LIVE FROM HUAWEI CONNECT, SHANGHAI: As the connected world shifts to data-hungry Internet of Things applications, Intel CEO Brian Krzanich forecasts a data flood that can only be supported by 5G networks.

Today’s networks are built for “people devices”, with average daily data usage at around 600-700Mb, he said. That is forecast to jump to 1.5Gb per day by 2020.

But that is just the start as more and more machines and IoT devices are connected. “The average smart hospital will produce 3,000Gb a day, and the average autonomous car will produce 4,000Gb a day,” predicted Krzanich, who delivered a keynote at the event this morning.

The average connected car in 2020 will create about as much data as 3,000 people. “That’s a huge network load but also huge opportunity to optimise networks and systems and create value.”

He said Intel believes this will require a 5G network, which is more than an evolutionary step forward from 4G.

Huawei rotating CEO Guo Ping agreed, saying “we need to have many breakthroughs to achieve this capability. Without these we won’t see a move to autonomous vehicles.”

Intel is collaborating with Huawei on a broad range of 5G technologies for the ecosystems of the future. “Using Intel technology, Huawei is building end-to-end cloud-based 5G infrastructure to support global 5G trials,” Krzanich said.

Indeed, Intel has been something of a laggard in the 3G and 4G mobile world, and knows that future success is dependent on playing a key role in development of 5G technology.

Cloud have-nots
A world where everything is smart and connected represents a huge opportunity for industries and companies that embrace digital transformation requiring the cloud, Krzanich said.

“I believe that the world will be divided between those who embrace the connected, data-driven cloud world and those that don’t. Those that do will have access to more information and will grow fast, and those that don’t will fall behind.”

Demonstrating Intel’s capability from the edge of the network to the cloud, he explained that Intel’s Curie module has all the elements needed for edge-sensing capability. Previously a microcontroller at the edge would stream data from sensors and another computer elsewhere would process and interpret that data. “But today with Curie everything is done inside the chip.”

It not only has the ability to gather sensor information and to connect to other sensors, it has a pattern-matching engine and cloud tool, called Knowledge Builder, he said. This allows any developer to train Curie to identify specific movements and process analytics on site.