Advanced processors customized for AI are converging with novel self-learning algorithms and robotics to take machine learning to the next level. The developments may disrupt incumbent chip producers while opening a potential door for China to build out its own semiconductor industry with new tech. A recent McKinsey study concludes AI-related technologies could add $3.5 trillion to $5.8 trillion annually to the global economy.
Consider the following:
Next-generation AI chips will supercharge machine learning. In recent years, neural networks running on GPU (graphical processing units) chips, originally designed for video games, have produced major advances in AI as hardware accelerators to digest reams of data. Cloud computing has made it cheap to acquire enormous amounts of data, and businesses want to leverage it with AI. However, hardware has become a bottleneck as demand for processing power used in AI projects has been doubling every 3.5 months since 2012.
Read more here...