Intel Betting Machines Will Need Its Chips Rather Than Web

Intel Corp., the world’s largest semiconductor maker, is betting that the growth of Internet connections in everything from passenger jets to vending machines will help reduce its reliance on chips for personal computers and data centers.

More than a third of the 15 billion devices connected to the Web by 2015 will be non-standard computing machines, according to Ton Steenman, Intel’s vice president for embedded devices. They will generate a flood of data that can’t be processed in remote computing centers, and will need powerful semiconductors to analyze the information.

“If you look at the growth of the network and the growth of storage, the amount of data is far outpacing that,” Steenman said at a company conference in San Francisco.

Intel’s fastest-growing business is selling server chips to cloud services providers, yet computing over the Internet won’t be the answer in every situation, Steenman said. That opens up an opportunity as demand shifts away from personal computers toward smartphones and tablets, areas where Intel has struggled to win orders, he said.

Intel, which trimmed its third-quarter sales prediction last week amid a slump in PC demand, will get $2 billion of its estimated $53 billion in revenue this year from chips used in so-called embedded systems, he said.

The embedded division is one of Chief Executive Officer Paul Otellini’s attempts to reduce Intel’s reliance on chips for personal computers and data centers, which provided 84 percent of revenue last year.

Local Analytics

Intel, which uses the annual three-day Developer Forum to show off technology it wants computer makers to adopt, demonstrated an Web-connected Coca-Cola Co. vending machine based on one of its i7 series processors. Sporting a large touch screen display, the machine shows products and offers to take and e-mail digital photographs. While it could have worked with a less powerful chip to track inventory remotely, the more powerful chip enables the vending machine to do sales work, said Steenman.

Other examples included an intelligent traffic light that crunches sensor data and diverts traffic when an emergency services vehicle is approaching and greenhouse watering systems. Jet engines generate a vast amount of data that can’t be carried over any Internet connection, Steenman said.

“We have to find a different answer than moving it up to the cloud and trying to analyze it there,” he said. “If you need real-time response, you can’t just send the data to the cloud and hope -- you need a lot of local analytics.”

Shares of the Santa Clara, California-based company rose less than 1 percent to $23.36 at the close in New York, and are down 6.9 percent since the day before the company revised its earnings outlook on Sept. 7.

Before it's here, it's on the Bloomberg Terminal. LEARN MORE