It would be same playbook that NVIDIA did CUDA where was market 2010 when it was research labs and hobbyists doing vector calculations.