towards consolidation, leading to a few large companies dominating technology development. Consolidation is often a result of high development and production costs, the proliferation of standards, and the necessity for developer communities, and startups just can’t compete. (examples like Qualcomm, Intel, Google, AWS) The path to profitability is often very long, and opportunities for building meaningful businesses are rare. Most will be acquired early. Others will unfortunately fail. However, when there are sea-change moments in technology development as there currently is with AI, many new companies have the opportunity to “own” certain parts of the stack _before_ it’s established. Then valuations soar, and incumbents get itchy to acquire new “business units” in fear that they might miss the boat.

Following the wave of infrastructure development, there’s a boom of **applications** built on top. As application companies are often building for customers with picky tastes (your product has to work!), the death rate rapidly increases, and often considered a risky business to start. However, since there are typically 100x — 1000x more application companies built than infrastructure companies because the barriers aren’t nearly as high. For the right founders, who know their markets, know how to build solutions (not products) and know how to distribute them, the opportunities are enormous.

Until the infrastructure changes again.

![](https://hackernoon.com/hn-images/1*5XmAjBn2f9gLIKX_S56Nyg.png)

Image Source: [NY Times Guilbert Gates](https://www.nytimes.com/interactive/2016/12/14/technology/how-self-driving-cars-work.html) | Source: Google | Note: Car is a Lexus model modified by Google.

**Mobileye has the opportunity to become an infrastructure supplier. This fits perfectly with Intel’s positioning.**

Intel has been THE infrastructure supplier for PCs & servers basically since they were invented. However, with Qualcomm ahead with mobile and NVIDIA’s GPUs dominating the machine learning cloud platforms, Intel had to find a way to ensure it didn’t miss the next major revolution: **intelligent machines**. Movidius and Nervana both provided opportunities for the next generation of computation, at the edge or in the cloud.

But Mobileye provided something neither of them could.

In addition to chip technology, Mobileye had the credibility, channels, and uniqueness to sell to auto-makers. They are the ONLY commercial-scale autonomy or driver assistance technology supplier. More important is the community of trust and supplier relationships (read: integrations) it had painstakingly built up over past few of years.

At the same time, Intel has the infrastructure, having produced the world’s best computer chips for decades now, making it the ideal partner for Mobileye to meet the growing demand. Their success is just riding on the two teams being able to cooperate effectively across the globe.

**So how do we get EVERYTHING Autonomous?**

Mobileye and Intel are just the start. Their combined dominance will definitely be a force to be reckoned with. Intel’s production chops and management ability combined with Mobileye’s algorithm and supplier relationships make them a powerhouse for anyone trying to build autonomous vehicles. But the race is far from over.

Startups building more specialized solutions (perhaps on Mobileye’s stack in the future) will slowly start to find that the common platform or solution is not perfect for every use case. Gradually, a new generation of technology companies will appear providing even more specialized “infrastructure” layers and you can bet that Intel isn’t going to pay $15 billion again any time soon. Intel better be pumping R&D now while they still can.

![](https://hackernoon.com/hn-images/1*j_bhQv-YM2kEwXEgbXCNWQ.png)

Source: [Intel & Mobileye Acquisition Announcement](http://intelandmobileye.transactionannouncement.com/)

**What are the implications of this acquisition from an investor’s perspective? Will it scare off startups?**

Because applications for our Transportation Lab are in full swing for a couple months now, we’ve been meeting with both startups and potential corporate partners to get a better sense of their personal goals and how the infrastructure around f self-driving cars is evolving

Not too long ago when self-driving cars were just entering the mainstream, startups were attempting full-stack autonomous solutions, meaning they had developed the entire system bumper to bumper. Large OEMs (like Ford, GM, Audi, and others) quickly saw this as an opportunity to jumpstart their internal engineering teams and snapped up the startups as quickly as they could. Now, we’re sensing the trend that OEMs are going to slow (if not stop completely) acquiring full-stack startups and instead aim for startups that are highly specialized. Why would they buy another full-stack startup, if their team is already working on a solution of their own?

Enter Intel and Mobileye. Since its beginning, Mobileye has had only one focus, which is to completely master the chip inside the car that enables computer vision. They’re now seeing that their specialization strategy has really, really, really paid off.

Intel isn’t an OEM, but they do have a lot to gain from being in the infrastructure game. They can use Mobileye’s distribution channels to distribute their other products as well.

We think the acquisition is going to send a signal to other startups (there are already a ton just here in the Bay Area) that focusing on one specific component or function is where most of the opportunities are. The fact that Intel isn’t an OEM is also noteworthy — the range of acquirers has broadened beyond car brands. Major car part suppliers we’ve spoken to recently are now far more interested in startups than they’ve ever been. They know that if they can take advantage of what the startups have accomplished, they can scale that technology massively with their existing foundation and make big bucks with their OEM clients.

Looking ahead, expect to see more startups specializing in intelligent components. Car part suppliers and other companies further up on the supply chain will be seeking to acquire and jump in the fray. In the process, we’ll see a major redesign of auto infrastructure as we move closer to autonomous cars.

If this story provoked some thoughts, hit the recommend button, and subscribe to [the Comet Labs newsletter](http://eepurl.com/bHt88L) for more interesting AI and robotics stories.

Chain

Google

Intel

NVIDIA

Qualcomm

Tesla

Too Long; Didn't Read

Get hands-on learning from ML experts on Coursera 

With Intel + Mobileye, Industry Infrastructure Everywhere Is Changing

Too Long; Didn't Read

Companies Mentioned

Coin Mentioned

@comet_labs

RELATED STORIES