BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Intel Furthers Its IoT Vision With Aquisition

Following
This article is more than 7 years old.

Intel took one big step further away from its past late last month with the acquisition of a Russian software company that helps computers see.

Surely you remember Intel. It is still the largest U.S. semiconductor maker, but has curiously faded from prominence. It was once at the top of the heap, riding the wave of desktop computer upgrades and exploiting its part in the WinTel alliance for all it was worth. Yet those days have passed. Businesses began migrating to a thin client, cloud-based computing model years ago and consumers moved to mobile. While Intel has built a modest cloud microprocessor business, companies in the vanguard have been adept at hobbling together custom hardware to suit their needs and snubbing the giant. Intel's efforts in mobile failed miserably, culminating with a humiliating exit in April.

To move into the future, Intel has trained its focus on cloud computing and the Internet of Things. Last year it introduced series of low power microcontroller chips called Quark to augment its powerful cloud computing processors. The addition of Itseez, a startup led by Intel alumni Victor Erukhimov and Sergey Molinov will help it develop facial recognition software for security solutions and image recognition for drones and self-driving cars. In fact, Itseez algorithms have made considerable progress identifying traffic signs, lane markers, and pedestrians to help autonomous cars avoid accidents.

Doing that algorithmically is much more difficult that you would think. Humans quickly recognize the difference between obstacles that are dangerous and those that are harmless. For example, humans easily figure out that rain or bugs hitting the windshield are harmless. A computer has to first identify the obstacle, classify its threat value based on all that it knows about rain or insects in the real world, make an informed decision and then automate. And that’s an easy example. Consider the dilemma when the obstacle is live, like a deer or worse, a human being. These are complex problems we humans, for all of our foibles, are better at solving than computer algorithms.

Morgan Stanley estimates self-driving cars will lead to $507 billion in annual productivity gains worldwide so competition is fierce. Carmakers are scrambling to implement systems. Intel, with its longtime support of Linux, held an early advantage given the automotive industry preference for open standards. The Linux Foundation release of Automotive Grade Linux should further that lead. The most notable rivals are QualcommNXP Semiconductors and most importantly, Nvidia. Intel's strong history in computer graphics and modular approach is gaining traction in the industry with carmakers such as Tesla.

Yet the industry giant still lags behind. While Qualcomm and Intel are building software applications on top of their powerful modems and processors, Nvidia is offering a plug-n-play solution called Drive PX-2, essentially a supercomputer bolted into the car. Flipping a switch begins monitoring of the car’s position, the objects around it and its speed. All of this data is then shipped back to a cloud where it can be analyzed and used to build new models.

Although Intel has had a tough time making the transition from desktop PCs to the cloud and Internet of Things, it’s a move the company must make. Self-driving cars are going to be one of the most important pieces of that transition. Itseez helps Intel remain competitive with its faster rival Nvidia. Intel shares are quiet now, but keep an eye on them because in the past they have had a habit of resurrection when least expected.

Check out my website