“How much does Watson cost?” I often get asked this question whenever I give a lecture or speak with clients about cognitive computing.
My response is always the same: “What is the price of not knowing?” And depending on who I am speaking to, the context could be not knowing when the next devastating storm will arrive, what the next maintenance failure on an aircraft will be, or even what’s going to be the next fashion trend.
Right now, we don’t know a lot. Petabytes of data are being created every second ofevery day, in large part by the Internet of Things (IoT). Yet, 90 percent of it, which is often referred to as data on the edge, is never captured, analyzed or acted upon — it essentially can’t be read by computers. For example, only 1 percent of the data collected from an oil rig with 30,000 sensors is examined for anomalies. To me that translates into a lost opportunity, or worse a pending disaster.
Today, I’m in Munich, Germany, where IBM is changing all of this as we bet big on a future IoT powered by IBM Watson and cognitive computing. Munich is where this new business unit will be headquartered, bringing together 1,000 IBM developers, consultants, designers and researchers — representing IBM’s largest investment in Europe in more than two decades. As the vice president for Research in Europe and director of IBM’s lab in Zurich, I’ll be supporting this new unit globally to drive IoT research across all 12 of our global labs.
Of course we aren’t starting from scratch. In March 2015, we announced a $3B commitment to IoT. More recently we announced plans to acquire the Weather Company’s IT assets. It includes a remarkable three billion weather forecast reference points, and a cloud platform supporting 26 billion inquiries per day, mostly from mobile apps. Combined with the dozens of Watson APIs available on IBM’s Watson IoT Cloud, this platform will bring the promise of IoT to both small businesses and large enterprises alike.
In China, our Green Horizon Project is using sensor data and analytics to generate high-resolution 1km-by-1km pollution forecasts 72 hours in advance. It can also make pollution trend predictions up to 10 days into the future to warn citizens and to help officials take action. Similar projects are taking place in South Africa and India as well.
Back home in Switzerland, my colleagues in Zurich are working with IBM researchers in Dublin to bring the IoT to automobiles. Cars make excellent sensor platforms since they have strong batteries and are idle 95 percent of the day. The sensors can be used to find available parking spaces or perhaps even more critically, they can detect gas leaks in a city’sinfrastructure.
So, what is the price of using cognitive computing and IoT to predict a gas leak before it can do any harm? Priceless.