The technology industry is exploring new models of innovation To boost technological developments as provided for Moore's Law

Gordon Moore, cofounder of Intel and Fairchild Semiconductor engineer, claimed in 1965 that the number of transistors per same size circuit would double, at constant prices, every year.
In 1975, Moore reevaluated its prediction by assuming that the number of transistors on a dual microprocessors silicon chip every two years. Between 1971 and 2001, his theory has been verified. Accordingly, the electronic machines have become less expensive and more powerful. However, he said in 1997 that the growth performance of the chips would run around 2017 to a physical limit: the size of atoms.

Today, reality shows, the literal definition of Moore's Law reached its limits. Processor speed is no longer increasing at the same rate, and it becomes more expensive to get on silicon performance. The single silicon scale no longer provides cost effective innovation and can not meet the performance requirements of the Cloud, Big Data and analysis.

Of course, Moore's Law is not dead, but the IT industry think the concept existed long before the theory formulated by Gordon Moore. It then evolved over time and will continue its evolution after the end of a silicon era.

Some companies have therefore met to develop a new model of innovation. The idea is to find innovations to ensure exponential growth as envisaged by Moore's Law through different technologies.

The initiative, led by OpenPower Foundation is built around a new innovative chip that offers high performance that customers need. By collaborating on new solutions that can be customized in data centers for new workloads such as cloud services or data analysis, OpenPower Foundation creates an ability to customize systems to meet specific needs .

For example, OpenPower has developed a high-speed interconnect technology to connect its graphics processing unit (GPU) to the microprocessor from IBM, allowing data to flow in both directions between them and extremely fast. By granting the license to members of the OpenPower Foundation, this could lead to new systems that are up to 100 times faster than today.

Likewise, opening a new standard I / O that allows communication with flash devices with a much lower software overhead, partners and IBM OpenPower (Redis Labs, Altera and Nallatech) created a database server NoSQL data that can reduce about 5 times the cost of storage by replacing DRAM technology by flash technology to the same levels of performance.

Efforts are underway to find out what will come after silicon. And these efforts could enable us, in 10 years, to have systems completely different from what we know today.